case study on demand of coffee

How Starbucks Became Everyone's Cup Of Coffee

Table of contents.

Starbucks Coffee Company boasts impressive stats:

• Owns 40% share of US Coffee Market
• Earns $24,72 billion worldwide
• Has 29,324 stores worldwide in 72 countries
• Over 14,000 of total stores in the United States / over 27,000 worldwide
• Conducts over 90 million transactions per week
• So popular in China, a new store opens every 15 hours
• Following McDonald's as the most valuable fast food brand worldwide (valued at $44.5 billion)

It will be very hard to achieve something Starbucks did since 1971 when the company started. There’s a lot of firsts when it comes to the company. First to introduce the new coffee culture, the first privately owned company which offered all their employees health insurance AND the share of the company.

The CEO, Howard Schultz, who might even run for president at some point , achieved something that is almost impossible — appeal to shareholders, employees, and customers at the same time. This is my giant case study on how to achieve world domination in case you want to bring an old product to the new market.

{{cta('eed3a6a3-0c12-4c96-9964-ac5329a94a27')}}

The Starbucks Idea

The coffee culture in the United States before the 80s was nonexistent. 

Americans were used to huge cans of ground coffee and they couldn’t care less about the flavor. Even if you’d go outside your household to a dinner you would be met with a generic drip coffee or styrofoam cups of foul-tasting joe at the workplace. No one even thought about the flavor, the origin, or anything more sophisticated tied to the drink.

The 70s coffee culture didn’t exist at all.

In 1970 three college friends: Gordon Bowker, Jerry Baldwin, Zev Siegl went into the coffee business together. They set up a shop and sold roasted beans. They received the knowledge from a man named Alfred Peet (if that rings a bell, yes he is the owner of Peet’s Coffee). Alfred was one of the most knowledgeable people in the country about coffee. He knew where to source it, how to roast it. He was the first to introduce dark roasted and french roasted beans.

In 1971, the three friends opened the roastery and bean shop in Pike’s Place, Seattle’s famous tourist destination known for the Pike’s Public Market Center. Peet helped the young entrepreneurs by providing them with beans and connecting them with reliable bean providers.

The name Starbucks stuck because it’s easy to say, impossible to misspell, and has a vaguely British overtone to it. Really, we picked it because our lawyer called and told us we had to submit papers and needed a name. We didn’t know at the time, but Starbuck is the name of the first mate on the Pequod in Moby Dick. That might explain the siren logo. Some might even say it comes from Mount Rainier's Mining company Starbo . According to Gordon Bowker, they were initially going for the name Cargo House Coffee .

The business was successful enough for the trio so they opened 4 more shops in Seattle. However, no coffee drinks were being served. This was still a roasted bean retail shop intended for home use.

At that time Starbucks was competing against instant coffee cans. The quality was stark and thus the business went well. Things were about to change when the founders hired the head of marketing and sales, Howard Schultz in 1982.

The Inclination for Grit and Determination Fix Social Injustice

Howard Schultz was a child raised in poverty. After seeing his father injuring himself doing grueling manual labor, he decided he wanted to get rid of the injustice of the working class. An idea of creating and striving for an environment where employees are fairly compensated and taken care of has been set in.

In Masters of Scale interview with Reid Hoffman, Schultz described seeing his father stretched out on the sofa after suffering an injury. Howard Schultz swore to himself to make a company his father had never worked for.

“I saw my father losing his sense of dignity and self-respect. I am sure that this was caused mostly by the fact that he has been treated as an ordinary working man.” – Howard Schultz, AstrumPeople article

Schultz started working at the age of 12 selling newspapers. Since he was being athletic, Howard earned an athletic scholarship at Northern Michigan University where he received his Bachelor’s degree in Communications in 1975.

After his graduation, Howard Schultz spent three years as a sales manager at Xerox, and then he started working at a Swedish company Hammarplast , where he was selling home appliances, including coffee grinders to businesses like Starbucks.

The Starbucks founder trio took him amidst to grow the company.

In 1983, Howard Schultz gets an epiphany. He travels to Milan, Italy for some sort of conference and what he sees there changes his perception of coffee forever. In certain European countries, especially Italy, coffee was one of the more important things in life. It served as a social lubricant and the third place of dwelling between home and work. Schultz discovered what it means to have a high-quality espresso served in a proper way in a relaxed environment.

He was determined to bring this piece of coffee culture back to the United States. The founders gave in after continuous pressure from Schultz to open an espresso bar. Eventually, they gave him an opportunity to open up a coffee bar inside a store. It was incredibly popular. But the owners didn’t want to turn the coffee retail business into a cafe.

“After Milan I flew back to the United States, excited to share what I experienced. But my bosses, the first founders of Starbucks, for whom I had tremendous respect, did not share my dream of re-creating the coffee bar experience in Seattle. I was crushed, but my belief was so powerful that, in April 1986, I left Starbucks and raised money from local investors to found my own retail coffee company. I named it Il Giornale after Milan’s daily newspaper.

In 1985 Howard Schultz opened his own cafe chain - Il Giornale . He wanted to pursue the dream and went back to Starbucks owners and offered to buy all 6 stores that were operational at that time. With the help of raised venture capital he succeeded and became the CEO after the successful acquisition with $3.6M.

The hyper-growth began.

Key takeaway #1 — change is good

The determination and unrelenting belief to change the current situation is not just a helpful attribute but a prerequisite for cultural change. Staying true to the “one thing” without flinching will be the cause and the driver of change.

“An Old Product in the New Market”

Whenever something works out on an incredible scale in one market, there’s a potential of seeing it succeed in a new one. This is called introducing an old product to a new market.

For example, Uber and Lyft built an incredible business about ride-sharing. Because they have to contain the growth before they are spread too thin, that gives the opportunity to copy-cats in different markets. In the United Arab Emirates, you have Careem ( just recently acquired by Uber ), in Croatia you’ve got have Cammeo and in India, you’ve got sRide .

After something experiences great success, there is only a matter of time before someone else sees the potential and brings it back to the new market, and starts eating out the market share

Coffee was a big opportunity in the United States at that time. Howard Schultz saw with his own eyes how effective and important it is in Italy and he knew he could do something similar in the United States. To perform a similar innovative (for the new market) service you would need to take the entire concept and localize it to the new market.

Even the trends from 2004 to this day shows an upward trend in coffee:

This go-to-market product strategy was first introduced In 1957 by Russian American mathematician and business manager Igor Ansoff. The Ansoff Matrix was published in Harvard Business Review in the article “Strategies for Diversification”. In his opinion, there are only two ways to develop a growth strategy — varying what is sold (product growth) and to whom it is sold (market growth).

Market development — new market, existing product

The Starbucks go-to strategy was to bring the already established product in different cultural and geographical spaces into the new market — the coffee-culture deprived United States.

Howard Schultz’s task was to closely observe how Italians treat the product and figure out a way to bring it home with minor changes. It was impossible to expect that the new market is going to slurp macchiatos from tiny espresso cups but everyone could understand comfort and better quality. That was going to be Starbucks’s trump card.

Market penetration — old market, old product

The most obvious strategy is to sell the existing product to the existing market. With this concept there’s a little risk since the companies don’t have to educate the market with the new product, however, the growth is inhibited by competition or decreasing trends.

Diversification - new market, new product

By far the riskiest approach is introducing a new product in new markets. Not only the product needs to provide clear values, but it also has to educate its use in the new market.

Imagine bringing augmented reality technology to a country where there’s no practical use for it yet. Since there’s a great risk, it can also result in amazing success where you’re the only provider in the blue ocean market.

Most of the startups are banking on this strategy.

Product development - old market, new product

This strategy is most often used by established brands that are already known as leaders in their field. If a washing machine company introduces a new technology that also folds your clothes after washing and drying, that would be much easier to understand and adapt to their existing users.

Key takeaway #2 — do market research

When developing the new market, learn as much as possible about the product itself in the location where it’s mostly used and established. Identify all the major benefits and think of the most significant values that would succeed in the new market

Eco-Conscious, Friendly People, and Profitable — Starbucks’ Triple balancing act

Howard Schultz had an idea to build something that is almost impossible to imagine and can exist only in Utopia. From the start, he wanted to serve with equal importance towards customers and employees.

This is almost impossible to achieve since on one end the business investors want to see money coming in, which in most cases means lean running staff with lower wages and higher-priced products. The staff, or “ partners ” as Howard Schultz calls its employees, are not only compensated a fair wage ( between $10 to $15/hour according to Glassdoor ) but also have healthcare insurance and discounted stock options for company shares.

Howard went even further, offering full tuition coverage through Arizona State University's online degree program .

This idea was most likely outrageous to shareholders. Everyone will get a piece of the company’s pie?

In a Tim Ferriss interview with Jim Collins, the author of Built to Last and Good to Great mentioned the final lesson of his mentor and all-around management superhero Peter Drucker:

“The management isn’t about being more efficient all the time, but it’s also being more humane at the same time.

Striving for workplace quality for the employees was thus one of the main values the CEO implemented in the company.

The interesting analogy is Jordan Peterson’s theory of order and chaos (yin and yang) where one side represents the profit that company must achieve by ruthlessly cutting back the cost in the workforce and the other side where the conscience of doing the right thing for your people brings satisfaction and peace to the workplace which is a proven necessity for customer-facing businesses.

Key takeaway #3 — happy employees make happy clients

Treat your people well. When you’re in the service industry the customer satisfaction and treatment are at times more important than the actual product. And happy employees make happy clients.

The Product

Better coffee.

To coffee drinkers, there are not a lot of things more important than a good coffee in the morning or during the day. By today’s standards, Starbucks drinks aren’t at the level of barista artisans and coffee aficionados. But when the shops started opening in the early 70s, 80s, and 90s, the espressos and lattes were vastly different from all the other stuff people were drinking.

Coffee is generally roasted in three ways: light, medium, or dark, depending on the time dedicated to the coffee beans’ roasting.

In a light roast, you would notice a fruity and acidic taste. Coffee beans are actually considered fruit and are sometimes called cherries. That is the reason you taste light roast as acidic with fruit notes.

In Medium roast, the coffee tastes the sweetest. The glucose levels reach the point where the glucose starts to break. Coffee roasters would say the medium roast is the most balanced since it’s not bitter nor acidic but something in between.

In dark roast, you can taste the bitterness due to burned beans.

Coffee quality comparison

Starbucks predominantly use dark roast coffee which also represents the majority of the coffee that is being consumed in North America. As mentioned, the coffee quality was much better than instant abominations in the early 80s; however, it definitely cannot measure up to artisan roasters.

There are two main reasons:

1— Dark roast is cheaper and can be produced in mass quantities. Similarly to green tea, the light roast-worthy beans are grown in shady, high-altitudes where it produces the most sweetness. High-quality matcha (powdered green tea leaves) is intentionally kept in the shade so it produces more photosynthesis and better taste. Since Starbucks has to supply tens of thousands of shops, they have to bring the mass supply to the cafes. Brian Stoffel from El Toledo roastery in Costa Rica says: “It would be financially stupid for a large chain to buy high-quality coffee beans and use them for dark roast coffee.”

This brings us to…

2— The coffee has to taste the same across the cafes to guarantee uniformity. With dark roast, the flavors of the beans are getting covered up in the same way as overseasoning a dish or overcooking a steak.

But it wasn’t just about the coffee alone. The branding kicks in and people pay for something they want to eventually become. Drinking Starbucks drinks meant they are sophisticated, culturally progressive individuals who enjoyed the premium experience of coffee-drinking culture from fashionable Milano streets.

The slim and elegant takeaway cups proudly wore the green siren logo so the passers-by noticed the person drinking that exact coffee. These cups were different from styrofoam cups in the office or fast food joints.

A similar tactic was used by Apple with the launch of iPods and white earbuds. The iPod was a cool new gadget you had to wear to be relevant in modern society. Apple made it in such a way that people noticed which users had iPods — because they plugged white earbuds into them.

This was a genius idea because the users were immediately differentiated from other less-cool mp3 gadget-using people. Secondly, this was a perfect silent word-of-mouth strategy. If local influencers were seen using white earbuds, everyone else wanted to get on that trend. This strategy is viral in concept and is used by many companies; however, it’s harder to implement it on a distinctive level.

Later on, Starbucks adapted to the marketing with something called “horizontal offer”. It wasn’t just about the dark roast and espresso shots. Young budding students wanted something sweet and mocha just hit the note between coffee and rich chocolate fudge. Why not having both in one product?

Later on, Starbucks started offering teas and snacks. Snack is bringing in a substantial amount of revenue. The shops are using the display of sweet pastry or savory egg sandwiches like any expert pastry shop in Europe. And there are not many people who can resist a croissant, cinnamon roll, or blueberry muffin with their americano or latte.

The food is bringing in more than 20% of all revenue . The pasty was the start, but the company followed up by offering breakfast sandwiches. The adaptation to the market goes even further.

With the recent diet trends in health and fitness, Starbucks has you covered with gluten-free, protein-rich snacks.

With all the addition and expansions to serve a larger audience, it’s inevitable to create resistance groups who blame Starbucks as a commodity coffee provider. And they would be right, it has become that because their system of sourcing beans has to ensure the stock supplies for thousands of shops. But by becoming the main coffee dealer to the masses all the micro-roasters and man-bun wearing, tattoo-sleeved barista artists can fall on their knees and thank the mighty green Siren for creating a market for them.

The need for coffee has increased substantially with the introduction of better coffee, so it created another pocket of niche providers of premium roasted bean roasters.

The price of a cup

Most of the coffee shops live well because they can afford hefty margins. An 80% markup is a standard in the coffee business, especially on the higher-end brews. According to the Small Business Development Center’s 2012 report, food costs take up about 15 percent of revenues on average. The average coffee shop then has a gross margin of 85 percent.

Starbucks margins must be pretty loaded then since they buy tons of coffee from a few sources. According to Coffee Makers USA, the actual coffee in a grande Starbucks cappuccino costs about 31 cents.

For a commodity product such as coffee, Starbucks drinks are quite up there on the more expensive tier ranging from $2.15 for a tall drip to $5.95 for a seasonal frappuccino concoction. But taking into consideration the physical positioning ( Chapter 5 — Coffee Locations ), paying off employees and staff the actual margin per coffee sold are 7% .

Historically, Starbucks has been raising the prices per cup over the years. Since it has poured a lot of equity into maintaining the brand image, it can afford to have a steeper price than its competitors (McDonald's and Dunkin Donuts). Instead of losing the price-sensitive customers, Starbucks differentiates itself from the before-mentioned companies and thus keeping the brand image of a premium java provider.

However, as Tucker Dawson from PriceIntelligently mentions, the prices aren’t increased across the whole product offering . The high-margin items have stayed the same.

Product differentiation

By having a strong and recognizable brand, the company can afford to put out merchandise. Starbucks holiday-themed mugs and localized artwork on them are a big part of the exposure. The merch cabinets and tables are usually near the counters or areas where there’s a longer dwelling time.

The revenue isn’t coming just from the beverages alone. Starbucks did an amazing job of offering non-caffeinated beverages including kids drinks and teas which were introduced after partnerships or acquisitions of Tazo and Teavana.

Starbucks started to diversify its products, pushed them into retail space, and also added teas.

The big drivers are also snacks, wholesale beans, before-mentioned merch, and coffee equipment.

Key takeaway #4 — diversify and expand

While the product is one of the key components of a successful business think about its potential upgrades. Keeping the core you can diversify the offering (and acquire new revenue channels) by expanding into different verticals but staying inside your core company values.

Experience is More Important Than The Product Itself

With a distinctive brand identity, Starbucks shops are easily recognizable anywhere in the World. For a global brand, this is one of the mandatory elements. Each franchise is slightly different than the other — Starbucks in the posh downtown area will have a different feel than the one on the Student campus or at an airport.

But each store follows certain guidelines which are prescribed. In tech and startups, product development follows a concept called minimal acceptance criteria . In other words, what are the lowest common denominators the dev team needs to do before it can be rolled out as a published version.

For Starbucks Cafes, even though the store managers have a certain freedom to run and maintain the facility, they have to ensure to deliver the core Starbucks qualities.

• Indie playing music
• Comfortable (community) tables for remote work
• Reliable wireless connection
• Charging Outlets

These shouldn’t just be taken for granted. People love some sense of predictability in their lives. How many times have you been on the lookout for Starbucks when visiting a new country just to take advantage of their wi-fi connection and use of restroom? From that perspective, Starbucks serves as a transactional facility offering other services which don’t have much to do with coffee.

The main idea is, coffee is not the product that is being sold at Starbucks cafes — the whole thing is a social experiment of creating a meeting place between people. It serves as some sort of oasis for meeting up with friends, having a snack and a cup of coffee in a comfy chair while listening to the latest Indie playlists . Starbucks is less in the coffee business as is in people’s business as well.

“It’s not Starbucks coffee you are getting, it’s the Starbucks experience. “

By calling your name and writing it on the cup, it doesn’t just inform the customer that their drink is ready. It allows a more personalized service since we love hearing and seeing our name.

Smells and sounds

Starbucks Sounds

Chances are when you go to Starbucks you don’t ever hear the music. But it plays an important role nevertheless. Starbucks playlists are carefully curated to help create that ambiance of a neighborhood coffee shop.

It has been a piece of the Starbucks experience for over 40 years already . The songs and tracks are carefully curated way ahead of time. These handcrafted playlists usually consist of indie, feel-good songs, pop, alt-country to season-themed or even classical playlists during holidays.

In 1999, Starbucks even acquired a Bay Area music store to launch its own branded coffeehouse and later on, even a record label. In the early 2000s, Starbucks sold CDs in the store until the format decline. In 2016, Starbucks partnered with Spotify . Through the mobile app integration, Spotify plays music as part of the app. In-store listeners can take a look inside to identify the artists and save the tracks to their playlists.

Holly Hinton and David Legry, the in-house music curators, are responsible for what gets played. What sounds like the best job in the world, actually is. Their sole work is searching for the right tracks and artists that they can see are fit to be played in the coffee shop.

In an interview with Fast Company , Holy Hinton said:

“We want our customers to walk in and have a ‘What’s that song?’ moment. We want them to hear interesting, cool music that they might not hear when they turn the radio on. It’s music that we think is cool and would sound beautiful in the coffee shop. It’s the music that we’d want to hear on Sunday morning when we’re reading the paper and drinking coffee. It’s a friend-to-friend personal. And we’re lucky to be able to be a part of that.”

To localize the experience, every region is slightly customized regarding the music, while still carrying the same vibe Starbucks customers are used to. This way, whenever a customer comes to the cafe, within the first few seconds, they feel accustomed based on the music alone.

The interior design

Every piece of furniture and interior is carefully planned to conform to the standards of the homey coffee place.

To get their store right, Starbucks interviewed hundreds of coffee drinkers to get as much information which they could use to build a perfect coffee shop. The overwhelming consensus actually had nothing to do with coffee; what consumers sought was a place of relaxation, a place of belonging.

If we go back to Howard Schultz’s deciding moment from the Milanese coffee shops, it shows he managed to do just that. Create a community space as a second home. It’s somewhere where people meet, it’s where you can take someone for a first date or even get some work done at the large community table.

In the book Starbucked, freelance journalist Taylor Clark claims, that “The round tables in a Starbucks store were strategically created in an effort to protect self-esteem for those coffee-drinkers flying solo. After all, there are no “empty” seats at a round table.”

If we looked at the interior, the counters, chairs, and wardrobes are built out of natural materials like warm woods and stone. In some stores, you would find cozy armchairs as well. With the Shared Planet initiative , they doubled down with environmental sustainability in mind and employing local craftsmen to do the job. The stores are built from reused and recycled materials wherever possible.

Most of the new stores that are being built are a part of the LEED Certification program (Leadership in Energy and Environmental Design).

Starbucks differentiates from three general looks with the addition of concept designs:

• Heritage coffee houses reflect the history of the place where the store is located. At Pike Place, the coffee shop reflects the merchant trading roots with worn wood, stained concrete or tiled floors, metal stools, and factory-inspired lighting. Even more sophisticated is the New Orleans inspired coffeehouse showing the rich music history.
• A “Louisian merchant in the early 1900s” inspired heritage coffeehouse with vintage trombones light fixtures. Located in French Quarter, New Orleans.
• Artisan stores echo the industrial past of urban markets, taking inspiration from the Modernism of the 1930s. This motif celebrates simple materials like exposed steel beams, masonry walls, factory casement glass, and hand-polished woodwork in a creative gathering place for culture and the arts.

• Regional Modern are localized stylized coffee shops. The interior is spacious. comfortable and welcoming. The bright, loft-like, light-filled spaces punctuated with regionally inspired furniture and culturally relevant fabrics create a calm and contemporary respite from the clamor of the fast-paced world.

• Experimental — with growth and a plethora of locations comes more daring and innovative designs. Unique designs such as the reimagined drive-thru in Colorado , the Swiss Train contemporary mobile coffee space from Geneva Airport to St. Gallen or one of the beautiful Shinto shrine-inspired coffee shops in Japan

‍ Starbucks Reserve

To combat the upscale coffee market which ironically has to thank Starbucks for creating fertile grounds of demand for premium coffee, Starbucks started opening up so-called Starbucks Reserve stores. These are luxurious, beautiful, and magnificent stores where they roast premium, rare beans and experiment with different brewing techniques.

CNN Money described the store concept as "an open, marketplace-style" with a Princi bakery counter, a full liquor bar, and a Reserve coffee bar, with tables, lounge areas, and two fireplaces.

"Our Reserve store takes the best of coffee craft as well as artisan baking and layers in a marketplace-style customer experience creating a space that has both energy and moments of intimacy," — Liz Muller, VP of Creative, Global Design & Innovation at Starbucks

Coffee shop locations

In any high-traffic area in the city where Starbucks is located, you almost have a feeling their shops are everywhere. You would be partially right — Starbucks are strategically located in areas with high appeal. Similarly to Walgreens, Starbucks chose the concept of the convenience store, always located in an area of larger foot-traffic .

Source | A snapshot of Starbucks shops in Seattle

Arthur Rubinfeld who is responsible for Starbucks’ location selection, explained there are about 20 or so analytic experts around the world who are assessing different factors of the appropriate area for the new Starbucks shop .

Key takeaway #5 — spoil your customers

Think beyond the product and identify what else can you do for the customer to add you in their daily, weekly routine. Customer support excellence is mandatory, so think further and in the direction of the place’s ambiance including smells and sounds.

Breaking down the Brand and Messaging

Bill Macaitis, former CMO of Slack said it best - “The brand is the sum of all customer touchpoints your customers have with you at any point”. With the food and beverage category, this is even more important.

By introducing and creating a culture of coffee drinking, Starbucks had a major opportunity to create intimacy with the customer. In Italy, coffee culture is a part of every day and the same culture was slowly getting familiar to the new audience.

Because of the personal nature of coffee and frequency of visits, this relationship-bonding happened much faster than in other fast-food joints, especially since in the early years of Starbucks there was no competition.

Brand and product

The bright white cups with the green siren are the first noticeable brand. But it goes beyond that. You will notice that Starbucks never offers any sort of discounts or actions like buy-one-get-one-free. That’s sort of action dilutes the premium feel of the brand. You can get a free coffee drink for your birthday, but the underlying reason for that is for a customer to develop a positive connection with the brand and company.

The most valuable assets of the regular Starbucks coffee shop can be broken down:

☕Free reliable Wifi - besides oxygen, water, and sleep, the online connection has become a necessity in modern civilization. Whenever you’re in a new place and you need to connect, one of the first options would be a Starbucks shop.

☕Comfortable seats and community tables - whether you’re there to take a breather or putting some hours of online work or organizing an impromptu study group, there’s a Starbucks location that can provide those demands. Most of the Starbucks are generously equipped with charging outlets as well, so you can get another drink after your focus is starting to drop… and then another… And another...

☕Friendly baristas - customer service is ingrained in the retail work description yet rarely done the right way. With L.A.T.T.E. method (Chapter 8 - Disciplined Action) and general training of Starbucks partners , each interaction with the customer is there to provide a positive experience. Calling people by their name, timely service, and the patience of crafting ridiculously complex drink orders .

☕Brand colors and materials — the nature-influenced interior with dark colors and wood finishes are giving a feel of hominess. Sometimes a Starbuck visit is just a pause you take in a day to relax your eyes.

☕Music and smells — coffee and snacks just smell amazing. Let’s take that for granted. The music serves a purpose as well as bringing an ambiance that is great for having a conversation or focusing on work (or your date).

Key takeaway #6 — positive interactions

The brand is the sum of all touchpoints the customer has with the company. This goes beyond the product and customer service. Think about every single interaction customers have with you and make them positive.

Starbucks Master Example of Mobile Retention and App Rewards

Starbucks mastered the mobile game at the right time. Dabbling with mobile technology since 2007, Adam Brotman spearheaded the platform to maximize the effect. The big challenge was to align it with the brand.

“We don't look at mobile in a vacuum. We have an overall digital strategy that's all about building relationships with our customers, and that strategy runs across a number of digital touchpoints. We're looking at mobile, Web and social to think more holistically about how we engage with our customers and tell our story." — Adam Brotman, Chief Digital Officer

In the Manifest survey in 2018, 500 smartphone owners rated their satisfaction using food apps. Starbucks had the most popular and regularly used loyalty rewards app — 48% of users used it on a daily basis.

Four years later, Starbucks remains one of the most popular apps, ranking number 6 on the list of most downloaded Food & Drink apps. 

The mobile switch paid dividends with time. Instead of support and enhancing physical visits to the store, the channel began bringing in 23% of all the revenue.

Ordering ahead of time and user experience

For a food mobile app to be successful, it must bring value to the user, be easy or even fun to use and it should have entertaining, dynamic content.

The design has to adhere to rules of the brand, achieve a consistent visual look and continuity across all touchpoints.

The mobile app design is no different than the rest of the materials Starbucks uses.

Digital Engagement paid tremendous dividends for the company.

Starbucks CFO Scott Maw said almost all of the company’s same-store sales growth has come from customers that have digital relationships with the company and those that are in the Starbucks Rewards program.

User-friendly design

This is the minimal and easiest thing to leverage on. With a strong brand, it should not be hard to create an appealing visual interface and create logic flow and transitions or continuation to the desired action.

Engaging loyalty program

Retention is the name of the game. If a customer trusts you well enough to download your app, you have a unique opportunity to convert him or her to be a regular user.

Starbucks has a similar strategy with the reward system. Every day there’s a slight reward, whether it’s collecting points or showing the current mouth-watering warm drink inside the app. It’s sticky and you can’t help but wish for a warm beverage.

Mobile pay and ordering

The North American market is known for heavy mobile use . By prepaying and using the device to quickly go through the ordering process, the customers feel more efficient and slightly more an advantage than the other poor souls who still buy their coffee with credit cards or cash.

Integration with other platforms and services

Partnerships are ways to get tons of new users with one big swoop. Spotify acquired one million users a few days after partnering with Facebook (Source) and Facebook had one sexy product update from it as well. For similar reasons, Starbucks used Spotify to enrich the experience of the mobile app.

Now playing highlight in Starbucks stores (Music is a big part of the brand and having perennial "Shazam" embedded brings seemingly insignificant, yet positive experience.

UX/UI — breaking down the mobile app design

Out of this world personalized experience.

The app remembers your favorite order. This is ingenious. We’ve mentioned how coffee represents a daily habit - if Starbucks manages to infiltrate itself into your habit loop, they’ve won. They have become a part of your daily routine. Stacy always stops at the same drive-through Starbucks, orders her Grande Latte with Soy Milk at 6:15 am before she checks-in at her job. When that’s her daily or even only a few time per week routine, the LTV for that kind of customer is absolutely amazing!

Every little detail counts. For instance, here’s the customized greeting each time a user opens the app’s Home tab.

Gamification

Most addictive phone games always give you something to do if you’re not using them for a while. From Candy Crush Saga to Supercell’s engineered mobile drugs like Clash of Clans and Boom Beach, the mechanics of engagement are carefully predicted for maximum time and cash spend. These games start with low difficulty. They are fun, colorful, and offer an entertaining introduction to their mechanics. But you can play all day, and after a while (on a free tier) you’re locked out of the game.

To continue playing, you can either (literally) buy your time or increase your chances of success with extra loot, power levels, or something similar.

Starbucks uses a similar principle of gamifying its mobile app. There’s a lot of value upfront (pay with a card, skip the line, earn credits for free drinks) but it serves the company’s profit. You get hooked to those stars (credits) which are stacking in your beautifully designed mobile app.

There are also challenges for extra Starbucks points (who can say no to double credit days?)

With the app, Starbucks gets you to try new products and thus increase the range of products you are consuming AND it gives the company an opportunity to increase the average order revenue per customer.

There’s a thin line between being overbearing and being just enough engaging. And at the same time, they have to be very strategic on the number of features offered. Sean Ellis , the OG Growth hacker said the product is ready to ship once all the unnecessary features are taken away (kind of the same mentality as per good design). Luckily with MILLIONS of users, Starbucks can apply some Data Science magic and figure those timings for every type of person.

Personalization goes even further - it tries to give a similar experience as to visiting the store ( source )

Starbucks Loyalty Program on triple-caffeine nitro power

The Starbucks Rewards are dead simple - the more you spend the more stars you get. Besides the stars, the rewards program offers birthday rewards, phone payments, paying ahead, free in-store refills and special offers and events for members. As expected the experience is personalized for each user.

The Rewards work like gangbusters! More than 14.2 million active members in the U.S. are invested in the loyalty program and the mobile strategy has seen an 11% growth in users in Q2 2018 . The gamification of the program and “spend more, earn more” in some cases represent 39% of the entire chain's sales .

Here’s what’s ingenious about the mobile program. Even though there are people who prefer to have the minimum number of apps on their phone and think twice before opening the doors for the elite club on their smartphone storage, the Starbucks app is a trojan horse of benefits - even if you don’t care about collecting stars, it’s tough to say no to the free birthday drink or the convenient mobile pay.

Online Ordering and easy payments flatten the friction of getting the product. Just like the Amazon 1-click purchase or Slack’s onboarding sequence , the same goes for picking up a mocha and Petite Vanilla Bean Scone. At first, Starbucks had some issues, since the mobile members had to wait in line just like the others, but Starbucks responded by adding dedicated stations for mobile order-ahead customers.

Members can skip the waiting line and enjoy the jealous looks while feeling elite of themselves.

The beauty of the app isn’t giving one big benefit of a quicker caffeine shot to the member, but it serves as an upsell marketing tool. The Starbucks app is a delivery method for presenting new items ahead of time. These generate interest and coupled with email notifications, it gives their customers something to look forward to.

To keep the retention flat, the Rewards program has “punishment” traits tied into it. If you’re not using the stars for visiting the cafes you start losing them. This psychological trick, known as The Endowment Effect , helps to nudge those people who are affected more about losing something they already have.

The Mobile part is one of the main drivers of customer retention and has proven to raise the average order size per customer. Since the frequency of orders and visits is so high, the LTV per customer contributes to that impressive double-digit growth in the first years.

Key takeaway #7 - APP A mobile app for a product that is being used on a daily basis and is in the lifestyle category is not a nice to have, but almost mandatory. If you want to stay a part of your customer's daily lives, bring the entertainment, rewards, and gamification to keep retention and customer satisfaction high. You will be rewarded with increased LTV.

The Success Flywheel of Starbucks

The easiest way to figure out and identify the success of a company is to apply the try-and-true framework. Jim Collins, the author of Good to Great, Built to Last claims all mega-successful companies have to figure out the Flywheel principle .

To become an unstoppable juggernaut in its own field, Starbucks had to align 5-6 different elements in three categories:

• Disciplined People
• Level 5 Leadership
• First Who… Then What
• Disciplined Thought
• Face the Reality
• Hedgehog Concept
• Disciplined Action

Culture of Discipline

• Leveraging the Technology

Imagine the concepts as drivers of one giant flywheel. Let’s say you’d want to move a giant stone wheel that sits on an axle. It would take a lot of effort to get it moving at first. After gaining speed it would need less and less power to keep it going. After gaining momentum, the same wheel would run on its own with little interaction. Just like the extremely simplified quote says; “If it ain’t broke, don’t fix it.”

Disciplined people - Starbucks Level 5 Leadership

Excerpt from Good to Great -> “Level 5 leaders display a powerful mixture of personal humility and indomitable will. They're incredibly ambitious, but their ambition is first and foremost for the cause, for the organization and its purpose, not themselves. While Level 5 leaders can come in many personality packages, they are often self-effacing, quiet, reserved, and even shy. Every good-to-great transition in our research began with a Level 5 leader who motivated the enterprise more with inspired standards than inspiring personality.”

There’s no doubt, Starbucks CEO Howard Schultz possesses the characteristics and personality traits of a Level 5 leader. The ambition alone to introduce a new cultural concept in a new market sounds incredibly daunting, but to play it right with the shareholders, customers and their own people sounds impossible.

But that was the initial idea, a moral standard. The mission statement of Starbucks is:

“to inspire and nurture the human spirit – one person, one cup and one neighborhood at a time.”

Let’s break this down into two pieces.

Inspire and nurture the human spirit .

The people, customers, and partners (staff) are the most important assets of any company. The first part of the mission statement explains that in a split-second. The relationships within the company have to be nurtured and supported while exuding warmth and friendliness.

Howard Schultz has shown respect for the mission by developing programs for their own people, which include free education, health insurance and even a share in the Starbucks company.

“One person, one cup and one neighborhood at a time .”

The second part stresses the importance of gradual improvement. Each interaction with a customer, each cup of coffee made hold a large amount of responsibility to deliver the right experience. The neighborhood part reminds the staff and the customers that the stores pay special respect and attention to the place where they are located.

In the article Inside Starbucks’s $35 Million Mission , author Sarah Kessler describes how Starbucks runs the “ Leadership Lab ” — part leadership, part training conference for 10,000 store managers.

Disciplined thought

Face the Reality — When stuff gets hard, leaders don’t turn away from the problem or worse, get busy with mundane tasks, deceiving themselves they are working. Closing your eyes to the reality means you’re on a great way to a downward spiral.

In 2008, Howard Schultz got reinstated by the board as CEO. The sales and shares were dropping. The brand and the culture of Starbucks were deteriorating rapidly. The magical experience was a shadow of its former self.

Schultz decided on a radical idea to close all the stores and retrain in order to inflict the importance of the Starbucks vision and mission. Tied into this transition was closing numerous shops and letting go of hundreds of employees. The ordeal cost the company 6 to 7 million dollars .

In 2018, Starbucks closed the doors again in order to put the staff through racial anti-bias training. The temporary closure cost the company between $15 - $20 million dollars

But it was necessary and long needed. The company picked up from the bottom just like in Drake’s song and has been rapidly growing in the world’s map as well as on index stock charts.

The Hedgehog Concept

The term Hedgehog concept introduced by Collins is some sort of a marriage consisting of a Venn diagram and three major ideas. Jim Collins thinks that in order to have a chance to be the best in the world you have to possess all three:

• The Elite Skill - You will have to be the best in your area of expertise. Constant learning, innovating and moving the boundaries are expected from the movers and shakers of the world.
• Deep Passion - Someone who grows a business will eventually (and continuously) encounter major obstacles where the skill isn’t going to be enough. The grit, powered with a deep passion and a reason why is arguably even more important than the knowledge alone.
• Ability to generate revenue - Understanding of what drives the economic engine is the third piece of the puzzle that completes the concept. No business can survive without sustaining itself and its people financially.

Schultz possesses all three: the Stanford education armed him to become shrewd and dangerous in the business world with a deep understanding of the economic machine while he stayed in love with the company and continued to deeply care for its people and the customers.

The second part of the hedgehog concept is the sheer simplicity of your objective. When it comes to specializing and becoming the best in the world, you need one clear statement which completely prevails over all the others.

The hedgehog is the exact opposite of the fox concept. Foxes are cunning, smart and resourceful animals who take any opportunity to get ahead using any tactic they can think off. Yet when they encounter and attack the hedgehog, the hedgehog simply rolls up into a ball and protects itself with its spiky hide.

The hedgehog companies have one major driving goal that is ingrained as the cornerstone of its business. In Starbucks, it’s not the coffee quality, but it’s the deep desire to create an experience for their customers. Everything is tied into this.

Sometimes, achieving massive rapid growth for the growth sake reveals cracks in the system if it’s not solid. In 2008, when the company was on the decline, Schulz looked at the strategy of the past few years and, in a letter penned company-wide, explained that Starbucks had “invested in infrastructure ahead of the growth curve” and it was time to “shift our emphasis back onto customer-facing initiatives.”

Imagine, the Starbucks insane growth pace required to hire 1,500 new employees a week.

Disciplined action

The success of anything in our lives is in the hands of people. It always is the #1 element in any company.

“In determining the right people, the good-to-great companies placed greater weight on character attributes than on specific educational background, practical skills, specialized knowledge, or work experience.”

When the quality of the work started slipping. Schultz had to close down hundreds of shops for a training day. It was a necessary decision to refocus, restructure and boost Starbucks employees to work and deliver on the right things and to deliver the experience as it was intended in the first place.

When faced with a difficult customer or a problem, the Starbucks partners (employees) are taught customer service by using a L.A.T.T.E. system. The acronym helps baristas deal with any situation in the store.

• L isten to the customer
• A cknowledge the problem/situation
• T ake actions and solve the problem
• T hank the customer
• E xplain what you did

The simple system isn’t there just to provide clear guidelines but it also boosts motivation and willpower among employers. In the book, The Power of Habit , Charles Duhigg wrote that the LATTE system prevented the customer service meltdown , and sustained willpower throughout the day.

In the end, customer service is there to deliver and exceed the experience which is tied to the brand. Nothing is as important as delivering the service. 

“[Employees] are the true ambassadors of our brand, the real merchants of romance and theater, and as such the primary catalysts for delighting customers. Give them reasons to believe in their work and that they’re part of a larger mission, the theory goes, and they’ll in turn personally elevate the experience for each customer–something you can hardly accomplish with a billboard or a 30-second spot.” — Excerpt from book Onward, Howard Schultz

Technology Accelerators

For a globally recognizable brand like Starbucks technology plays a major role in the expansion. The Starbucks app and the emails alone played a significant role in the company’s growth.

According to Collins, technology accelerators have to be carefully selected. Companies had to sift through the emerging technology, identify and select the right ones and gradually introduce them in the business model.

The Hedgehog Concept would drive the use of technology, not the other way around — Jim Collins

Companies that jumped the gun burned badly.

In fact, Jim Collins discovered that more than 80% of great companies didn't rank technology as one of the top five ranking factors for success.

“Those that stay true to these fundamentals and maintain their balance, even in times of great change and disruption, will accumulate the momentum that creates breakthrough momentum. — Jim Collins

Down to the core, Starbucks has one secret ingredient to thank for — knowing their customers. Data analytics. According to Starbucks, this function uses “ methodologies ranging from ethnography to big data analytics … that help support Starbucks pricing strategy, real estate development planning, product development, trade promotion optimization, and marketing strategy.”

Starbucks contracts with a location-analytics company called Esri to use its technology platform that helps analyze maps and retail locations. It uses data like population density, average incomes, and traffic patterns to identify target areas for a new store.

The Crawl, Walk, Run Concept

The gradual introduction of technology is a part of the hedgehog concept. Technology is a major proponent of business growth however if it doesn’t tie into the one simple concept , the company has to be disciplined enough to say no to new opportunities.

Eventually, they can adapt the technology in their concept which turns the massive flywheel forward.

In Starbucks sense, they seem like they embrace technology. They started out with gift cards and pay-ahead mobile purchases. The next step was adding the Starbucks Rewards program to cultivate upsells and raise the LTV per customer. And today with big data, AI, and predictable algorithms they maximize the relationship with the customers.

Key takeaway #8 — the flywheel concept

Successful companies that persevered and thrived with time have found and adopted the Flywheel concept. Focusing on the essentials of the business, working with the right people in the right places, and maintaining discipline is the only way for continued sustainable growth.

Starbucks Vs the World

Competitors.

Starbucks enjoyed the blue ocean marketplace as a premium coffee culture experience provider. 

But as soon as competitors noticed Starbucks discovering a new opportunity they had to react quickly. McDonald's and Dunkin’ Donuts were the big ones that introduced their own versions of coffee-to-go. Better than instant coffee and convenient while on the go, the two competitors did enjoy new revenue stream of introducing coffee; however, as companies, they had to keep the focus on what they are good at — McDonald's with their fast food burgers and fries and Dunkin’ Donuts with well… donuts. DD does serve coffee but had no intention to put more emphasis on it until the late 1990s .

Starbucks kept the lead in the coffee concept because of its focus on the coffee culture and holistic concept of their brand, especially customer service. This point can be seen as soon as you look at international markets. Dunkin' Donuts’ international revenue in 2018 contributed less than 4% of total sales, while roughly 30% of Starbucks' consolidated net revenues in the same period were attributed to markets outside America.

When international expansion goes right

When you get it right and you know you have the brand, processes, and culture down, you can move outside. When Starbucks expanded its adopted “Coffee culture” to new markets it could follow its own tracks again. In many countries, especially Asian nations the idea of a coffee culture was new, fresh, and exciting.

To overcome the culture gap, Starbucks sought partnership through direct investments and joint ventures instead of direct franchising . This solved two major problems.

First, they relied on local retailers who already had experience and experience in the local markets. They married the coffee culture idea with market research of the new areas to discover regional customers’ tastes and preferences. After that, they just had to deliver the employee training, workflows, and the product itself.

Secondly, they acquired and absorbed the entire pieces of coffee markets , such as Coffee Partners in Thailand and Bonstar in Singapore. All in one big swoop.

But even today a Starbucks café is opened every 15th hour in China. It already operates more than 3,000 stores in China and plans to add 2,000 more by 2021 . Seoul has the most Starbucks cafes in any city ( 284 ).

Starbucks is present in 6 continents and in more than 72 countries and territories. But it wasn’t always smooth sailing for the old Starbuck.

And when it doesn’t go so well

While Starbucks had amazing success in Asian countries, they hit a snag in Australia.

In 2008, they closed two-thirds of all stores.

The reason?

Australia is already known as one of the hardest markets to get into in the first place and they are very proud of their coffee culture. The flat whites, coffee art in ceramic lattes have been served for dozens of years at beloved local cafes and by baristas who knew what they are doing.

What Starbucks was doing in the United States was introducing the coffee culture in the new market because it was non-existent before. But in Australia, this model didn’t fit in at all.

In 2008, Starbucks closed two-thirds of all the stores. The prices of Starbucks’ relatively common-tasting coffee (compared to established coffee shops) were pricier than the local solutions and managed by young students who didn’t have the level of appreciation of either the coffee culture and/or Starbucks as a brand.

Key takeaway #9 — establish yourself

Follow the winning formula of developing the markets first and turning into a product innovator after you have established yourself. Forcing the innovation where it’s not perceived as such, is waging a losing battle.

Starbucks on Social Media

The website is simply designed with an intention to present the latest seasonal product in the Starbucks shops in the first fold. The focus of the homepage is also on advertising the Starbucks Rewards program.

According to SimilarWeb, it attracts 18.9M visits per month, with an average of 2 minutes and 3.2 page views per session. Starbucks site is the 9th top ranked site for Food and Drink category in the world

The Youtube channel was established at the end of 2005. After 16 years it managed to acquire 335,000 subscribers, which isn't’ that much if we take the size of the company into consideration.

The most successful videos are close to 10 million views; however, they are short, 15-seconds clips of the product. The channel moderators are not participating in the comment sections.

Luckily there’s not much competition on YouTube; however, as a highly visual channel, Starbucks could advertise their mobile app and Starbucks reward program using socially-conscious values, product innovation, or sustainability programs.

On the other hand, Instagram is doing absolutely amazing. Naturally, since the best Starbucks customers are the ones who have been using their mobile devices for ordering and participating in the Starbucks Rewards program

Starbucks Instagram uses a mix of images and video clips mostly displaying their well-designed cups. The posts are mostly re-shared (“regrams”) of other Instagram users. With this tactic, Starbucks incentivizes UGC (user-generated content), since Instagram users have the chance to be regrammed and have their Starbucks shot seen by 17.8 million followers.

Pinterest is another great visual platform where images are split into different categories: from coffee recipes, coffee photography to store designs and world-recognized Starbucks cups.

Pinterest receives 10+ million monthly views and has 443,600 followers.

Even though their daily support is dropping, Facebook is still being used as one of the channels where Starbucks shows its videos and posts.

On Twitter , Starbucks shares its globally conscious ideas, news, and stories about the company and its products. Twitter also serves as a chance to (as in Instagram) retweet other users’ posts.

Starbucks likes to reshare the positive messages of happy users who had a positive experience at one of their stores

Since Starbucks' success mainly lies in their visual branding, they use social media for their brand awareness and in a Facebook sense, pushing the mobile app downloads.

Key takeaway #10 — delegate your resources

When using social media, identify which social media platform brings the best results. If your users are primarily on mobile devices, Instagram would be a smart choice. Delegate your resources to the best-performing channel.

Starbucks Corp. has become a worldwide success by sticking to its hedgehog concept. The realization of being customer-centric in the practical, not just theoretical sense laid the foundation of expansion in North American markets as well as international ones.

When all of the decisions are catered to the concept of serving their customers, including using technology as accelerators, there’s nothing to worry about in their future.

Starbucks Case Study - How Starbucks Conquered The Coffee Industry?

Devashish Shrivastava

Starbucks Corporation is an American coffee chain that was established in 1971 in Seattle, Washington. By mid-2019, the organization had a presence in over 30,000 areas around the world. Starbucks has been depicted as the fundamental delegate of "second wave espresso," a reflectively-named development that advanced high-quality espresso and specially simmered coffee. Starbucks now uses robotized coffee machines for proficiency and well-being.

Starbucks serves hot and cold beverages, entire bean espresso, micro-ground moment espresso known as VIA, coffee, caffe latte, full-and free leaf teas such as Teavana tea products, Evolution Fresh squeezes, Frappuccino refreshments, La Boulange baked goods, and bites (for example, chips and wafers); some offerings such as the Pumpkin Spice Latte are explicit to the territory of the store. Numerous Starbucks outlets sell pre-bundled nourishment items, sweltering and cold sandwiches, and drinkware such as cups and tumblers. Furthermore, there are Select "Starbucks Evenings" areas that offer brew, wine, and appetizers.

Starbucks first ended up productive in Seattle in the mid-1980s. Despite an underlying financial downturn with its venture into the Midwest and British Columbia in the late 1980s, the organization experienced rejuvenated success with its entrance into California in the mid-1990s. Starbucks opened an average of two new stores every day between 1987 and 2007. On December 1, 2016, Howard Schultz reported he would leave his position as the CEO and would be supplanted by Kevin Johnson. Johnson accepted the role of the CEO of Starbucks on April 3, 2017, and Howard Schultz resigned to end up as the 'Chairman Emeritus', effective from June 26, 2018. Kevin Johnson is currently serving as the CEO and President of Starbucks.

Starbucks - Company Highlights

Startup Story Of Starbucks Corporation History Of Starbucks Corporation Starbucks - Name and Logo Starbucks Expansion Journey Starbucks Corporation in India Business Strategy Of Starbucks In India Products Of Starbucks Corporation Business Growth Of Starbucks Corporation Over The Years Future Plans Of Starbucks Corporation

Startup Story Of Starbucks Corporation

If you are wondering how did Starbucks start? Then, the story of Starbucks started back in 1971, when the company was a roaster and retailer of whole bean and ground coffee, tea and spices with a single store in Seattle’s Pike Place Market.

Zev Siegel stated that at that time he knew the coffee industry inside and out, he was well-versed, especially with the gourmet end of the industry. Besides, he was also known as the most educated coffee guy in the country at that time. So, the three college friends - Zev Siegel, Jerry Baldwin and Gordon, started out with their coffee bean shop and roastery at Seattle’s famous Pike Place Market in 1971. Eventually, they found a mentor in Alfred Peet, who was the founder of Peet’s Coffee and the man responsible for bringing custom coffee roasting to the U.S. and started with the coffee business in full swing. Starbucks initially began by selling coffee beans that were roasted by Peet's, a gourmet coffee company in Berkeley, California, and later on, started roasting on their own.

History Of Starbucks Corporation

The first Starbucks store was initiated in 1971 in Washington by 3 individuals who met while they were studying at the University of San Francisco: English educator Hun Baldwin, history educator Zev Siegl, and author Gordon Bowker. The trio was encouraged to sell top-notch espresso beans and hardware after businessman Alfred Peet showed them his style of simmering beans.

During this time, the organization sold simmered, entire espresso beans. During its first year of activity, Starbucks bought green espresso beans from Peet's, and then started purchasing legitimately from producers.

Starbucks - Name and Logo

Bowker reviews that Terry Heckler, with whom Bowker claimed a publicizing office, thought words starting with "st" were ground-breaking. The organizers conceptualized a rundown of words starting with "st" and in the long run arrived on "Strabo," a mining town in the Cascade Range. The team then finalized on "Starbuck," the name of the young chief mate in the book "Moby-Dick".

Starbucks has given too many slogans/taglines already among which the most popular one is - " Brewed for those who love coffee".

Starbucks Expansion Journey

In 1984, the first proprietors of Starbucks, driven by Jerry Baldwin, acquired Peet's. During the 1980s, all-out offers of espresso in the US were falling. However, offers of strength espresso expanded, shaping 10% of the market in 1989; it stood at just 3% in terms of market share in 1983. By 1986, the organization worked six stores in Seattle and had just barely started to sell coffee.

In 1987, the first proprietors sold the Starbucks chain to the previous manager Howard Schultz, who rebranded his II Giornale espresso outlets as Starbucks and immediately extended. Starbucks then launched its outlets outside Seattle at Waterfront Station in Vancouver, British Columbia, and Chicago, Illinois. By 1989, 46 stores existed over the Northwest and Midwest, and every year Starbucks was simmering more than 2,000,000 pounds (907,185 kg) of coffee. At the hour of its first sale of stock (IPO) on the financial exchange in June 1992, Starbucks had 140 outlets with an income of $73.5 million, up from $1.3 million in 1987.

The organization's fairly estimated worth was $271 million at this point. The 12% segment of the organization that was sold raised around $25 million for the organization, which encouraged a multiplying of the number of stores throughout the following two years. By September 1992, Starbucks' offer cost had ascended by 70% to more than multiple times the income per portion of the past year. In July 2013, over 10% of in-store buys were made on the client's cell phones utilizing the Starbucks app.

The organization used the versatile social media stage when it propelled the "Tweet-a-Coffee" campaign in October 2013. People had the option to buy a $5 gift voucher for a companion by entering both "@tweetacoffee" and the companion's handle in a tweet. Research firm Keyhole observed the advancement of the event and a media article from December 2013 detailed that Starbucks had discovered that 27,000 individuals had taken an interest and $180,000 of buys were made to date.

As of 2018, Starbucks is positioned 132nd on the Fortune 500 rundown of the biggest United States organizations by revenue. In July 2019, Starbucks announced a "monetary second from last quarter total compensation of $1.37 billion, or $1.12 per share, up from $852.5 million, or 61 pennies for each offer, a year sooner." The organization's fairly estimated worth of $110.2 billion expanded by 41% in the middle of 2019. The income per share in quarter three was recorded at 78 pennies, considerably more than the estimate of 72 cents.

Starbucks Corporation in India

In January 2011, Starbucks Corporation and Tata Coffee reported designs to start opening Starbucks outlets in India. Despite a bogus beginning in 2007, in January 2012, Starbucks declared a 50:50 joint endeavour with Tata Global Beverages, called Tata Starbucks Ltd. , which would possess and work outlets marked "Starbucks, A Tata Alliance". Starbucks had endeavoured to enter the Indian market in 2007. However, it didn't provide any explanation behind its withdrawal of it.

It was on October 19, 2012 that Starbucks opened its first store, a 4,500 sq ft store in Elphinstone Building, Horniman Circle, Mumbai. Starbucks opened its first cooking and bundling plant in Coorg, Karnataka in 2013 to supply its Indian outlets. The company extended its reach to Delhi on 24 January 2013 by opening 2 outlets. Tata Global Beverages declared in 2013 that they would have 50 areas before the end of the year, with a venture of ₹4 billion ($58 million). The organization did open its 50th store in India on July 8, 2014.

The third city of India to get a Starbucks outlet was Pune, where the organization opened an outlet at Koregaon Park on 8 September 2013. Starbucks opened a 3,000-square-foot lead store at Koramangala, Bangalore on 22 November 2013, making it the fourth city to have an outlet. Starbucks opened the biggest espresso-forward store in the nation at Vittal Mallya Road, Bangalore on 18 March 2019. The store is estimated at 3,000 sq ft and is Starbucks' 140th outlet in India.

Tata Starbucks opened 25 stores between 2017 and 2018, which went up to 30 during 2018-19. On 21 February 2019, CEO Navin Gurnaney reported that Tata Starbucks would use only compostable and recyclable bundling materials over the entirety of its stores from June 2020.

Starbucks reported its entrance in Gujarat on 7 August 2019. The organization opened five stores in Surat and Ahmedabad the following day. Starbucks' leader store in the state is situated at Prahlad Nagar, Ahmedabad, and offers more vegan alternatives than other Indian outlets. CEO Navin Gurnaney expressed that the organization would open more than 30 stores in the 2019-20 financial year, of which 11 have already been opened.

Business Strategy Of Starbucks In India

Starbucks' strategies for business in India seemed rock-solid but the brand wasn't completely immune still. In any case, the world's biggest bistro chain is building its position cautiously via a progression of well-picked steps. Numerous worldwide brands have entered India since the 1990s, being pulled in by its developing and optimistic customer base. Yet, not all have succeeded.

Starbucks isn't the primary contestant in India's composed espresso showcase; so it doesn't have any first-participant advantage. Cafe Coffee Day (CCD) is the market head while Barista Lavazza was the main espresso chain to open for business. Both are valued by the white-collar class. Costa Coffee, Coffee Bean and Tea Leaf (CBTL), and Gloria Jean are valued by the rich group in India.

India is customarily a tea-drinking nation, so espresso chains have concentrated on giving a feel where individuals can unwind and invest energy with one another. This setup implies higher capital expenses. It is different from the US, where the vast majority have a liking for espresso. The Indian buyer base has likewise advanced in the recent decade. What can worldwide brands like Starbucks do to augment their odds of achievement in India? Here are a few thoughts:

Picking a Local Partner

Worldwide brands face the difficult choice of either going solo or tying up with a nearby accomplice. Starbucks' choice to team up with India's TATA Global Beverages demonstrates attention to utilizing different advantages. The TATA Group is one of India's morally determined brands, an observation passed on about Starbucks India too.

Given that India produces espresso beans in just a couple of spots, the other sourcing alternative was bringing in the beans. Be that as it may, this would have raised costs fundamentally.

Tata's espresso plant in Karnataka has been contracted to supply beans to Starbucks' universally, making common cooperative energies. It has contracted to take into account TATA's TAJ SATS, which supplies to TATA's top-notch lodging network – The TAJ. The TATAs are put into the retail part with store brands like Westside, Tanishq, Croma, Star Bazaar, and so forth. Starbucks can use them for information sharing on Indian land, territory points of interest, and handling land administrations. This would enable its very own development to outline. This strategy gives scope for store-in-store deals.

Consistency in Store Arrangements

This keeps up the one-of-a-kind selling purpose of customer experience and allows to pick up economies of scale on CAPEX. Starbucks plans to have a similar store group crosswise over India. However, the size can change depending on financial matters. This is how it works all around. Starbucks wants to provide an agreeable 'café' experience. Having a similar organization gives clients the solace of accepting the equivalent 'Starbucks' vibe any place they go throughout the world.

Keeping the store designs steady means it needs to pick and open new areas stringently, to such an extent that the area can yield a throughput by the venture. Its methodology in-store arrangement is different from CCD, which has picked various configurations to tap the potential interest in any region. CCD has opened a couple of premium outlets dependent on the area's customer profile . It has additionally gone for non-store organizations like takeaway booths and candy machines. Be that as it may, Starbucks may expect that such non-store configurations may weaken its image esteem.

Estimating the Pace of Expansion

India is the place where an inability to screen primary concerns has tossed numerous organizations out of the rigging. So, a top-line just approach doesn't work here. Since Starbucks needs to pick new areas stringently by its equivalent configuration approach, it has decided on a deliberate pace of extension. It is concentrating on the budgetary feasibility of every outlet, as opposed to going for an aggressive development plan which may have brought about rehashed calls for capital.

This operational process is different from its system in the USA and China where it has fabricated scale by opening stores in pretty much every area – being the main port-of-call for espresso by basically being all over the place. CCD's methodology behind adaptable store organizations was to guarantee there is a CCD bistro at a simple reach. It is intriguing to check its normal store gainfulness given its scale.

Guaranteeing Top-Authority Backing and Responsibility

Top initiative responsibility from the two sides of the organization, Tata and Starbucks, has been plentifully clear. Starbucks took as much time as is needed to enter the market (6 years), recognizing that India was a mind-boggling market and required cautious passage arranging. The two sides have spoken finally about their dedication and shared their future plans to give their business a new direction toward growth.

Altering Contributions to Suit Indian Market and Client Needs

Being adjusted to Indian culture, tastes, and inclinations conveyed at a suitable "esteem" guarantees customer importance, construct, and continued utilization. Starbucks mirrors this comprehension – as observed through a blend of western staples, a wide scope of intriguing Indian tidbits similar to confined refreshments on the idea. Since its experience ( and item as well, however to a lesser degree) is its image guarantee, its test lies in conveying an all-around steady, yet locally significant brand experience.

The stores, or the "third spot" as Starbucks calls them, have been altered likewise. The stores don't pursue the worldwide layout and appear to have been planned with consideration, with neighbourhood contacts consolidated. Stores in various urban communities have been structured unexpectedly, mirroring the neighbourhood culture – for e.g., New Delhi's store has ropes and chat on the dividers and henna designs on the floor, though the Pune store has a rich showcase of collectables and copper.

There appears to be sufficient utilization of shading – something missing in the US. The stores have been intended to convey a particular, premium café experience, predictable, and in a state of harmony with the one conveyed over the rest of the world.

Making Inventive and Restricted Plan of Action

Starbucks appears to have made a confined plan of action, planned for conveying a universally reliable item and involvement with locally-focused costs. The Tata group conveys a major sourcing advantage (attributable to its quality over the generation chain, developing, broiling, and exchanging espresso), yet it has just gone past that to develop and support associations with nearby espresso cultivators – putting resources into structure economical cultivating rehearses. All of Starbucks' espresso is sourced locally, a first-ever for the organization.

Scaling up using Arrangements and Organizations

The Tata organization is the genuine overthrow in the Starbucks passage story. Having Tata as an accomplice is gigantically profitable, not due to the validity and strength it offers, or because it coordinates the scale and stature of Starbucks as an organization.

It offers numerous advantages catalyzing pretty much every market section achievement variable - for example, The Tata group has involvement in the retail business , a solid reputation in advancing new pursuits, gives a sourcing advantage through Tata espresso, offers access to high traffic areas using its lodgings and other retail outlets, guarantee excellent nourishment and refreshment supply through its F&B business and so forth.

Furthermore, the potential for an effective organization is amazingly high given Starbucks' and Tata's mutual qualities – the two of them have a solid social inner voice and are resolved to "give back" to the general public and network.

Influencing India for Worldwide Items

Not long after it finished its first year, Starbucks reported that it was serving top-quality Indian Arabica espresso as "Indian coffee" in different markets. Another world-class office for cooking and bundling has just been initiated in Coorg, Karnataka; the results of which are to be analyzed in India and abroad.

Overseeing Discernment and Guidelines

This viewpoint is tied in with structure, a solid positive observation and a picture for the business and brand crosswise over key outer partners and crowds – incorporating the administration, corporate accomplices, networks inside the eco-framework, and customers on the loose. Given what Starbucks has figured out how to accomplish in a year and a half since dispatch, it appears to be genuinely evident that its thought combined with the Tata advantage (critical reach and impact) has helped in developing solid connections and a positive picture with key outside partners and voting demographics.

Engage Nearby Association

Starbucks is by all accounts constructing a nation-explicit activity with nearby individuals in charge and overall unmistakable customer interface focuses, giving them the necessary position to coordinate and work. There is overwhelming interest in enlisting the perfect individuals and giving the essential preparation – to install and instil the organization's culture and administration models.

Along these lines, how has Starbucks fared against the McKinsey spread out variables for long-haul India achievement? Its accomplishments against the scorecard look noteworthy. With thorough vigorous passage arranging and brilliant and quick execution, the multi-month-old endeavour appears to have impressive force, making purchaser and network-driven ventures and focused on sustaining its centre business and brand. It appears to be very much set to "win" in India.

Whether Starbucks will collect a huge piece of the overall industry and accomplish its objective of India being among its best 5 markets over the long haul is not yet clear. It's still early days, yet for the organization, this appears to be an incredible beginning and a great globalization model for multinationals looking for an India section.

Products Of Starbucks Corporation

Aside from the typical items offered globally, Starbucks in India has some Indian-style item contributions, for example, Tandoori Paneer Roll, Chocolate Rossomalai Mousse, Malai Chom Tiramisu, Elaichi Mewa Croissant, Chicken Kathi Roll, and Murg Tikka Panini to suit Indian customers. All coffees sold in Indian outlets are produced using Indian broiled espressos by Tata Coffee. Starbucks additionally sells Himalayan packaged mineral water. Free Wi-Fi is accessible at all Starbucks stores.

In January 2017, Tata Starbucks presented Starbucks' tea image "Teavana". Teavana offers 18 unique assortments of tea in India. One of the assortments called the India Spice Majesty Blend was explicitly created for the Indian market and is just accessible in India. India Spice Majesty Blend is a mix of full leaf Assam dark tea injected with entire cinnamon, cardamom, cloves, pepper, star anise, and ginger. On 15 June 2015, Tata Starbucks reported that it was suspending the utilization of fixings that had not been affirmed by the Food Safety and Standards Authority of India (FSSAI).

The organization didn't indicate what the fixings were or which items they were utilized in. The organization additionally expressed that it was applying for FSSAI endorsement for these ingredients.

As per the Latte Index positioning of the expense of a tall hot latte at Starbucks in 44 nations, India was the fifth most costly nation to buy the drink dependent on January 2016 costs. The record distributed by US-based buyer research firm ValuePenguin found that a tall hot latte cost $7.99 in India, far higher than the $2.75 it costs in the least expensive nation, the United States, yet much lower than the $12.32 in the most costly nation, Russia .

Tata Starbucks propelled the Starbucks Delivers program in mid-2019. The administration offers home conveyance from Starbucks outlets through an organization with Swiggy. The administration was first propelled in Mumbai, with designs to turn it out to other cities.

In its menu, the Tata Starbucks company has launched ice-creams as their new products. The frozen delights are available even in flavours like java chip and caramel macchiato among others and will come in takeaway tubs and single scoops. The ice-creams are now available in 50-60% of the Starbucks stores.

Business Growth Of Starbucks Corporation Over The Years

Tata Starbucks, a 50:50 joint endeavour between Tata Global Beverages and Starbucks Coffee of the US, has announced a 30%  top-line development in financial 2018-19, driven by new store openings and improved execution. Tata Starbucks, which is hoping to make back the initial investment in the current money, has opened 146 stores to date. Tata Starbucks announced "twofold digit top-line development - 30% for the entire year, driven by new stores and improved store execution," Tata Global Beverages Ltd (TGBL) said in a financial specialists' introduction. Tata Starbuck's income for 2018-19 is required to be approximately INR 450 crores.

TGBL said Tata Starbucks opened 30 outlets in the past financial year, out of which 15 new stores were opened during the last quarter of the money-related year. The organization claimed detailed benefits at the store level; all urban areas were likewise productive, and additionally saw an ascend in nourishment share in general deals.

The Starbucks company has added around 40 stores in FY21 but the company had recorded a 33% Y-O-Y  fall in its revenues during the same fiscal. According to the Sushant Dash, CEO of Tata Starbucks, the recovery that the company has seen after the second wave of COVID-19 was better than what it saw after the first wave of the deadly pandemic. The quarterly growth after Q2 FY22 was 120% more than what it saw during the same period in the previous fiscal. The company has hugely focused on home deliveries ever since the pandemic broke out. It has already addressed concerns associated with the spillage and other challenges pertaining to home delivery, which contributed to over 18% of the total sales that the company witnessed this fiscal, as per the reports in November 2021. Furthermore, the company has also added ice-creams to their menu in flavours like java chip and caramel macchiato. The Sanjeev Kapoor menu is another thing that has been freshly launched by Tata Starbucks. Besides, the company also launched a one-litre freshly brewed beverage and at-home coffee.  

Future Plans Of Starbucks Corporation

Tata Starbucks Pvt. Ltd. is looking to forcefully grow its impression in the Indian market with its eyes on the quickly spreading "espresso culture" among the twenty to thirty-year-olds and upwardly versatile customers. Tata Starbucks, a JV between US-based Starbucks Coffee Company and Tata Global Beverages Ltd, hopes to set up altogether more number stores this monetary than it did previously.

Starbucks is hopeful about solid business development in India throughout the following year as it means to leave red in monetary numbers after 2020. "Our proceeded with development in topline and reasonable methodology towards extension will enable us to accomplish make back the initial investment by March 2020," Navin Gurnaney, CEO, Tata Starbucks disclosed to Business Line in the wake of declaring five new stores in Gujarat - three in Ahmedabad and two in Surat. Gurney likewise included, "First time in quite a while, we are opening five stores in any state in one go.

Gujarat is a significant market for us. In the wake of opening these five stores on Thursday, the all outnumber of hides away goes up to 157 in India." Starbucks entered India with its first store opened at Mumbai in 2012. Of the 157, the organization has opened all out 11 stores so far in this financial, as against complete 30 stores opened during 2018-19. It takes into account 270,000 clients each week in India. The organization had announced a turnover of INR 442 crores for the monetary 2018-19.

"Espresso business in India is developing significantly. The espresso culture is being initiated by recent college grads, upwardly versatile, and individuals who travel and get brand. Two years back, we set up 25 stores (in a year). During the last financial 2018-19, we included 30 stores.

This year we will beat that number considerably and by end of March 2020, we will have included a lot a greater number of stores than we included in the past," Gurney said. With per store venture prerequisites being evaluated at INR 1.7-2 crores, the complete CAPEX plan by the organization works out in overabundance of INR 50 crores during current monetary on the off chance that it opens more number of stores than a year ago. Be that as it may, Gurnaney ceased from giving venture figures for 2019-20.

The organization is likewise open to different open doors for development including inorganic development through acquisitions. Be that as it may, when tested about any probability of a venture plan in the espresso chain Cafe Coffe Day (CCD), Gurnaney denied estimating any discussions for securing. "We are very hopeful about India. We will be attentively forceful (to extend). (At present) we are not in discussions with anyone for obtaining.

In any case, we are hoping to develop constantly," he included. With an end goal to upgrade the client experience, Starbucks is presenting new nourishment things, taking into account all client needs including breakfast and lunch. The income share from nourishment things is right now around 25%, even as it keeps on developing with new things to meet the client's needs.

Who founded Starbucks?

Starbucks was started by Hun Baldwin, Zev Siegl, and Gordon Bowker in 1971.

Where was the first Starbucks started?

Starbucks was started in Pike Place Market, Seattle, Washington, United States.

When was Starbucks started in India?

Starbucks was launched in India in 2012.

What is the revenue of Starbucks?

Starbucks revenue was recorded $29.02 billion in 2021.

How many Starbucks stores are there worldwide?

There are 33,830 Starbucks stores in the world as of 2021.

Must have tools for startups - Recommended by StartupTalky

• Convert Visitors into Leads- SeizeLead
• Payment Gateway- Razorpay
• Spy on your Competitors- Adspyder
• Manage your business smoothly- Google Workspace

Top 15 Best Money-Earning Mobile Apps That Actually Work

In this world of constant updating, where everything has a cost, money has become an essential part of survival. In the past, people had part-time jobs because they had to be deserving, but today, mobile phones are a means of earning large sums of money without putting in a lot

Varco Leg Care Secures $0.5 Million Seed Funding Led by Sunicon Maiden Fund

* The funds will be used to grow R&D, marketing and offline presence. * The company’s valuation has grown by 2x in the past 10 months. 16th April 2024, National: New Delhi-based Varco Leg Care, the world’s first dedicated leg care brand, has raised a cumulative of US$ 0.

Surge in the E-commerce Industry: WorkIndia Provides In-Depth Insights on its Market Dynamics

National, 16 th April 2024: The e-commerce sector, a driving force in modern commerce, continues to demonstrate resilience and growth. WorkIndia, a leading tech-driven blue-collar recruitment platform reveals key insights into its intricate dynamics across the key aspects of this sector's composition and growth trajectory. In terms of geographic concentration,

Internshala Jobs Unveils New Report: Navigating GenZ Workplace Expectations in 2024: Insights for HR Professionals

67% GenZ job seekers prioritise learning and growth opportunities over salary packages; Genz is highly optimistic in finding Internshala Jobs, the jobs arm of career-tech platform, Internshala, has released a pioneering report titled "Navigating GenZ Workplace Expectations in 2024: Insights for HR Professionals". This report provides an in-depth exploration of

• Search Menu
• Browse content in A - General Economics and Teaching
• Browse content in A1 - General Economics
• A13 - Relation of Economics to Social Values
• Browse content in B - History of Economic Thought, Methodology, and Heterodox Approaches
• B2 - History of Economic Thought since 1925
• B4 - Economic Methodology
• Browse content in B5 - Current Heterodox Approaches
• B52 - Institutional; Evolutionary
• B59 - Other
• Browse content in C - Mathematical and Quantitative Methods
• Browse content in C0 - General
• C01 - Econometrics
• Browse content in C1 - Econometric and Statistical Methods and Methodology: General
• C11 - Bayesian Analysis: General
• C12 - Hypothesis Testing: General
• C13 - Estimation: General
• C14 - Semiparametric and Nonparametric Methods: General
• C15 - Statistical Simulation Methods: General
• Browse content in C2 - Single Equation Models; Single Variables
• C21 - Cross-Sectional Models; Spatial Models; Treatment Effect Models; Quantile Regressions
• C22 - Time-Series Models; Dynamic Quantile Regressions; Dynamic Treatment Effect Models; Diffusion Processes
• C23 - Panel Data Models; Spatio-temporal Models
• C24 - Truncated and Censored Models; Switching Regression Models; Threshold Regression Models
• C25 - Discrete Regression and Qualitative Choice Models; Discrete Regressors; Proportions; Probabilities
• C26 - Instrumental Variables (IV) Estimation
• Browse content in C3 - Multiple or Simultaneous Equation Models; Multiple Variables
• C30 - General
• C32 - Time-Series Models; Dynamic Quantile Regressions; Dynamic Treatment Effect Models; Diffusion Processes; State Space Models
• C33 - Panel Data Models; Spatio-temporal Models
• C34 - Truncated and Censored Models; Switching Regression Models
• C35 - Discrete Regression and Qualitative Choice Models; Discrete Regressors; Proportions
• C36 - Instrumental Variables (IV) Estimation
• Browse content in C4 - Econometric and Statistical Methods: Special Topics
• C41 - Duration Analysis; Optimal Timing Strategies
• C43 - Index Numbers and Aggregation
• C45 - Neural Networks and Related Topics
• C46 - Specific Distributions; Specific Statistics
• C49 - Other
• Browse content in C5 - Econometric Modeling
• C50 - General
• C52 - Model Evaluation, Validation, and Selection
• C53 - Forecasting and Prediction Methods; Simulation Methods
• C54 - Quantitative Policy Modeling
• C55 - Large Data Sets: Modeling and Analysis
• C57 - Econometrics of Games and Auctions
• C58 - Financial Econometrics
• Browse content in C6 - Mathematical Methods; Programming Models; Mathematical and Simulation Modeling
• C61 - Optimization Techniques; Programming Models; Dynamic Analysis
• C63 - Computational Techniques; Simulation Modeling
• C68 - Computable General Equilibrium Models
• Browse content in C7 - Game Theory and Bargaining Theory
• C72 - Noncooperative Games
• C73 - Stochastic and Dynamic Games; Evolutionary Games; Repeated Games
• Browse content in C8 - Data Collection and Data Estimation Methodology; Computer Programs
• C81 - Methodology for Collecting, Estimating, and Organizing Microeconomic Data; Data Access
• C82 - Methodology for Collecting, Estimating, and Organizing Macroeconomic Data; Data Access
• C83 - Survey Methods; Sampling Methods
• Browse content in C9 - Design of Experiments
• C90 - General
• C91 - Laboratory, Individual Behavior
• C92 - Laboratory, Group Behavior
• C93 - Field Experiments
• C99 - Other
• Browse content in D - Microeconomics
• Browse content in D0 - General
• D01 - Microeconomic Behavior: Underlying Principles
• D02 - Institutions: Design, Formation, Operations, and Impact
• D03 - Behavioral Microeconomics: Underlying Principles
• D04 - Microeconomic Policy: Formulation; Implementation, and Evaluation
• Browse content in D1 - Household Behavior and Family Economics
• D11 - Consumer Economics: Theory
• D12 - Consumer Economics: Empirical Analysis
• D13 - Household Production and Intrahousehold Allocation
• D14 - Household Saving; Personal Finance
• D18 - Consumer Protection
• D19 - Other
• Browse content in D2 - Production and Organizations
• D20 - General
• D21 - Firm Behavior: Theory
• D22 - Firm Behavior: Empirical Analysis
• D23 - Organizational Behavior; Transaction Costs; Property Rights
• D24 - Production; Cost; Capital; Capital, Total Factor, and Multifactor Productivity; Capacity
• D25 - Intertemporal Firm Choice: Investment, Capacity, and Financing
• Browse content in D3 - Distribution
• D30 - General
• D31 - Personal Income, Wealth, and Their Distributions
• Browse content in D4 - Market Structure, Pricing, and Design
• D40 - General
• D43 - Oligopoly and Other Forms of Market Imperfection
• D44 - Auctions
• D45 - Rationing; Licensing
• Browse content in D5 - General Equilibrium and Disequilibrium
• D57 - Input-Output Tables and Analysis
• D58 - Computable and Other Applied General Equilibrium Models
• Browse content in D6 - Welfare Economics
• D60 - General
• D61 - Allocative Efficiency; Cost-Benefit Analysis
• D62 - Externalities
• D63 - Equity, Justice, Inequality, and Other Normative Criteria and Measurement
• D64 - Altruism; Philanthropy
• Browse content in D7 - Analysis of Collective Decision-Making
• D70 - General
• D71 - Social Choice; Clubs; Committees; Associations
• D72 - Political Processes: Rent-seeking, Lobbying, Elections, Legislatures, and Voting Behavior
• D78 - Positive Analysis of Policy Formulation and Implementation
• Browse content in D8 - Information, Knowledge, and Uncertainty
• D80 - General
• D81 - Criteria for Decision-Making under Risk and Uncertainty
• D82 - Asymmetric and Private Information; Mechanism Design
• D83 - Search; Learning; Information and Knowledge; Communication; Belief; Unawareness
• D84 - Expectations; Speculations
• D85 - Network Formation and Analysis: Theory
• D86 - Economics of Contract: Theory
• D89 - Other
• Browse content in D9 - Micro-Based Behavioral Economics
• D90 - General
• D91 - Role and Effects of Psychological, Emotional, Social, and Cognitive Factors on Decision Making
• D92 - Intertemporal Firm Choice, Investment, Capacity, and Financing
• Browse content in E - Macroeconomics and Monetary Economics
• Browse content in E3 - Prices, Business Fluctuations, and Cycles
• E31 - Price Level; Inflation; Deflation
• Browse content in E4 - Money and Interest Rates
• E43 - Interest Rates: Determination, Term Structure, and Effects
• Browse content in F - International Economics
• F0 - General
• Browse content in F1 - Trade
• F10 - General
• F11 - Neoclassical Models of Trade
• F12 - Models of Trade with Imperfect Competition and Scale Economies; Fragmentation
• F13 - Trade Policy; International Trade Organizations
• F14 - Empirical Studies of Trade
• F15 - Economic Integration
• F16 - Trade and Labor Market Interactions
• F17 - Trade Forecasting and Simulation
• F18 - Trade and Environment
• F19 - Other
• Browse content in F2 - International Factor Movements and International Business
• F22 - International Migration
• F23 - Multinational Firms; International Business
• Browse content in F3 - International Finance
• F36 - Financial Aspects of Economic Integration
• Browse content in F4 - Macroeconomic Aspects of International Trade and Finance
• F43 - Economic Growth of Open Economies
• F47 - Forecasting and Simulation: Models and Applications
• Browse content in F5 - International Relations, National Security, and International Political Economy
• F51 - International Conflicts; Negotiations; Sanctions
• F54 - Colonialism; Imperialism; Postcolonialism
• F55 - International Institutional Arrangements
• Browse content in F6 - Economic Impacts of Globalization
• F61 - Microeconomic Impacts
• F62 - Macroeconomic Impacts
• Browse content in G - Financial Economics
• Browse content in G0 - General
• G00 - General
• Browse content in G1 - General Financial Markets
• G12 - Asset Pricing; Trading volume; Bond Interest Rates
• G13 - Contingent Pricing; Futures Pricing
• G14 - Information and Market Efficiency; Event Studies; Insider Trading
• G15 - International Financial Markets
• Browse content in G2 - Financial Institutions and Services
• G22 - Insurance; Insurance Companies; Actuarial Studies
• G23 - Non-bank Financial Institutions; Financial Instruments; Institutional Investors
• Browse content in G3 - Corporate Finance and Governance
• G32 - Financing Policy; Financial Risk and Risk Management; Capital and Ownership Structure; Value of Firms; Goodwill
• G33 - Bankruptcy; Liquidation
• G34 - Mergers; Acquisitions; Restructuring; Corporate Governance
• Browse content in H - Public Economics
• Browse content in H1 - Structure and Scope of Government
• H12 - Crisis Management
• Browse content in H2 - Taxation, Subsidies, and Revenue
• H20 - General
• H22 - Incidence
• H23 - Externalities; Redistributive Effects; Environmental Taxes and Subsidies
• H25 - Business Taxes and Subsidies
• H27 - Other Sources of Revenue
• Browse content in H4 - Publicly Provided Goods
• H40 - General
• H41 - Public Goods
• H43 - Project Evaluation; Social Discount Rate
• Browse content in H5 - National Government Expenditures and Related Policies
• H55 - Social Security and Public Pensions
• Browse content in H8 - Miscellaneous Issues
• H83 - Public Administration; Public Sector Accounting and Audits
• Browse content in I - Health, Education, and Welfare
• Browse content in I1 - Health
• I10 - General
• I12 - Health Behavior
• I18 - Government Policy; Regulation; Public Health
• I19 - Other
• Browse content in I3 - Welfare, Well-Being, and Poverty
• I30 - General
• I31 - General Welfare
• I32 - Measurement and Analysis of Poverty
• I38 - Government Policy; Provision and Effects of Welfare Programs
• Browse content in J - Labor and Demographic Economics
• Browse content in J1 - Demographic Economics
• J10 - General
• Browse content in J2 - Demand and Supply of Labor
• J20 - General
• J21 - Labor Force and Employment, Size, and Structure
• J22 - Time Allocation and Labor Supply
• J24 - Human Capital; Skills; Occupational Choice; Labor Productivity
• Browse content in J4 - Particular Labor Markets
• J43 - Agricultural Labor Markets
• Browse content in J5 - Labor-Management Relations, Trade Unions, and Collective Bargaining
• J54 - Producer Cooperatives; Labor Managed Firms; Employee Ownership
• Browse content in J6 - Mobility, Unemployment, Vacancies, and Immigrant Workers
• J60 - General
• J61 - Geographic Labor Mobility; Immigrant Workers
• J62 - Job, Occupational, and Intergenerational Mobility
• Browse content in K - Law and Economics
• Browse content in K2 - Regulation and Business Law
• K20 - General
• Browse content in K3 - Other Substantive Areas of Law
• K37 - Immigration Law
• Browse content in L - Industrial Organization
• Browse content in L1 - Market Structure, Firm Strategy, and Market Performance
• L10 - General
• L11 - Production, Pricing, and Market Structure; Size Distribution of Firms
• L12 - Monopoly; Monopolization Strategies
• L13 - Oligopoly and Other Imperfect Markets
• L14 - Transactional Relationships; Contracts and Reputation; Networks
• L15 - Information and Product Quality; Standardization and Compatibility
• L19 - Other
• Browse content in L2 - Firm Objectives, Organization, and Behavior
• L21 - Business Objectives of the Firm
• L22 - Firm Organization and Market Structure
• L23 - Organization of Production
• L24 - Contracting Out; Joint Ventures; Technology Licensing
• L25 - Firm Performance: Size, Diversification, and Scope
• L26 - Entrepreneurship
• Browse content in L3 - Nonprofit Organizations and Public Enterprise
• L31 - Nonprofit Institutions; NGOs; Social Entrepreneurship
• Browse content in L4 - Antitrust Issues and Policies
• L41 - Monopolization; Horizontal Anticompetitive Practices
• L42 - Vertical Restraints; Resale Price Maintenance; Quantity Discounts
• L43 - Legal Monopolies and Regulation or Deregulation
• Browse content in L5 - Regulation and Industrial Policy
• L51 - Economics of Regulation
• L52 - Industrial Policy; Sectoral Planning Methods
• L59 - Other
• Browse content in L6 - Industry Studies: Manufacturing
• L66 - Food; Beverages; Cosmetics; Tobacco; Wine and Spirits
• Browse content in L8 - Industry Studies: Services
• L81 - Retail and Wholesale Trade; e-Commerce
• L89 - Other
• Browse content in M - Business Administration and Business Economics; Marketing; Accounting; Personnel Economics
• Browse content in M1 - Business Administration
• M11 - Production Management
• M13 - New Firms; Startups
• M14 - Corporate Culture; Social Responsibility
• Browse content in M2 - Business Economics
• M21 - Business Economics
• Browse content in M3 - Marketing and Advertising
• M30 - General
• M31 - Marketing
• M37 - Advertising
• M38 - Marketing and Advertising: Government Policy and Regulation
• M5 - Personnel Economics
• Browse content in N - Economic History
• Browse content in N5 - Agriculture, Natural Resources, Environment, and Extractive Industries
• N50 - General, International, or Comparative
• Browse content in O - Economic Development, Innovation, Technological Change, and Growth
• Browse content in O1 - Economic Development
• O12 - Microeconomic Analyses of Economic Development
• O13 - Agriculture; Natural Resources; Energy; Environment; Other Primary Products
• O14 - Industrialization; Manufacturing and Service Industries; Choice of Technology
• O16 - Financial Markets; Saving and Capital Investment; Corporate Finance and Governance
• O18 - Urban, Rural, Regional, and Transportation Analysis; Housing; Infrastructure
• O19 - International Linkages to Development; Role of International Organizations
• Browse content in O2 - Development Planning and Policy
• O20 - General
• O24 - Trade Policy; Factor Movement Policy; Foreign Exchange Policy
• Browse content in O3 - Innovation; Research and Development; Technological Change; Intellectual Property Rights
• O31 - Innovation and Invention: Processes and Incentives
• O32 - Management of Technological Innovation and R&D
• O33 - Technological Change: Choices and Consequences; Diffusion Processes
• O34 - Intellectual Property and Intellectual Capital
• Browse content in O4 - Economic Growth and Aggregate Productivity
• O44 - Environment and Growth
• Browse content in O5 - Economywide Country Studies
• O52 - Europe
• O54 - Latin America; Caribbean
• O55 - Africa
• Browse content in P - Economic Systems
• Browse content in P1 - Capitalist Systems
• P13 - Cooperative Enterprises
• P16 - Political Economy
• Browse content in P2 - Socialist Systems and Transitional Economies
• P22 - Prices
• P23 - Factor and Product Markets; Industry Studies; Population
• Browse content in P3 - Socialist Institutions and Their Transitions
• P31 - Socialist Enterprises and Their Transitions
• P32 - Collectives; Communes; Agriculture
• Browse content in Q - Agricultural and Natural Resource Economics; Environmental and Ecological Economics
• Browse content in Q0 - General
• Q00 - General
• Q01 - Sustainable Development
• Q02 - Commodity Markets
• Browse content in Q1 - Agriculture
• Q10 - General
• Q11 - Aggregate Supply and Demand Analysis; Prices
• Q12 - Micro Analysis of Farm Firms, Farm Households, and Farm Input Markets
• Q13 - Agricultural Markets and Marketing; Cooperatives; Agribusiness
• Q14 - Agricultural Finance
• Q15 - Land Ownership and Tenure; Land Reform; Land Use; Irrigation; Agriculture and Environment
• Q16 - R&D; Agricultural Technology; Biofuels; Agricultural Extension Services
• Q17 - Agriculture in International Trade
• Q18 - Agricultural Policy; Food Policy
• Q19 - Other
• Browse content in Q2 - Renewable Resources and Conservation
• Q20 - General
• Q21 - Demand and Supply; Prices
• Q22 - Fishery; Aquaculture
• Q23 - Forestry
• Q25 - Water
• Q26 - Recreational Aspects of Natural Resources
• Q27 - Issues in International Trade
• Q28 - Government Policy
• Browse content in Q3 - Nonrenewable Resources and Conservation
• Q34 - Natural Resources and Domestic and International Conflicts
• Browse content in Q4 - Energy
• Q41 - Demand and Supply; Prices
• Q42 - Alternative Energy Sources
• Q48 - Government Policy
• Browse content in Q5 - Environmental Economics
• Q50 - General
• Q51 - Valuation of Environmental Effects
• Q52 - Pollution Control Adoption Costs; Distributional Effects; Employment Effects
• Q53 - Air Pollution; Water Pollution; Noise; Hazardous Waste; Solid Waste; Recycling
• Q54 - Climate; Natural Disasters; Global Warming
• Q55 - Technological Innovation
• Q56 - Environment and Development; Environment and Trade; Sustainability; Environmental Accounts and Accounting; Environmental Equity; Population Growth
• Q57 - Ecological Economics: Ecosystem Services; Biodiversity Conservation; Bioeconomics; Industrial Ecology
• Q58 - Government Policy
• Q59 - Other
• Browse content in R - Urban, Rural, Regional, Real Estate, and Transportation Economics
• Browse content in R1 - General Regional Economics
• R11 - Regional Economic Activity: Growth, Development, Environmental Issues, and Changes
• R12 - Size and Spatial Distributions of Regional Economic Activity
• R14 - Land Use Patterns
• R15 - Econometric and Input-Output Models; Other Models
• Browse content in R2 - Household Analysis
• R23 - Regional Migration; Regional Labor Markets; Population; Neighborhood Characteristics
• Browse content in R3 - Real Estate Markets, Spatial Production Analysis, and Firm Location
• R32 - Other Spatial Production and Pricing Analysis
• R38 - Government Policy
• Browse content in R5 - Regional Government Analysis
• R52 - Land Use and Other Regulations
• R58 - Regional Development Planning and Policy
• Browse content in Y - Miscellaneous Categories
• Browse content in Y8 - Related Disciplines
• Y80 - Related Disciplines
• Browse content in Z - Other Special Topics
• Browse content in Z1 - Cultural Economics; Economic Sociology; Economic Anthropology
• Z13 - Economic Sociology; Economic Anthropology; Social and Economic Stratification
• Advance articles
• Editor's Choice
• Author Guidelines
• Submission Site
• Open Access
• Why Publish?
• About European Review of Agricultural Economics
• About European Agricultural and Applied Economics Publications Foundation
• Editorial Board
• Advertising and Corporate Services
• Journals Career Network
• Self-Archiving Policy
• Dispatch Dates
• Terms and Conditions
• Journals on Oxford Academic
• Books on Oxford Academic

Article Contents

1. introduction, 2. market power and price dynamics in the coffee market, 3. econometric model, 4. empirical analysis, 5. conclusion.

• < Previous

Coffee price dynamics: an analysis of the retail-international price margin

• Article contents
• Figures & tables
• Supplementary Data

Atanu Ghoshray, Sushil Mohan, Coffee price dynamics: an analysis of the retail-international price margin, European Review of Agricultural Economics , Volume 48, Issue 4, September 2021, Pages 983–1006, https://doi.org/10.1093/erae/jbab027

• Permissions Icon Permissions

We examine the dynamics of the margin between retail and international coffee prices from 1980 to 2018. We find no significant trend in the margin using a robust procedure for estimating a trend. Further, we establish that any deviations in the margin are transitory for the full sample as well as the periods prior to and after the demise of the ICA, but with asymmetric adjustment. One of the reasons for the observed asymmetry could be market concentration in the coffee supply chain at the coffee roasting level, which allows coffee roasters to keep a higher share of the profits.

Most agricultural commodities move through a complex processing and distribution supply chain. Coffee is no exception. The focus of this study is on retail and international coffee prices, which appear at the downstream level along the coffee supply chain. We examine the dynamics of the gap between retail and international coffee prices or in other words, the price margin. 1 We address key questions about the dynamics of how this margin is evolving over time; whether this margin is constant or increasing, and if there are any significant deviations from this margin, how the retail and international prices adjust to correct such deviations. This research is of significance as it helps policymakers and practitioners to address concerns being raised on the effects of growth in market concentration (consolidations and increased market power of multinational roasting companies) in the coffee roasting industry, which gives coffee roasters relatively higher market power to capture a larger share in the coffee supply chain.

The coffee value chain relates to all revenues generated by activities carried out along the coffee supply chain. In the coffee supply chain, the producer price is the cash price received at the ‘gate’ by coffee producers (referred to as producers hereafter). The international price is the price of coffee delivered at the first point of entry in coffee-consuming countries. 2 The international price therefore reflects the export price (c.i.f) 3 of coffee from coffee-producing countries. The retail price is the national urban US price of roasted ground coffee. 4 For the purpose of this paper, the retail price is the price of the green coffee equivalent of the roasted ground coffee 5 . The focus of the paper is to analyse the dynamics of the margin between retail and international coffee prices. Figure 1 shows how the retail and international prices of coffee have been evolving over the last four decades. We can see from the graph that the gap between the prices—which is the margin—tends to vary over time.

Retail (R) and international (W) coffee prices (in US cents/lb).

Prior to the 1990s, unilateral and multilateral interventions in coffee markets were common, the primary objective being price support and price stabilisation for the specific welfare of producers. The interventions were implemented by the International Coffee Agreement (ICA) in 1962, through a quota system to stabilise/support international prices and attenuate competition. The demise of the ICA in 1989 over a disagreement on quotas and the initiation of economic reforms in developing countries in the late 1980s and early 1990s resulted in most countries liberalising their coffee sector. As a result, the world coffee market has become more competitive and subject to market forces. It is widely felt that the end of the ICA regime in 1989 has resulted in a higher proportion of the income generated in the coffee supply chain retained in coffee-consuming countries. In an analysis of the value of the global coffee chain, the World Coffee Producers Forum, Colombia, 10–12 July 2017 , declared that the share reaching coffee-producing countries is very low, in contrast to that remaining in the hands of roasting companies in consuming countries. This is because of a shift in market dominance in favour of roasters over agents lower down in the coffee value chain. The paradox is while the coffee chain as a whole is profitable, the vast bulk of the profits are captured by the roasters, with adverse consequences for coffee-producing countries dependent on earnings from coffee exports. 6

However, not all agree that market concentration has contributed to the fall in the share in the value chain reaching coffee-producing countries, arguing that it is more of a rhetoric against multinational roasting companies. For example, Bettendorf and Verhoven (2000) , Feuerstein (2002) , Koerner (2002) and Durevall (2003 , 2018) conclude price transmission in the coffee market rejects the hypothesis that market power determines price transmission. These studies do accept that markets function imperfectly, and that price behaviour may not be an appropriate indicator of market power, given that a highly concentrated sector may be characterised by high price competition. Moreover, several studies argue that since the coffee markets have become more competitive, producers and exporters in the coffee supply chain have actually increased their returns from more efficient markets, rather than being worse off. 7 A possible reason for the different conclusions of the above-mentioned studies is the choice of the coffee price from the coffee supply chain employed in those studies. In this paper, we focus on the retail and international price of coffee because retail price is a relevant measure of the price charged by final processors of coffee in the retail market, and international price is the export price of coffee from coffee-producing countries after adding the cost of insurance and freight. The margin between the two prices includes the transfer costs as well as the profits of roasters. Therefore, the market power of roasters can be expected to be one of the factors affecting the dynamics of this margin.

We analyse the dynamics of the price margin by using a robust econometric model to test whether the margin between retail and international prices has increased over time and whether deviations from the margin are asymmetric. We use robust tests for estimating the trend that allows us to be agnostic of the order of integration of the data, a common problem found in agricultural prices (see Ghoshray, 2019 ). We further aim to determine whether retail and/or international prices respond to correct any deviation in the margin and whether retail prices adjust at a different rate compared to international prices. Accordingly, we aim to answer two broad research questions:

Question I: While the margin between retail and international prices is likely to fluctuate, are these fluctuations around a constant or a trend? In other words, can the margin be described as a long-run constant intertemporal equilibrium value? Or is it gradually increasing over time reflecting consolidation and increased market power over time in the roasting industry?

Question II: If such fluctuations were to occur, do they revert to this long-run intertemporal constant equilibrium value or the underlying equilibrium trend? If so, is the adjustment asymmetric thereby reflecting the presence of market power?

The rest of the paper is organised as follows. Section 2 presents an overview of the market concentration and price dynamics in the coffee market. Section 3 describes the econometric model, Section 4 describes the data and the empirical results and Section 5 concludes.

The general trend in the world coffee market has been the dominance and concentration of market power of multinational roasting companies in the coffee supply chain. In 1998, about two-thirds of the world’s coffee was purchased by five multinational companies who controlled nearly two-thirds of the world market share for roasted and instant coffees: Philip Morris (Kraft Jacob Suchard), Nestle, Procter and Gamble (P&G), Sara Lee and Tchibo. Table 1 shows the companies’ world market retail share for roasted and instant coffee in 1998 and 2014. The combined market share in 1998 of Nestle and Philip Morris was 49 per cent. The trend of market dominance and consolidation of coffee market continued until 2002, with Nestle and Kraft Jacob Suchard further consolidating their share of the world market for roasted and instant coffee ( Brown and Gibson, 2006 ; ActionAid and South Centre, 2008 ).

Share of global coffee market by roasters (1998 and 2014)

Source : Ponte (2002) ; Statistica (2016) .

Thereafter (post-2002), the roasting market has witnessed a gradual trend of embracing greater diversity, evident from the emergence of new players and a gradual fall in market share of roasters compared to 1998 (see Table 1 ). The bigger size of the coffee sector and larger geographical dispersion of consumption has allowed for the emergence of a large number of small roasters. In addition, the United States has seen the emergence of small specialty coffee roasters capturing a higher market share; this trend can also be seen in other European countries. Despite the gradual trend of market diversification, Nestle and Jacobs Douwe Egberts remain prominent players in the market, with 38 per cent combined market share of global roast and instant coffee in 2014, although lower than their share of 49 per cent in 1998 ( Statistica, 2016 ). 8 Since 2014, there have been reconsolidation efforts by Jacobs Douwe Egberts, going on a buying spree in its quest to challenge the long-standing market dominance of Nestle. Despite efforts of Nestle and Jacobs Douwe Egberts to hold on to their dominant position, there are signs of market diversification, albeit in a very gradual manner ( Grabs, 2017 ).

The overview shows high levels of dominance and concentration of market power in multinational roasting companies until around 2002 followed by a gradual dilution of this dominance due to some diversity and reorganisation in coffee roasting over the years. Despite the dilution, looking at Table 1 , one can say that there still is continued concentration of market power in multinational roasting companies. Given the market power in coffee supply chain, it is not surprising that large roasters are regularly alleged for leveraging their power to capture high share of the rents that accrue in the coffee value chain.

Market concentration can alter the marketing systems and can have an impact on the dynamics of price transmission in the coffee value chain. Market concentration could potentially weaken the ability of coffee-producing countries to influence international prices, while increasing the ability of coffee roasters in coffee-importing countries to influence international prices and the extent to which changes in international prices are passed on to retail prices (and vice versa), resulting in price transmission asymmetries in the coffee supply chain. As a case in point, there is evidence of price transmission asymmetries in supply chains for agricultural commodities. Various empirical studies focusing on food products find that increases in input (factor) prices are often transmitted more quickly to retail prices than decreases in these prices ( Serra and Goodwin, 2003 ; Meyer and Von Cramon-Taubadel, 2004 ; Lass, 2005 ; McLaughlin, 2006 ). The literature identifies market structure and the presence of non-competitive behaviour (i.e. market power) as the main cause for such asymmetry in price transmission ( Ward, 1982 ; Bacon, 1991 ; Borenstein, Cameron and Gilbert, 1997 ; Peltzman, 2000 ; Nakamura and Zerom, 2010 ). The literature identifies another explanation of imperfect transmission in the context of oligopolistic and monopsony markets. The risk of provoking a price war may make oligopolistic firms reluctant to lower their prices in response to a fall in input prices; therefore, price adjustment in response to the fall might be sluggish or take place only after time lags. In oligopolistic markets with unspoken collusion, oligopolistic firms will use price changes to signal the unspoken agreement ( Balke, Brown and Yücel, 1998 ; Brown and Yücel, 2000 ). When input prices rise, each oligopolistic firm will quickly adjust prices upwards to signal that collusion will be maintained, whereas the response of the oligopolistic firms will be slower when it comes to adjusting prices downwards when input prices fall to avoid undermining a tacit agreement. Further explanations by Kinnucan and Forker (1987) describe how government intervention can lead to asymmetric price transmission. Processors of agricultural commodities may believe that a reduction in input price may be temporary because it will trigger government intervention through support prices. In this context, processors will not react to a reduction in input prices, but they will quickly respond to increases in input prices because they will believe it is more likely to be long-lived.

The upshot from this discussion leads one to observe that asymmetric price adjustment behaviour can be relevant in the context of coffee. Where market power is in the hands of roasters, this would mean that increases in international prices would trigger a prompt increase in retail prices to ensure no reduction in roasters’ margins. However, decreases in international prices may not elicit the same response (i.e. prompt decrease in retail prices) as roasters are in a position to exploit their market power by keeping prices above the competitive level. There have been studies that analyse the impact of ICA termination on price transmission at various levels in the coffee supply chain ( Bohman, Jarvis and Barichello, 1996 ; Buccola and McCandlish, 1999 ; Krivonos, 2004 ; Mehta and Chavas, 2008 ; Fafchamps and Hill, 2008 ; Gómez, Lee and Koerner, 2009 ; Lee and Gómez, 2013 ). Shepherd (2005) examined the impact of the end of ICA on price transmission from producer to international prices and from international to retail prices employing a vector autoregression model. The results suggested that the ICA termination did not improve price transmission because of market power exerted by coffee roasters. Moreover, asymmetries in price transmission at all levels of the supply chain were identified, particularly during the post-ICA period. A set of studies (for example, Feuerstein, 2002 ; Shepherd, 2005 ) are concerned with short-term price transmission issues and seek evidence regarding the allegations of growing market power in the coffee roasting sector through the 1990s. They find that price transmission to the retail sector is asymmetric, with retail prices more responsive to increases than decreases.

Mehta and Chavas (2008) studied the price effects of the ICA termination and found that the short-run retail price response was greater for increases than for decreases in international prices during the post-ICA period. Lee and Gómez (2013) examine price transmission from international to retail coffee prices and find evidence of short-run asymmetries with differences among importing countries because of dissimilarities in market structures across countries. For example, in the United States, retail prices rise faster than they fall in response to changes in international prices, while in Germany and France, retail prices respond faster when international prices are falling. Leibtag et al. (2007) use price data over the period from 1997 to 2004 to study the path of raw material cost pass-through in the US coffee industry to gain insights on how changes in the input cost of coffee affect consumer (supermarket) prices of coffee. They do not find robust evidence that coffee prices respond more to increases than to decreases in coffee input costs. Nakamura and Zerom (2010) point out that firms have to pay a ‘menu cost’ to adjust the consumer prices which could result in price rigidity behaviour in terms of delayed and sluggish response to changes in input prices. However, the role of ‘menu cost’ is more valid in the short-run and for relatively small changes in input prices, in comparison to the long-run and for substantial changes in input prices where the role of menu costs is negligible. Subervie (2011) applies threshold cointegration to analyse the dynamics of international coffee price transmission to producers over the pre- and post-ICA periods. They find asymmetric price adjustments in the post-ICA period (large decreases in world prices being transmitted relatively quickly to producers as against increases) that can be seen as expressions of an unfavourable pricing influence over the post-ICA period. This could be because of emergence of new market structure over the post-ICA period, meaning that termination of the ICA may have failed to create competitive market structures in some cases.

In general, we cannot assume that high levels of roaster buyer power will necessarily lead to high international and retail price margin or even high-profit margins for roasters. This is because even if the higher buyer power allows roasters to exert pressure on keeping lower international prices, in a free market it can be expected that roasters would compete with each other from the seller side of the roasted coffee market. Furthermore, it is reasonable to expect that the margin between international and retail prices is likely to fluctuate, given the volatile nature of coffee prices in general. 9 The following econometric model is designed to capture these dynamic properties of the coffee price margin.

In this section, we lay out the framework for conducting the empirical test of coffee price margin adjustment. We particularly focus on the possibility that the cyclical adjustment of the coffee price margin around its long-term mean or trend might be asymmetric in the light of the arguments made in sections 1 and 2 . The speed or momentum at which price margins fluctuate around the long-run mean or trend may differ, depending on whether the prices are increasing or decreasing relative to the mean.

To obviate this problem of possible non-stationarity of the price margin, we employ the robust procedure of trend estimation according to Perron and Yabu (2009a) . This method allows one to be agnostic to the underlying order of integration of the data. A quasi-feasible generalised least squares (q-FGLS) procedure is applied to obtain the estimate of |$\beta $| the trend parameter, denoted |$\hat \beta $|⁠ , and construct the robust q-FGLS t -statistic for the unbiased and median unbiased estimate, that is, |$t_\beta ^{RQF}\left( {UB} \right)$| and |$t_\beta ^{RQF}\left( {MU} \right)$|⁠ , respectively (see Perron and Yabu, 2009a ). For completeness, we calculate both the unbiased and the median unbiased estimates.

To estimate the dynamics of the price margin, we employ the momentum threshold autoregressive (MTAR hereafter), which is a procedure proposed by Enders and Granger (1998) . This procedure is particularly attractive as it neatly fits in with the estimation and hypothesis testing as in this case of coffee price adjustment. As a prelude to determining whether adjustments of the margin are asymmetric or not, we need to establish that deviation of the margin is transitory in nature. As Enders and Granger (1998) stipulate, that rejection of the unit root null in the margin would allow us to test for asymmetric adjustment. This is reinforced by their view that testing for unit roots are mis-specified if the underlying adjustment is asymmetric.

Accordingly, in this study, we use the MTAR method by Enders and Granger (1998) .

We use the methodology proposed by Chan (1993) to estimate the threshold denoted |$\tau$|⁠ . 11 The estimated residual series are sorted in ascending order, that is, |${\Delta}{Z_1} \lt {\Delta}{Z_2} \lt \ldots \lt {\Delta}{Z_T}$|⁠ , where |$T$| denotes the number of usable observations. The largest and smallest 15 per cent of the |${\Delta}{Z_t}$| series are eliminated, and each of the remaining 70 per cent of the values were considered as possible thresholds. For each of the possible thresholds, the equation was estimated using equations  ( 2 ) and ( 3 ). The estimated threshold yielding the lowest residual sum of squares was deemed to be the appropriate estimate of the threshold.

The null hypothesis of a unit root in the coffee price margin is given by the following testable hypothesis, that is, |${H_0}:\left( {{\gamma _1} = {\gamma _2} = 0} \right)$|⁠ , which is obtained from estimating equation  ( 2 ) and comparing to the critical values computed by Enders and Granger (1998) , against the alternative |${H_A}: $| |${\gamma _1} \lt 0 $| and/or |${\gamma _2} \lt 0 $|⁠ . Note that under the null hypothesis, the margin is symmetric and there is a unit root in both regimes; while under the alternative hypothesis, there is a stationary process in at least one regime (see Enders and Granger, 1998 ). If we can reject the null hypothesis, it is possible to test for asymmetric adjustment, that is, |${H_0}:\left( {{\gamma _1} = {\gamma _2}} \right)$|⁠ , using the |$F$| statistic. If the test statistic is greater than the critical value from the |$F$| table, we can reject that the adjustment to any deviation is symmetric, enabling us to conclude that there is asymmetric adjustment. Diagnostic checking of the residuals is undertaken using the Ljung–Box |$Q$| tests to ascertain whether the |${\omega _t}$| series in equation  ( 2 ) is a white noise process, and to ensure the residuals are white noise, the right-hand side of equation  ( 2 ) is augmented by lagged variables given by |$\sum_{i = 1}^p {\phi _i}{\Delta}{Z_{t - i}}$|⁠ . The lag length |$p$| in equation  ( 2 ) is determined by the ‘General to Specific’ criterion.

Finally, we carry out innovation accounting by tracing out the responses to exogenous shocks to the MTAR model. Following Coakley, Fuertes and Zoega (2001) , we use the recursive nature of the model to create a history |${h_{t - 1}} = \left\{ {{Z_{t - 1}}, {Z_{t - 2}}, \ldots , } \right\}$| and time |$t$| shock |${\omega _t}$|⁠ , which serves as an initial condition, and |$n$| randomly selected shocks |${V_t} = \left\{ {{v_{t + 1}}, {v_{t + 2}}, \ldots ,{v_{t + n}}} \right\}$|⁠ . Using the parameter estimates of the MTAR model, we generate |$k$| sets of forecasts for the shocked model |$\left\{ {{Z_{t + i}}\left( {{h_{t - 1}},{\omega _t}} \right)} \right\}_{i = 0}^n$| and |$k$| sets of baseline forecasts |$\left\{ {{Z_{t + i}}\left( {{h_{t - 1}}} \right)} \right\}_{i = 0}^n$| using the same history and random future shocks in all the forecasts (see Coakley, Fuertes and Zoega, 2001 ). The generalised impulse response function can be defined as follows:

|$IRF\left( {i, {h_{t - 1}},{\omega _t}} \right) = E\left[ {{Z_{t + i}}|{h_{t - 1}},{\omega _t}} \right] - E\left[ {{Z_{t + i}}|{h_{t - 1}}} \right]$|⁠ ,   |$i = 0,1,2, \ldots ,n$|

The difference in the two forecasts is averaged over the |$k$| replications and repeated 100 times for different combinations of history and shocks.

This section is structured into three sub-sections, we first describe the data, followed by the robust estimation of the trend of the margin and then the estimation of the asymmetric price adjustment of the margin.

All price data are in nominal terms and measure the monthly average price in US cents per pound for the period from January 1980 to May 2018 (see Figure 1 ). As a measure of retail price, we use the monthly averages of the national urban US price of roasted ground coffee obtained from the US Bureau of Labor Statistics (2020) . 12 The data series for the period 1980–2018 has missing values for the monthly averages of September and October 2007 and of January 2008 to November 2009. However, the annual averages for the years 2007, 2008 and 2009 are available from the International Coffee Organization (ICO); we have interpolated the missing values using annual averages for the year. The retail price reflects only the price of roasted ground coffee; prices of whole bean gourmet coffee and coffee drinks are not reflected in this price (see Mehta and Chavas, 2008 ). We focus on roasted ground coffee since green beans is the input in the production process and the product has been homogenous over the years. By input we mean raw material input; the conversion process to roasted ground coffee does include other inputs such as labour and machinery. However, the other input costs are much lower for roasted ground coffee compared to instant coffee or other forms of coffee, which explains our choice of retail price to reflect the price of roasted ground coffee. In accordance with internationally accepted practices, the ICO has stipulated conversion factors to convert different types of coffee to green bean equivalent (GBE); converting roasted coffee to GBE requires multiplying the weight of roasted coffee by 1.19. Therefore, for the retail price to reflect the price of same quantity of coffee, we convert the retail price of ground coffee to GBE price by dividing the retail price by 1.19. 13

The bulk of the coffees used in roasted ground coffee blends are Brazilian Natural grade Arabica and Robusta (both Asian and Brazilian Robusta); rough estimates place the share of Arabica and Robusta coffee in the roasted ground coffee blends at 60 and 40 per cent, respectively. For international price to be comparable with the retail price of ground coffee, we use a weighted average price of Arabica (60 per cent) and Robusta (40 per cent) coffee. As a measure of international price, we use the weighted average of the ICO monthly Indicator Price for Brazilian Natural Arabica (60 per cent weight) and Robusta (40 per cent weight) for the period from 1980 to 2018. The prices are available on a monthly average basis from the ICO database and are calculated by weighting the ex-dock prices on the international markets in New York, Bremen/Hamburg and Le Havre/Marseilles markets ( ICO, 2020 ). 14

4.2. Estimating the trend of the price margin

The trend estimate |$\hat \beta = 0.355$| is positive in magnitude but statistically not different from zero as shown by the standard error 0.397 in parentheses. The associated t -statistic for the trend estimate is calculated to be 0.894, which falls below the conventional levels of significance. We repeat the robust trend estimation procedure using the median unbiased statistic and the results are identical. 15 We can conclude from this result that the trend in the coffee price margin is not significant. While the sign is positive, there is too much variability around this trend to concur that it is significantly positive. This result is not surprising, given the amount of variation we observe in the graph of the margin in Figure 2 . Besides, international prices can lead to higher volatility in retail prices with a lag (see Mehta and Chavas, 2008 ), which contributes to the large variation of the margin over time. Using the robust trend estimate of Perron and Yabu (2009a) , we calculate the 90 per cent confidence interval of the trend estimate to be (–0.299, 1.01), and the 95 per cent confidence interval is (–0.423, 1.13). Both confidence intervals cannot exclude zero, thereby rendering the trend to be insignificant.

Price margin of retail-international coffee prices.

Figure 2 tends to reflect that there may be a slight positive trend, but the huge amount of variability around the trend renders the trend estimate to be statistically insignificant. Our robust procedure is in line with the finding by Mehta and Chavas (2008) where they too find an insignificant trend estimate of the retail-international price margin using a ‘delta method’. This result departs from those studies that concluded an increasing margin such as Talbot (1997) , Calfat and Flores (2002) and Ponte (2002) . However, these latter studies do not use robust econometric methods.

However, one may argue that there could be one or more structural breaks in the trend, thereby causing the estimate to be insignificant. To address this, we make use of robust structural break tests. First, we apply the Perron and Yabu’s (2009b) quasi-feasible Wald (W-QF) test for a single structural break in the trend. Given the length of the data sample, we also apply the sequential test for multiple structural breaks using the robust Supremum F test (sup-F) procedure of Sobreira and Nunes (2016) . Both tests are robust, allowing us to be agnostic of the underlying order of integration of the data series. The results of the structural break tests are given in Table 2 .

Robust tests for structural breaks

Notes : The single break test by Perron and Yabu (2009b) is given by the quasi-Feasible Wald (W-QF) test. The test is carried out using the null hypothesis of no break against a single break or |$\left( {0|1} \right)$|⁠ . The critical values (c.v.) at the 5 and 10 per cent significance levels are reported alongside the test statistics. The multiple breaks by Sobreira and Nunes (2016) are given in the second column of results using the sup-F test. These are sequential tests that test for no break against one (0|1), two (0|2) and three (0|3) breaks separately.

Using the single structural break test by Perron and Yabu (2009b) , we find the estimated W-QF test statistic to be below the critical value at standard conventional levels. We therefore cannot reject the null hypothesis of no break against the alternative of a single break [that is, |$\left( {0|1} \right)$| ] in the trend. Based on the robust single break test, our conclusion is that there is no evidence of breaking trends. Further, as a confirmatory test, we apply the robust multiple break test procedure by Sobreira and Nunes (2016) allowing for up to 3 breaks based on the sample size. In each of the cases, we cannot reject the null hypothesis of no break against one break [that is, |$\left( {0|1} \right)$| ], followed by no break against two breaks [that is, |$ \left( {0|2} \right)$| ], and finally, no break against three breaks [ |$\left( {0|3} \right)$| ]. The upshot is that there is no evidence of any breaking trends; therefore, trend estimation is secular, but we find the estimate to be insignificant. Hereafter, the econometric analysis of the dynamics of the margin excludes the presence of the trend.

4.3. Estimating the dynamics of the price margin

We test for the dynamic behaviour of the price margin using the MTAR model. The delay parameter for the models is set as |$d = 1$|⁠ . The MTAR model is estimated using equations ( 3 ) and ( 4 ), and the non-zero threshold is calculated using the method by Chan (1993) . 16 The results are presented in Table 3 .

Results of the MTAR model for the full sample

Notes : a and b denote rejection of the null at the 1 and 5 per cent significance levels, respectively. The numbers in parentheses denote t -statistics, and the numbers in square brackets denote p -values. The critical value of the MTAR test at the 1 per cent significance level is 6.99. The results in Table 3 show that we can reject the null at the 1 per cent significance level.

In the first column of Table 3 , we set out the hypothesis of interest. First, we report the estimate of the |$ {\gamma _1}$| parameter, which provides the rate of adjustment when the change in retail prices is greater than the change in international prices, thereby creating a ‘positive deviation’ in the margin. The opposite deviation, labelled as a ‘negative deviation’ in the margin, returns the parameter estimate |$ {\gamma _2}$|⁠ , which provides the rate of adjustment when the retail prices are changing at a slower rate than international prices making the deviation shrink, which we will call ‘negative’. The null hypothesis of non-stationarity is given by |${H_0}:\left( {{\gamma _1} = {\gamma _2} = 0} \right)$|⁠ , which simply states that the price margin is a random walk. Subject to rejecting the null of non-stationarity, the null hypothesis of symmetry is given by |${H_0}:\left( {{\gamma _1} = {\gamma _2}} \right)$|⁠ , which states that there is no significant difference for the rates of adjustment given by |${\gamma _1}$| and |${\gamma _2}$|⁠ . This is followed by a diagnostic test to determine whether there is no serial correlation in the residuals of the regression equations given by equation  ( 3 ).

We can reject the null of non-stationarity, that is, |${H_0}:\left( {{\gamma _1} = {\gamma _2} = 0} \right)$| (given by the test statistic equal to 16.97, significant at the 1 per cent level) in the margin. Note that under the alternative hypothesis (that is, |${H_A}: $| |${\gamma _1} \lt 0 $| and/or |${\gamma _2} \lt 0 $|⁠ ), there is a stationary process in at least one regime (see Enders and Granger, 1998 ); this is found to be when there is a negative deviation. We find the negative discrepancy is corrected at the rate of 28.9 per cent every month based on the parameter estimate of |${\gamma _2} = - 0.289$|⁠ . However, for a positive discrepancy, we find no signs of adjustment. The parameter estimate of |${\gamma _1} = - 0.006$| is found to be statistically insignificant. The null hypothesis of symmetry, given by |${H_0}:\left( {{\gamma _1} = {\gamma _2}} \right)$|⁠ , is rejected as shown by the p -value of 0.004, thereby concluding that there is asymmetric adjustment. To sum up, we can conclude from the MTAR model that a positive deviation in the margin tends to persist, whereas a negative deviation in the margin is rapidly corrected, thereby underscoring the case for asymmetric adjustment.

Although there is no evidence of single or multiple structural breaks in the margin, we conduct further estimations to determine whether the dynamics of price adjustment are the same prior and after the elimination of the export quota system in 1989 following the collapse of the ICA. Accordingly, we divide the full sample into two regimes: Regime I, from January 1980 to August 1989, and Regime II, from September 1989 to May 2018. We apply the MTAR model to both regimes to check whether the underlying price dynamics remain the same or change. The results are presented in Table 4 .

Results of the MTAR model for the two regimes

Notes : a Denotes rejection of the null at the 1 per cent significance levels. The numbers in parentheses denote t -statistics, and the numbers in square brackets denote p-values. The critical value of the MTAR test at the 1 per cent significance level is 6.99. The results in Table 4 show that we can reject the null at the 1 per cent significance level.

In some respect, the empirical results lead to the same general conclusion for both regimes. For both regimes, we can reject the null of non-stationarity, that is, |${H_0}:\left( {{\gamma _1} = {\gamma _2} = 0} \right)$| in the margin (which is given by the test statistic equal to 30.26, significant at the 1 per cent level in Regime I, and equal to 10.15, significant at the 1 per cent level in Regime II). Note that under the alternative hypothesis (that is, |${H_A}: $| |${\gamma _1} \lt 0 $| and/or |${\gamma _2} \lt 0 $|⁠ ), there is a stationary process in at least one regime, which can be expected in the MTAR model (see Enders and Granger, 1998 ). In both cases, the adjustment takes place for the negative deviation given by the parameter estimate |${\gamma _2}$|⁠ . In Regime I, any deviation in the steady-state margin is corrected at the rate 89 per cent every month and in Regime II at the rate of 11 per cent every month. In this case, the discrepancy occurs when the margin shows a negative discrepancy. These conclusions on adjustment of price margins for each separate regime are no different from that of the full sample. When considering the parameter estimate |${\gamma _1}$|⁠ , we find that the rates of adjustment are statistically insignificant. That is, the parameter estimates of 0.003 in Regime I and 0.002 in Regime II are reported to be statistically insignificant, so that a positive deviation in the steady state margin is allowed to persist. We find that prior to the elimination of export quotas, adjustment to correct any negative discrepancy is relatively fast (i.e. 89 per cent of the deviation is corrected every month in Regime I), in comparison to the period when quotas were eliminated (where 11 per cent of the deviation is corrected every month in Regime II). We can confirm that there is asymmetric adjustment, by rejecting the null hypothesis of symmetry, given by |${H_0}:\left( {{\gamma _1} = {\gamma _2}} \right)$| as shown by the p -values of 0.06 in Regime I and 0.09 in Regime II, showing rejection at the 10 per cent significance level. Apart from the differences in the speed of adjustment, the dynamics for the full sample reflects that of the sub-samples corresponding to the regimes prior to and after the collapse of the ICA.

Since we find the margin to be stationary with MTAR adjustment, we proceed to estimate an asymmetric error correction model (AECM) as described by equations ( 5 ) and ( 6 ). The results of the AECM are given in Table 5 . The results include the full sample and the two regimes.

Results of the ECM for the full sample and the two regimes

The numbers in parentheses denote the t -statistics, while the numbers in square brackets are probability values.

We first consider the full sample results. During the phase, when there is a positive deviation in the margin, we find that such a deviation is corrected by the retail prices, albeit at a very slow rate of 1 per cent every month. International prices do not adjust to correct this deviation, as the speed of adjustment parameter (–0.003) is statistically insignificant. However, if there is negative discrepancy in the margin, then we find that such a deviation is corrected by both retail and international prices. International prices adjust at the rate of 16.6 per cent every month. In comparison, retail prices adjust at the rate of 8.7 per cent every month. These results allow us to determine how retail and international prices adjust, and we find the former adjusts at a sluggish rate to the latter, and both only correct any deviation during the phase when the price margin shows a negative discrepancy. In the short run, we find bidirectional causality, where a feedback effect is found to exist between international and retail prices. This implies changes in retail prices lead to a change in international prices and vice versa in the short run. The model is free from serial correlation as shown by the Ljung–Box Q test statistics.

When considering Regime I, we find the dynamics to be different. When the price margin shows a positive discrepancy, neither the retail nor international prices adjust to correct the deviation. When a negative discrepancy occurs, we find that such a deviation is corrected by both prices. We find the retail prices adjust at a relatively slower rate (35 per cent every month) compared to the international prices (47 per cent), and this adjustment occurs during the phase when the price margin shows a negative discrepancy. In this period, prior to the collapse of the ICA, we find the absolute value of the speed of adjustment to be faster when compared to the speed of adjustment coefficients for the full sample. The short run shows that the Granger causality is bidirectional, similar to the dynamics of the full sample. No problems with serial correlation are reported. In the case of Regime II, the correction to any deviation is distinct to the full sample results and Regime I. If there is a negative deviation in the margin, then only the retail prices adjust at the rate of 8.9 per cent every month to close the deviation. On the contrary, when there is a positive deviation in the margin, then only international prices adjust at the rate of 2.1 per cent every month to eliminate the deviation. In the short run, we find the Granger causality results to be no different to the other regimes, where we find feedback effects. The Ljung–Box Q test statistics show that there are no issues with serial correlation.

It has been well documented in the literature that supply shocks, such as adverse weather, can cause large variations in price, which can be further exacerbated if the demand functions for a commodity such as coffee are price inelastic, causing innovations in the price margin. To analyse the short-run adjustment in the margin, we conduct an innovation accounting exercise by making use of generalised impulse response analysis. As noted by Koop, Pesaran and Potter (1996) , the response to a price shock in models that show symmetric adjustment is independent of the history of the time series and the sign and magnitude of the given shock. However, for models that display asymmetric adjustment, which implies non-linearity, the impulse response functions are functions of the history of the price series and the sign and magnitude of the shock. Since we are conducting impulse responses on a MTAR model, we adopt the non-linear approach proposed by Koop, Pesaran and Potter (1996) where they make use of a generalised impulse response function. We follow the procedure as described in the section 3 by Coakley, Fuertes and Zoega (2001) , where they highlight the superiority of the generalised impulse response function in a MTAR model.

The impulse response function is estimated by averaging the 100 individual draws. The response to a positive and negative shock are given in Figures 3 and 4 , respectively.

Generalised impulse response function to a positive change |${\Delta}{Z_{t - 1}} \ge 0$| shock.

Generalised impulse response function to a negative change |${\Delta}{Z_{t - 1}} \lt 0$| shock.

A response is computed for each phase, one phase being where the change in retail prices is greater than international prices (labelled as a positive deviation) shown in Figure 3 , and the other phase being where the change in retail prices is less than international prices (labelled as a negative deviation) shown in Figure 4 . In both cases, the responses for each phase are computed by randomly selecting histories from the data, such that |${\Delta}{Z_{t - 1}} \ge 0$| (positive change) and |${\Delta}{Z_{t - 1}} \lt 0$| (negative change). The horizon is set to be 12 months and all the responses are normalised so that the initial effect of the shock is unity for all histories. Figure 3 shows the response of the price margin to a positive unit shock. As expected, the response of a shock is insignificant as shown by the standard error bands containing the value zero over the entire time horizon. The response after a unit shock is generally flat, after a very slight increase in the month following the shock. This is expected, given the insignificant estimate that we obtain for |${\gamma _1}$| from the MTAR model (see Table 3 ). The impulse response that we obtain for a negative shock is shown in Figure 4 . We find that in response to a unit shock, the margin increases slightly in the next month but thereafter starts to gradually dissipate. The response is no longer significant after 7 months which underscores the rate of adjustment being significant in the MTAR model for negative deviations in the margin.

This paper examines the price dynamics of the margin between retail and international coffee prices. Analysing monthly data from 1980 to 2018, we make several contributions. First, we find no evidence of a statistically significant increasing trend in the margin between retail and international prices. Our findings lend support to those by Mehta and Chavas (2008) and are backed by robust tests for estimating the trend in the margin. We analysed this issue further by conducting tests for structural breaks, and we found no evidence of the possibility of any breaking trends. We conclude the trend estimate is statistically insignificant, and the variability in the margin dominates any possible underlying trend in the margin. Second, we establish that any deviations in the margin are transitory for the full sample as well as the periods prior to and after the demise of the ICA but with asymmetric adjustment. Positive discrepancies tend to persist, whereas negative discrepancies are corrected through time. Third, we uncover the short-run and long-run dynamics from the AECM. The long-run analysis shows that adjustments to the increase or decrease in the margin are asymmetric. In the case of the full sample, during the phase where the margin shows a positive discrepancy, only retail price adjusts at a very slow but significant rate to correct the discrepancy. During the phase when the margin has a negative discrepancy, retail and international prices adjust to correct the deviation. In this phase, when the adjustment takes place, the retail prices adjust at a relatively slower rate compared to the international prices. This pattern changes in the regime prior to the collapse of the ICA as well as the regime after the demise of the ICA. In the regime when the ICA was in place, retail and international prices only adjust to a negative deviation. In the post-ICA regime, retail prices adjust to correct a negative deviation while international prices adjust to correct a positive deviation. In each case we find when the margin shows a negative discrepancy, the retail prices adjust to correct the deviation, thereby attempting to maintain their margin. One of the reasons for the observed asymmetry we find for the entire sample could be market concentration (i.e. oligopsony power) in the coffee supply chain at the roasting level, which allows roasters to keep a higher share of the rents/profits by keeping the retail prices higher compared to the international prices. Our results lend support to the idea that power might be concentrated in the hands of large roasters and policies may be needed to be devised that help promote competition. For example, they could include promoting greater market diversification in the coffee roasting industry to increase competition in the coffee roasting sector; promoting coffee roasting facilities in large coffee-producing countries to reduce the distance between coffee suppliers and coffee roasters for greater coffee market integration; and providing easier access to credit finance and price risk management (financial) instruments to economic agents in the coffee supply chain to improve competition in the coffee market.

A limitation of the study is that the margin between retail and international prices is not simply a markup over marginal cost but includes costs of conversion of coffee to roasted ground coffee, and therefore, changes in the conversion costs will result in changes in the margin. As stated earlier, we choose roasted ground coffee since the product has been generally homogenous over the years. Although the conversion process from coffee to roasted ground coffee does include other inputs such as labour and machinery, these inputs are relatively lower for roasted ground coffee compared to other forms of coffee. Moreover, one can expect that increases in labour (wage) cost over the years are to some extent compensated by technological improvements in the conversion process. We should therefore keep this in mind in drawing conclusions relating to the trend in the margin and acknowledge that the price dynamics of the margin can be influenced by factors in addition to market power of roasters.

It may be worth mentioning that our study is restricted to mainstream coffee, while an important trend in the coffee market is the growth of niche and specialty markets (including sustainable coffee), making the coffee market highly differentiated. 17 Fitter and Kaplinsky (2001) argue that the benefits from the differentiated coffee market do not trickle down to coffee-producing countries because roasters are buying a more homogenous coffee in the mainstream market (at more or less the same price) and differentiate their offering through product proliferation to increase their returns. Not everyone accepts this view. The other view is that the niche and specialty market demands differentiated coffee (higher quality or coffee that meets particular production standards), and such coffee is usually in limited supply, which allows their suppliers to capture higher prices. We feel this is an area that calls for further research, for example, how coffee product differentiation affects coffee market competition and coffee price dynamics.

Unless specified otherwise, ‘coffee’ means green (raw or unroasted) beans and coffee prices imply prices of green beans.

The same definition of international coffee prices has been used in Shepherd (2005) , Fafchamps and Hill (2008) , Gómez, Lee and Koerner (2009) , Subervie (2011) and Lee and Gómez (2013) .

The export price includes the cost of insurance and freight and any applicable custom duties. This is the price actually paid for physical deliveries of coffee (green bean) on the dock at port of destination.

This definition of retail price has been used in Mehta and Chavas (2008) .

More details about retail price are found in Section 4.1 where we describe the data.

For studies on implications of market concentration for producers, see Ponte (2002) , Daviron and Ponte (2005) , Muradian and Pelupessy (2005) , ActionAid and South Centre (2008) , Hoekman and Martin (2012) , Sexton (2013) and Igami (2015) . For general studies on the low returns to producers and coffee-producing countries, see Calfat and Flores (2002) , McCorriston, Sexton and Sheldon (2004) , Shepherd (2005) , Gibbon (2007) , Levy (2008) and World Vision (2014) .

For studies on higher returns to economic agents in the coffee supply chain in coffee-producing countries, see, for example, Raffaeli (1995) , Bohman, Jarvis and Barichello (1996) , Gilbert (1996) , McIntire and Varangis (1999) , Krivonos (2004) , Jarvis (2005) , Gemech et al. (2011) and Russell, Mohan and Banerjee (2012) .

Jacobs Douwe Egberts is a Dutch privately owned company that owns numerous beverage brands. It was formed in 2012 following the merger of Philip Morris (the coffee division of Mondelez International) with Douwe Egberts.

The high volatility of coffee prices is due to the susceptibility of output to frosts, disease and droughts, magnified by the inelasticity of demand with respect to prices and income and price inelasticity of supply ( Mehta and Chavas, 2008 ).

This is based on a rule suggested by Schwert (1989) .

Enders (2001) points out that the demean series will be a biased estimator of the threshold |$\tau$| as |${\gamma _1}$| and |${\gamma _2}$| differ which motivates the estimation of the consistent threshold which is super-consistent using the grid search approach by Chan (1993) . Note, while calculating the critical value Enders (2001) contains an error which is pointed out by Cook and Manning (2003) where they show that the power of consistent threshold MTAR model is found to be higher than all forms of plausible alternatives using the newly designed critical values which we consider in this paper. Enders (2001) concedes in his paper that the results he obtains for the MTAR are counter-intuitive as the consistent threshold M-TAR model employs a consistent estimator of the threshold and therefore should have increased power.

Data can be accessed at US Bureau of Labor Statistics, Consumer Price Index, Average Price data, coffee, 100 per cent, ground roast, all sizes, per lb; Series Id: APU0000717311.

This adjustment is made simply for comparison purposes. The price only changes in scalar terms and is not affected when conducting the econometric analysis of the price dynamics.

Data can be accessed at ICO, composite & group indicator prices—monthly averages; data prior to 1990 are available on request from the ICO.

The results are not reported here but are available on request.

We employ a non-parametric approach by Tsay (1989) to identify the existence of threshold effects in the AR component of the model. This test returns an F statistic to test the null hypothesis of no changes in the parameter estimates of the AR representation of the data. We find that the null can be rejected suggesting the existence of a threshold, prompting us to use a threshold model instead of a simple AR. For brevity, we do not report the details, and the results are available on request.

There is no universally accepted agreement on what constitutes as speciality coffee, so it is difficult to exactly quantify the share of specialty coffee of total coffee sold in retail outlets, although industry reports estimate the share to be over 20 per cent in the US and European market.

ActionAid and South Centre . ( 2008 ). Commodity dependence and development: suggestions to tackle the commodities problem . Report prepared for the ActionAid and the South Centre, April . Johannesburg and Geneva .

Bacon R. W. ( 1991 ). Rockets and feathers: the asymmetric speed of adjustment of UK retail gasoline prices to cost changes . Energy Economics 13 ( 3 ): 211 – 218 . doi: 10.1016/0140-9883(91)90022-R

Google Scholar

Balke N. S. , Brown S. P. A. and Yücel M. K. ( 1998 ). Crude oil and gasoline prices: an asymmetric relationship? Federal Reserve Bank of Dallas Economic Review , First Quarter 1998 : 2 – 11 .

Bettendorf L. and Verboven F. ( 2000 ). Incomplete transmission of coffee bean prices: evidence from the Netherlands . European Review of Agricultural Economics 27 ( 1 ): 1 – 16 . doi: 10.1093/erae/27.1.1

Bohman M. , Jarvis L. and Barichello R. ( 1996 ). Rent seeking and international commodity agreements: the case of coffee . Economic Development and Cultural Change 44 ( 2 ): 379 – 404 . doi: 10.1086/452219

Borenstein S. , Cameron A. C. and Gilbert R. ( 1997 ). Do gasoline prices respond asymmetrically to crude oil price changes? The Quarterly Journal of Economics 112 ( 1 ): 305 – 339 . doi: 10.1162/003355397555118

Brown O. and Gibson J. ( 2006 ). Boom or Bust: Developing Countries’ Rough Ride on the Commodity Price Rollercoaster . Manitoba, Canada : International Institute for Sustainable Development .

Google Preview

Brown S. P. A. and Yücel M. K. ( 2000 ). Gasoline and crude oil prices: why the asymmetry? Economic and Financial Policy Review Federal Reserve Bank of Dallas, issue Q3 : 23 – 29 .

Buccola S. T. and McCandlish J. E. ( 1999 ). Rent seeking and rent dissipation in state enterprises . Review of Agricultural Economics 21 ( 2 ): 358 – 373 .

Calfat G. and Flores R. ( 2002 ). Government actions to support coffee producers an investigation of possible measures from the European Union side . Economics Working Paper , vol. 448 . Brazil : Getulio Vargas Foundations .

Chan K. S. ( 1993 ). Consistency and limiting distribution of the least squares estimator of a threshold autoregressive model . Annals of Statistics 21 : 520 – 533 . doi: 10.1214/aos/1176349040

Coakley J. , Fuertes A. M. and Zoega G. ( 2001 ). Evaluating the persistence and structuralist theories of unemployment from a nonlinear perspective . Studies in Nonlinear Dynamics and Econometrics 5 ( 3 ): 1 – 25 . doi: 10.2202/1558-3708.1078

Cook S. and Manning N. ( 2003 ). The power of asymmetric unit root tests under threshold and consistent-threshold estimation . Applied Economics 35 ( 14 ): 1543 – 1550 . doi: 10.1080/0003684032000125105

Daviron B. and Ponte S. ( 2005 ). The Coffee Paradox: Global Markets, Commodity Trade and the Elusive Promise of Development . London : ZED Books .

Durevall D. ( 2003 ). Competition and pricing: an analysis of the market for roasted coffee. Chapter 5. In: High Prices in Sweden – a Result of Poor Competition? Swedish Competition Authority Report .

Durevall D. ( 2018 ). Cost pass-through in the Swedish coffee market . European Review of Agricultural Economics 45 ( 4 ): 505 – 529 . doi: 10.1093/erae/jby002

Enders W. ( 2001 ). Improved critical values for the Enders-Granger unit-root test . Applied Economics Letters 8 ( 4 ): 257 – 261 . doi: 10.1080/135048501750104033

Enders W. and Granger C. W. J. ( 1998 ). Unit-root tests and asymmetric adjustment with an example using the term structure of interest rates . Journal of Business and Economic Statistics 16 : 304 – 311 .

Fafchamps M. and Hill R. V. ( 2008 ). Price transmission and trader entry in domestic commodity markets . Economic Development and Cultural Change 56 ( 4 ): 729 – 766 . doi: 10.1086/588155

Feuerstein S. ( 2002 ). Do coffee roasters benefit from high prices of green coffee? International Journal of Industrial Organization 20 : 89 – 118 . doi: 10.1016/S0167-7187(00)00067-9

Fitter R. and Kaplinsky R. ( 2001 ). Who gains from product rents as the coffee market becomes more differentiated? A value-chain analysis observed in other agricultural-based value chains . IDS Bulletin 32 ( 3 ): 69 – 82 . doi: 10.1111/j.1759-5436.2001.mp32003008.x

Gemech F. , Mohan S. , Reeves A. and Struthers J. ( 2011 ). Market-based price-risk management: welfare gains for coffee producers from efficient allocation of resources . Oxford Development Studies 39 ( 1 ): 49 – 68 . doi: 10.1080/13600818.2010.550399

Ghoshray A. ( 2019 ). Are shocks transitory or permanent? An inquiry into agricultural commodity prices . Journal of Agricultural Economics 70 ( 1 ): 26 – 43 . doi: 10.1111/1477-9552.12277

Gibbon P. ( 2007 ). Africa, tropical commodity policy and the WTO Doha Round . Development Policy Review 25 ( 1 ): 43 – 70 . doi: 10.1111/j.1467-7679.2007.00359.x

Gilbert C. ( 1996 ). International commodity agreements: an obituary notice . World Development 24 ( 1 ): 1 – 19 . doi: 10.1016/0305-750X(95)00121-R

Gómez M. I. , Lee J. and Koerner J. ( 2009 ). Do retail coffee prices rise faster than they fall? Asymmetric price transmission in France, Germany and the United States . Journal of International Agricultural Trade and Development 6 : 175 – 196 .

Grabs J. ( 2017 ). The rise of buyer-driven sustainability governance: emerging trends in the global coffee sector . Zentra Working Papers in Transnational Studies No. 73/2017 . Center for Transnational Studies (ZenTra) of the Universities of Bremen and Oldenburg .

Hoekman B. and Martin W. 2012 . Reducing distortions in international commodity markets: an agenda for multilateral cooperation . Policy Research Working Paper no. WPS 5928 . World Bank .

ICO . ( 2020 ). Coffee statistics . London : International Coffee Organization . http://www.ico.org . Accessed 15 September 2020.

Igami M. ( 2015 ). Oligopoly in international commodity markets: the case of coffee . Journal of Industrial Economics 63 ( 2 ): 225 – 248 . doi: 10.1111/joie.12076

Jarvis L. S. ( 2005 ). The rise and decline of rent-seeking activity in the Brazilian coffee sector: lessons from the imposition and removal of Coffee Export Quotas . UC Davis Agricultural and Resource Economics Working Paper No. 04-004 . USA : Department of Agricultural and Resource Economics, University of California .

Kinnucan H. W. and Forker O. D. ( 1987 ). Asymmetry in farm-retail price transmission for major dairy products . American Journal of Agricultural Economics 69 ( 2 ): 285 – 292 . doi: 10.2307/1242278

Koerner J. ( 2002 ). A good cup of Joe? Market power in the German and US coffee market . Working Paper EWP 0205 . Department of Food Economics and Consumption Studies, University of Kiel .

Koop G. , Pesaran M. H. and Potter S. M. ( 1996 ). Impulse response analysis in nonlinear multivariate models . Journal of Econometrics 74 ( 1 ): 119 – 147 . doi: 10.1016/0304-4076(95)01753-4

Krivonos E. ( 2004 ). The impact of coffee market reforms on producer prices and price transmission . Policy Research Working Paper Series: 3358 . World Bank .

Lass D. A. ( 2005 ). Asymmetric response of retail milk prices in the northeast revisited . Agribusiness 21 ( 4 ): 493 – 508 . doi: 10.1002/agr.20061

Lee J. and Gómez M. ( 2013 ). Impacts of the end of the coffee export quota system on international-to-retail price transmission . Journal of Agricultural Economics 64 ( 2 ): 343 – 362 . doi: 10.1111/j.1477-9552.2012.00372.x

Leibtag E. , Nakamura A. O. , Nakamura E. and Zerom D. ( 2007 ). Cost pass-through in the US coffee industry . USDA-ERS Economic Research Report . 38.

Levy D. L. ( 2008 ). Political contestation in global production networks . Academy of Management Review 33 ( 4 ): 943 – 963 . doi: 10.5465/amr.2008.34422006

McCorriston S. , Sexton R. J. and Sheldon I. M. ( 2004 ). Vertical market structure, commodity exports and trade reform . In: Paper presented at 7th Annual Conference on Global Economic Analysis . World Bank , June 2004 , Washington DC, USA .

McIntire J. and Varangis P. ( 1999 ). Reforming Cote d’Ivoire’s cocoa marketing and pricing system . Policy Research Working Paper No. 2081, March . World Bank , Washington DC .

McLaughlin E. W. ( 2006 ). U.S. retailing: more competitive than in Europe? Interact 1 : 22 – 25 .

Mehta A. and Chavas J.-P. ( 2008 ). Responding to the coffee crisis: what can we learn from price dynamics? Journal of Development Economics 85 ( 1–2 ): 282 – 311 . doi: 10.1016/j.jdeveco.2006.07.006

Meyer J. and Von Cramon-Taubadel S. ( 2004 ). Asymmetric price transmission: a survey . Journal of Agricultural Economics 55 ( 2004 ): 581 – 611 . doi: 10.1111/j.1477-9552.2004.tb00116.x

Muradian R. and Pelupessy W. ( 2005 ). Governing the coffee chain: the role of voluntary regulatory systems . World Development 33 ( 12 ): 2029 – 2044 . doi: 10.1016/j.worlddev.2005.06.007

Nakamura E. and Zerom D. ( 2010 ). Accounting for incomplete pass-through . The Review of Economic Studies 77 ( 3 ): 1192 – 1230 . doi: 10.1111/j.1467-937X.2009.589.x

Ng S. and Perron P. ( 2001 ). Lag length selection and the construction of unit root tests with good size and power . Econometrica 69 ( 6 ): 1519 – 1554 . doi: 10.1111/1468-0262.00256

Peltzman S. ( 2000 ). Prices rise faster than they fall . Journal of Political Economy 108 ( 3 ): 466 – 502 . doi: 10.1086/262126

Perron P. ( 1988 ). Trends and random walks in macroeconomic time series: further evidence from a new approach . Journal of Economic Dynamics and Control 12 : 297 – 332 . doi: 10.1016/0165-1889(88)90043-7

Perron P. and Yabu T. ( 2009a ). Estimating deterministic trends with an integrated or stationary noise component . Journal of Econometrics 151 ( 1 ): 56 – 69 . doi: 10.1016/j.jeconom.2009.03.011

Perron P. and Yabu T. ( 2009b ). Testing for shifts in trend with an integrated or stationary noise component . Journal of Business and Economic Statistics 27 ( 3 ): 369 – 396 . doi: 10.1198/jbes.2009.07268

Ponte S. ( 2002 ). The ‘Latte Revolution’? Regulation, markets and consumption in the global coffee chain . World Development 30 ( 7 ): 1099 – 1122 . doi: 10.1016/S0305-750X(02)00032-3

Raffaeli M. ( 1995 ). Rise and Demise of Commodity Agreements: An Investigation into the Breakdown of International Commodity Agreements . Cambridge : Woodhead .

Russell B. , Mohan S. and Banerjee A. ( 2012 ). Coffee market liberalisation and the implications for producers in Brazil, Guatemala and India . The World Bank Economic Review 26 ( 3 ): 514 – 539 . doi: 10.1093/wber/lhr055

Schwert G. W. ( 1989 ). Tests for unit roots: a Monte Carlo investigation . Journal of Business and Economic Statistics 7 : 147 – 159 .

Serra T. and Goodwin B. K. ( 2003 ). Transmission and asymmetric adjustment in the Spanish dairy sector . Applied Economics 35 ( 18 ): 1889 – 1899 . doi: 10.1080/00036840310001628774

Sexton R. J. ( 2013 ). Market power, misconceptions, and modern agricultural markets . American Journal of Agricultural Economics 95 ( 2 ): 209 – 219 . doi: 10.1093/ajae/aas102

Shepherd B. ( 2005 ). Market Power in International Commodity Processing Chains: Preliminary Results from the Coffee Market . Paris : Groupe d’Economie Mondiale de Sciences Po .

Sobreira N. and Nunes L. ( 2016 ). Tests for multiple breaks in the trend with stationary or integrated shocks . Oxford Bulletin of Economics and Statistics 78 : 394 – 411 . doi: 10.1111/obes.12116

Statistica . ( 2016 ). Coffee retail market share worldwide in 2014, by company . Statistica . https://www.statista.com/statistics/323254/global-retail-coffee-market-share-by-company/ . Accessed 20 November 2018.

Subervie J. ( 2011 ). Producer price adjustment to commodity price shocks: an application of threshold cointegration . Economic Modelling 28 : 2239 – 2246 . doi: 10.1016/j.econmod.2011.06.010

Talbot J. M. ( 1997 ). Where does your coffee dollar go? The division of income and surplus along the coffee commodity chain . Studies in Comparative International Development 32 ( 1 ): 56 – 91 . doi: 10.1007/BF02696306

Tsay R. S. ( 1989 ). Testing and modeling threshold autoregressive processes . Journal of the American Statistical Association 84 ( 405 ): 231 – 240 . doi: 10.1080/01621459.1989.10478760

US Bureau of Labor Statistics . ( 2020 ). Consumer price index – average price data; Series Id: APU0000717311 . Washington DC : United States Department of Labor . https://data.bls.gov . Accessed 15 September 2020.

Ward R. W. ( 1982 ). Asymmetry in retail, whole, and shipping point pricing for fresh vegetables . American Journal of Agricultural Economics 64 ( 2 ): 205 – 212 . doi: 10.2307/1241124

World Vision . ( 2014 ). Impact and resilience: World Vision International Annual Review 2014 . UK : World Vision . https://www.wvi.org/international/publication/world-vision-international-annual-review-2014 . Accessed 5 December 2018.

Email alerts

Citing articles via.

• Recommend to your Library

Affiliations

• Online ISSN 1464-3618
• Print ISSN 0165-1587
• Copyright © 2024 Oxford University Press and the European Agricultural and Applied Economics Publications Foundation 2016
• About Oxford Academic
• Publish journals with us
• University press partners
• What we publish
• New features  
• Open access
• Institutional account management
• Rights and permissions
• Get help with access
• Accessibility
• Advertising
• Media enquiries
• Oxford University Press
• Oxford Languages
• University of Oxford

Oxford University Press is a department of the University of Oxford. It furthers the University's objective of excellence in research, scholarship, and education by publishing worldwide

• Copyright © 2024 Oxford University Press
• Cookie settings
• Cookie policy
• Privacy policy
• Legal notice

This Feature Is Available To Subscribers Only

Sign In or Create an Account

This PDF is available to Subscribers Only

For full access to this pdf, sign in to an existing account, or purchase an annual subscription.

International Conference on Modelling, Computation and Optimization in Information Systems and Management Sciences

MCO 2021: Modelling, Computation and Optimization in Information Systems and Management Sciences pp 247–258 Cite as

The Effect of Machine Learning Demand Forecasting on Supply Chain Performance - The Case Study of Coffee in Vietnam

• Thi Thuy Hanh Nguyen   ORCID: orcid.org/0000-0003-4582-8795 12 ,
• Abdelghani Bekrar 12 ,
• Thi Muoi Le 13 &
• Mourad Abed 12  
• Conference paper
• First Online: 08 December 2021

508 Accesses

Part of the book series: Lecture Notes in Networks and Systems ((LNNS,volume 363))

Demand forecasting methods are one of the variables that have a considerable influence on supply chain performance. However, there is a lack of empirical proof on the magnitude of savings as observable supply chain performance results. In the literature, most scholars have paid more attention to non-financial performance while ignoring financial performance. This study compared the effect of two famous forecasting models on the operational and financial performance of the supply chain. ARIMAX (Auto-Regressive Integrated Moving Average with exogenous factors as a traditional model) and LSTM (Long Short-Term Memory as machine learning model) have been chosen. These two models were tested on Vietnamese coffee demand data. The results demonstrated that traditional and machine learning forecasting methods have different impacts on supply chain performance. The machine learning forecasting method outperformed the traditional method regarding operational and financial metrics. Three relevant operating and one financial metrics are selected, such as bullwhip effect (BWE), net stock amplification (NSAmp), and transportation cost (TC), and inventory turn (IT), respectively.

This is a preview of subscription content, log in via an institution .

Buying options

• Available as PDF
• Read on any device
• Instant download
• Own it forever
• Available as EPUB and PDF
• Compact, lightweight edition
• Dispatched in 3 to 5 business days
• Free shipping worldwide - see info

Tax calculation will be finalised at checkout

Purchases are for personal use only

Perera, H.N., Hurley, J., Fahimnia, B., Reisi, M.: The human factor in supply chain forecasting: a systematic review. Eur. J. Oper. Res. 274 (2), 574–600 (2019)

Article   MathSciNet   Google Scholar  

Benkachcha, S., Benhra, J., El Hassani, H.: Demand forecasting in supply chain: comparing multiple linear regression and artificial neural networks approaches. Int. Rev. Model. Simul. 7 (2), 279–286 (2014)

Google Scholar  

Moon, M.A., Mentzer, J.T., Smith, C.D.: Conducting a sales forecasting audit. Int. J. Forecast. 19 (1), 5–25 (2003)

Article   Google Scholar  

Zotteri, G., Kalchschmidt, M.: Forecasting practices: empirical evidence and a framework for research. Int. J. Prod. Econ. 108 (12), 84–99 (2007)

Abolghasemi, M., Beh, E., Tarr, G., Gerlach, R.: Demand forecasting in supply chain: the impact of demand volatility in the presence of promotion. Comput. Ind. Eng. 142 , 1–12 (2020)

Moon, M.A., Mentzer, J.T., Smith, C.D., Garver, M.S.: Seven keys to better forecasting. Bus. Horiz. 41 (5), 44–52 (1998)

Danese, P., Kalchschmidt, M.: The role of the forecasting process in improving forecast accuracy and operational performance. Int. J. Prod. Econ. 131 (1), 204–214 (2011)

George, J., Madhusudanan Pillai, V.: A study of factors affecting supply chain performance. J. Phys: Conf. Ser. 1355 (1), 1–8 (2019)

Jaipuria, S., Mahapatra, S.S.: An improved demand forecasting method to reduce bullwhip effect in supply chains. Expert Syst. Appl. 41 (5), 2395–2408 (2014)

Feizabadi, J.: Machine learning demand forecasting and supply chain performance. International Journal of Logistics Research and Applications, pp. 1–24 (2020)

Kantasa-ard, A., Nouiri, M., Bekrar, A., Ait el cadi, A., Sallez, Y.: Machine learning for demand forecasting in the physical internet: a case study of agricultural products in Thailand. International Journal of Production Research, pp. 1–25 (2020)

Kilimci, Z.H., et al.: An improved demand forecasting model using deep learning approach and proposed decision integration strategy for supply chain. Complexity, pp. 1–16 (2019)

Aburto, L., Weber, R.: Improved supply chain management based on hybrid demand forecasts. Appl. Soft Comput. 7 (1), 136–144 (2007)

Soy Temür, A., Akgün, M., Temür, G.: Predicting housing sales in turkey using arima, lstm and hybrid models. J. Bus. Econ. Manag. 20 (5), 920–938 (2019)

Khashei, M., Bijari, M.: A novel hybridization of artificial neural networks and ARIMA models for time series forecasting. Appl. Soft Comput. J. 11 (2), 2664–2675 (2011)

Naveena, K., Singh, S., Rathod, S., Singh, A.: Hybrid ARIMA-ANN modelling for forecasting the price of robusta coffee in India. Int. J. Curr. Microbiol. App. Sci. 6 (7), 1721–1726 (2017)

Mentzer, J.T., Kahn, K.B.: Forecasting technique familiarity, satisfaction, usage, and application. J. Forecast. 14 (5), 465–476 (1995)

Mccarthy, T.M., Davis, D.F., Golicic, S.L., Mentzer, J.T.: The evolution of sales forecasting management: a 20-year longitudinal study of forecasting practices. J. Forecast. 25 (5), 303–324 (2006)

Chen, I.J., Paulraj, A.: Understanding supply chain management: critical research and a theoretical framework. Int. J. Prod. Res. 42 (1), 131–163 (2004)

Siami-Namini, S., Tavakoli, N., Siami Namin, A.: A comparison of ARIMA and LSTM in forecasting time series. In: Proceedings - 17th IEEE International Conference on Machine Learning and Applications, pp. 1394–1401. IEEE (2019)

Choi, H.K.: Stock price correlation coefficient prediction with ARIMA-LSTM hybrid model. Seoul, Korea: Korea University (2018). https://arxiv.org/pdf/1808.01560v5.pdf . Accessed 03 Mar 2021

Abbasimehr, H., Shabani, M., Yousefi, M.: An optimized model using LSTM network for demand forecasting. Comput. Ind. Eng. 143 , 1–13 (2020)

Dave, E., Leonardo, A., Jeanice, M., Hanafiah, N.: Science direct forecasting indonesia exports using a hybrid model ARIMA- LSTM. Procedia Comput. Sci. 179 (2020), 480–487 (2021)

Vietnam Logistics Report 2020. https://hdgroup.vn/wp-content/uploads/2020/12/Ba%CC%81o-ca%CC%81o-Logistics-Vie%CC%A3%CC%82t-Nam-2020.pdf . Accessed 02 Feb 2021

Coffee Market Report. https://ico.org/news/cmr-1020-e.pdf . Accessed 02 Jan 2021

Kwak, J.K.: Analysis of inventory turnover as a performance measure in manufacturing industry. Processes 7 (10), 1–11 (2019)

Luong, H.T.: Measure of bullwhip effect in supply chains with autoregressive demand process. Eur. J. Oper. Res. 180 (3), 1086–1097 (2007)

Luong, H.T., Phien, N.H.: Measure of bullwhip effect in supply chains: the case of high order autoregressive demand process. Eur. J. Oper. Res. 183 (1), 197–209 (2007)

Boute, R.N., Lambrecht, M.R.: Exploring the bullwhip effect by means of spreadsheet simulation. INFORMS Trans. Educ. 10 (1), 1–9 (2009)

Luong, Q.V., Tauer, L.W.: A real options analysis of coffee planting in Vietnam. Agric. Econ. 35 (1), 49–57 (2006)

Download references

Author information

Authors and affiliations.

LAMIH, Université Polytechnique Hauts-de-France, Valenciennes, France

Thi Thuy Hanh Nguyen, Abdelghani Bekrar & Mourad Abed

CRISS, Université Polytechnique Hauts-de-France, Valenciennes, France

Thi Muoi Le

You can also search for this author in PubMed   Google Scholar

Corresponding author

Correspondence to Thi Thuy Hanh Nguyen .

Editor information

Editors and affiliations.

Computer science and Applications Department, LGIPM, University of Lorraine, Metz Cedex, France

Hoai An Le Thi

Laboratory of Mathematics, National Institute for Applied Sciences - Rouen, Saint-Etienne-du-Rouvray Cedex, France

Tao Pham Dinh

Hoai Minh Le

Rights and permissions

Reprints and permissions

Copyright information

© 2022 The Author(s), under exclusive license to Springer Nature Switzerland AG

About this paper

Cite this paper.

Nguyen, T.T.H., Bekrar, A., Le, T.M., Abed, M. (2022). The Effect of Machine Learning Demand Forecasting on Supply Chain Performance - The Case Study of Coffee in Vietnam. In: Le Thi, H.A., Pham Dinh, T., Le, H.M. (eds) Modelling, Computation and Optimization in Information Systems and Management Sciences. MCO 2021. Lecture Notes in Networks and Systems, vol 363. Springer, Cham. https://doi.org/10.1007/978-3-030-92666-3_21

Download citation

DOI : https://doi.org/10.1007/978-3-030-92666-3_21

Published : 08 December 2021

Publisher Name : Springer, Cham

Print ISBN : 978-3-030-92665-6

Online ISBN : 978-3-030-92666-3

eBook Packages : Intelligent Technologies and Robotics Intelligent Technologies and Robotics (R0)

Share this paper

Anyone you share the following link with will be able to read this content:

Sorry, a shareable link is not currently available for this article.

Provided by the Springer Nature SharedIt content-sharing initiative

• Publish with us

Policies and ethics

• Find a journal
• Track your research

Between Drinks

Demand for instant coffee on the rise, study shows, younger generation influences global demand for instant coffee.

Being an avid coffee enthusiast, I can say that I am one of those people who firmly believes that drinking at least a cup or two in the morning is crucial for starting my day. Having started my relationship with coffee as a teenager in high school, today, I still turn to brewing myself a fresh pot every morning, despite all of the more recent options offered, like ready-to-drink (RTD) options, cold brews, etc. 

Although I’ve never really turned to instant coffee for my coffee fix, interestingly, today’s younger generation is not only turning to options like cold brews, but also instant coffee to get their fix, influencing the global instant coffee market, according to a recent Fact.MR study.

“Growing urbanization and changing lifestyles are some pivotal factors influencing the demand for instant coffee, especially among youngsters, around the world. These new customers find instant coffee more attractive,” Fact.MR states. “Modernization is another significant factor, which is fueling the rising demand for instant coffee.”

Further, in its global market study, “Instant Coffee Market To Accumulate $69.2 Billion By 2033,” Fact.MR notes that the global instant coffee market is forecasted to expand at a compound annual growth rate (CAGR) of 5% during the forecast period 2023 to 2033.

Moreover, noting the increasing demand for in-home coffee consumption in developing economies, Fact.MR points to the convenience in preparation as a factor causing consumers to adopt instant coffee. 

“Coffee is consumed for reducing fat, increasing microbiome diversity, providing essential nutrients, and improving energy levels. Many people are addicted to coffee and need it to stay awake at work,” it states. “Thus, owing to the hectic work culture, organizations and places of work ensure easy access to instant coffee facilities for employees. Consumers also prefer to carry their coffee with them due to the rising trend of on-the-go food.” 

Some of the key takeaways from the Fact.MR study, are as follows:

• The global instant coffee market is valued at $42.5 billion in 2023.
• Demand for instant coffee is estimated to reach a market valuation of $69.2 billion by the end of 2033.
• Sales of instant coffee are projected to increase at a CAGR of 5% from 2023 to 2033. 
• Demand for instant coffee in Germany is forecasted to increase at a CAGR of 3.6%. 
• Sales of instant coffee in Canada are projected to expand at a CAGR of 4% through 2033. 

Highlighting some of the prominent manufacturers of instant coffee, such as Starbucks Corp., The J.M. Smucker Co. and The Kraft Heinz Co., Fact.MR also notes that prominent market players now are involved in expanding their footprints overseas, introducing new products to enhance sales of instant coffee during the coming 10 years. 

With so much activity on the horizon, maybe I’ll trade in my coffee pot for some instant packets.

Share This Story

Lauren Sabetta, managing editor for Beverage Industry , writes for the magazine’s print and online components. She earned her Bachelor of Science in Communication, Journalism from Appalachian State University.

Restricted Content

You must have JavaScript enabled to enjoy a limited number of articles over the next 30 days.

Related Articles

Study shows coffee preferences shine light on consumer demographics

Consumer demand for premium spirits soars in on-premise, CGA analysis reveals

Ready-to-drink, low/no alcohol categories on the rise

Get our new emagazine delivered to your inbox every month., stay in the know on the latest beverage industry marketplace trends..

Copyright ©2024. All Rights Reserved BNP Media.

Design, CMS, Hosting & Web Development :: ePublishing

• Español – América Latina
• Português – Brasil

Costa Coffee: Helping consumers find their closest cup

About Costa Coffee

Founded in London in 1971, Costa Coffee is now the UK’s largest coffee store chain and the world’s second largest, with 4,000+ stores in 31 countries. Its latest digital innovation is ordering via the Costa Coffee and delivery partner apps.

Tell us your challenge. We're here to help.

About Snowdrop Solutions

A Google Cloud Premier Partner, Snowdrop Solutions delivers location intelligence for big brands.

Costa Coffee navigates consumers to their most convenient eat-in, takeaway, drive-thru, partner stores and delivery platforms.

• Ensures click-and-collect requests are made to the correct Costa Coffee store, saving time and costs by preventing mistakes
• Provides accurate information on store opening times and locations, cutting down on waiting time for consumers
• Helps store team members to prepare for demand by combining navigation tools with online ordering

Costa app helps consumers find their nearest Costa location

When you can count the number of stores you have on one hand, it’s easy to list their addresses on a website and show your consumers how to reach them. With more than 4,000 coffee stores spread across 31 countries, Costa Coffee recognized it needed a sophisticated location intelligence solution to intuitively direct its consumers to their nearest Costa Coffee and highlight which services are available where and when.

Since it was founded in 1971, technology has played a crucial role in taking Costa Coffee from a single roastery on London’s Fenchurch Street to a brand found on highstreets, grocery stores, and store cupboards in homes across the world. In 2020, when the coffee store brand wanted to ensure that its consumers were able to continue enjoying their cup of Costa Coffee amidst COVID-19 travel restrictions, the company once again turned to technology to bring new ideas and services to life.

The company wanted to make consumers aware that Costa Express machines remained operating in petrol stations and forecourts up and down the country, even when Costa Coffee stores closed. It also wanted consumers to know that their nearest store offered the facility to order and pay online (and complete a minimal contact collection), and that local stores offered delivery services via partners. Costa Coffee recognized that accurate location data was required to keep consumers informed and safe during an unusual year.

When Covid hit, the company focussed on rolling out more ways to order to more locations. Having used Google Maps Platform in its website and mobile app for years, it knew it could be the solution.

Making it as easy as possible for consumers to get what they need

When Costa Coffee launched its mobile app in 2017, it wanted to provide location intelligence features that would make its consumers’ journeys quicker and easier. When a customer imputed their location in the app, for example, it would pop up the address of their nearest store and its opening hours. But Costa Coffee wanted more. How about an interactive element that gave consumers the option to order online and show up when their order was ready for collection? These ideas would decrease waiting times, which in 2020, due to COVID-19 safety measures, was a matter of security and precaution, in addition to convenience.

To bring these ideas to life, Costa Coffee turned to its long-standing Google Cloud Premier Partner, Snowdrop Solutions . "We decided to leverage more Google Maps Platform capabilities because our developers find it easy to work with, and it supports all the functionalities we need, from routes that assist consumers to our stores to services such as click and collect," explains Gordon Lucas, Global Head of Digital Engineering at Costa Coffee.

"We decided to leverage more Google Maps Platform capabilities because our developers find it easy to work with, and it supports all the functionalities we need, from routes that assist consumers to our stores to services such as click and collect."

With a small web development team based in its UK headquarters, it was also important for Costa Coffee to be able to build in functionality that can be seamlessly rolled out across its multiple territories around the world. To that end, Costa Coffee embedded Google Maps Platform within its mobile app and international website, using the Geocoding API to create visual place markers illustrating Costa Coffee locations on its maps.

To save users time, they also leveraged Place Autocomplete to predict addresses and check their accuracy when consumers are making an online order. "The Places API powers our click-and-serve feature, where baristas bring orders to consumers’ cars. Now consumers can see how long their drive to the nearest service is," explains Lucas.

Relying on accurate location data to deliver services safely during COVID-19

When the pandemic and lockdowns forced stores to adapt in 2020, the new Google Maps Platform solutions employed by Costa Coffee online became even more invaluable for the business and its consumers. "Our Store Locator became crucial during lockdown, when opening times changed and restrictions looked different depending on where stores are based, each complying with their local lockdown measures," explains Lucas. "Having our Google Maps Platform solutions in place then meant that we were always able to show our consumers which stores or drive-thrus near them were open and safely serving great coffee during a time when, more than ever, they sought comfort in their favourite Costa Coffee beverage. That’s why our Store Locator quickly became the most visited page on our website."

"The Places API powers our click-and-serve feature, where baristas bring orders to consumers’ cars. Now consumers can see how long their drive to the nearest service is."

Although not part of the development specification, Google Maps Platform quickly became a vital part of Costa Coffee’s internal COVID-19 hygiene initiative too. While hundreds of its stores were forced to temporarily close, Costa Coffee’s 9,000 Express machines in grocery stores and petrol stations were working overtime when they were some of the few places selling cups of coffee. "Our mobile ordering system lets you find an Express machine, scan your order request, and pick up your drink without touching the screen," says Lucas.

The hands-free order-and-collect system was adapted for standalone stores too, allowing users to find and select a store, place an order, avoid queues, and pay and collect products, speeding up the transaction process and minimizing the risk of transmission between consumers and baristas.

"At the end of the first lockdown, for example, we saw huge consumption numbers on our mapping technology as people were trying to find their nearest Costa Coffee and eager to enjoy the experience of buying a coffee or snack, something that was normal pre-lockdown but had by then become a ‘treat’," says Lucas.

Location intelligence is also used to improve the experience of consumers from when they first order a cappuccino and cake, to when they arrive at a Costa Coffee store to enjoy it.

"Google Maps Platform is now a key part of our digital operations. It helps us to make our communications more relevant to consumers and boost sales by ensuring that consumers always know where and when to get their cup of Costa Coffee."

Bringing consumers around the world closer to Costa Coffee

Costa Coffee has not changed its Mocha Italia Signature Blend in 50 years, but it is continually using the latest technology to improve its distribution, accessibility, and customer service. In its latest development, it is trialling robotic coffee bars in the US, and Google Maps Platform will be integral to raising users' awareness of their whereabouts too.

"Google Maps Platform is now a key part of our digital operations. It helps us to make our communications more relevant to consumers and boost sales by ensuring that consumers always know where and when to get their cup of Costa Coffee," says Lucas. "That’s why Costa Coffee continues to innovate and experiment with Google Maps Platform to ensure that wherever our consumers are in the world, they are never far from a Costa Coffee Americano or classic Latte."

Recent Announcements

Biden-harris administration opens application period to support hydropower facilities across the country.

Bipartisan Infrastructure Law’s Hydroelectric Incentives Invest in Maintaining and Enhancing Hydroelectric Facilities to Ensure Generators Continue to Provide Clean Power

On March 14, 2024, the Biden-Harris Administration, through the U.S. Department of Energy (DOE), announced the opening of the 2024 application period for the Hydroelectric Production Incentive, which supports the modernization and expansion of hydroelectric power throughout America through the  Hydroelectric Incentives Program . Supported by the Bipartisan Infrastructure Law passed under President Biden, the Hydroelectric Incentives Program is focused on maintaining and enhancing hydroelectric facilities to ensure generators continue to provide clean power, while improving dam safety and reducing environmental impacts.

Hydroelectric Production Incentive applications for electricity generated and sold in calendar year 2023 are due by 5:00 p.m. ET on April 23, 2024 .

Open the video to learn more about the impact of Section 242: the Hydroelectric Production Incentive Program.

Video courtesy of the U.S. Department of Energy, Grid Deployment Office.

Learn more about the Hydroelectric Production Incentives application period to support hydropower facilities across the country .

DOE Reports Chart Path for East Coast Offshore Wind to Support a Reliable, Affordable Electricity System

Study Finds Interregional Offshore Transmission Network Would Support Long-Term Growth of Atlantic Offshore Wind, Action Plan Identifies Immediate Steps and Extended Efforts

On March 21 , DOE released findings from the Atlantic Offshore Wind Transmission Action Plan and Study. Offshore wind is projected to be a key part of a low-carbon future for East Coast states and this report is the most thorough analysis to date of options to bring Atlantic offshore wind energy to American communities. While immediate projects will connect individually to the onshore grid, the study finds that after 2030, strategically linking some offshore wind energy projects via offshore transmission networks will help lower electricity production costs, enhance U.S. grid reliability, and reduce dependence on fossil fuels, while ensuring disruptions to oceanic ecosystems are minimal.

The Atlantic Offshore Wind Transmission Study is a two-year study evaluating transmission options to support offshore wind energy deployment along the Atlantic Coast of the United States.

The study informed the  Atlantic Offshore Wind Transmission Action Plan which outlines immediate actions needed to connect the first generation of Atlantic offshore wind projects to the electric grid, as well as longer-term efforts to increase transmission over the next several decades.

Learn more about how DOE is working to bring Atlantic offshore wind energy to American communities .

DOE Invests More Than $18 Million for Clean Energy-Related Economic Development in Low-Income Communities

Technical Assistance Will Support 30 Projects Across US, Including Energy Resilience Projects in Puerto Rico

DOE announced on March 22 the second cohort of communities selected as part of the  Communities Local Energy Action Program (Communities LEAP) , a unique technical assistance initiative designed to help disadvantaged communities and those with historical ties to fossil fuel industries take direct control of their clean energy future. The 30 selected communities will receive a total of $18 million worth of technical assistance to create tailored community-wide action plans that reduce local air pollution, increase energy resilience, lower utility costs and energy burdens, and create well-paying jobs.

The 30 selected applicants will work with DOE and its network of technical assistance providers, as well as local coalition partners including local and tribal governments; community-based organizations; utilities; and environmental justice, economic development, and equitable investment organizations to develop roadmaps that utilize clean energy as a tool for economic development.

DOE selected projects for Communities LEAP’s Puerto Rico Community Resilience category located in the following municipalities: 

• Adjuntas, Jayuya, Lares, Maricao, & Utuado, Puerto Rico
• Yauco (Quebradas Neighborhood), Puerto Rico

Learn more about the second cohort of LEAP communities .

News and Updates

Doe publishes pr100 study results and possible pathways to puerto rico’s 100% renewable energy future.

The full PR100 Study is now available on PR100.gov.

DOE and the Federal Emergency Management Agency (FEMA) released the final Puerto Rico Grid Resilience and Transitions to 100% Renewable Energy Study (PR100) . The two-year study concludes that Puerto Rico can successfully meet its projected electricity needs with 100% renewable energy by 2050. PR100 outlines possible pathways to achieving clean energy success, including grid stabilization measures and the deployment of distributed energy resources.

Launched in February 2022 with funding from FEMA, PR100 deployed the world-class expertise of DOE's National Labs to engage diverse local stakeholders regarding their priorities for their energy future, gather and generate valuable data sets, and create advanced models to project scenarios for achieving 100% renewable energy. PR100 also includes an Implementation Roadmap, providing decision-makers in Puerto Rico with specific actions that can be taken to stabilize the grid, improve resilience, and achieve local renewable energy goals. The two-year PR100 Study relied on extensive stakeholder input to ensure the results and Implementation Roadmap reflect local priorities for Puerto Rico’s energy transition.

View the full study and interactive data visualizations at PR100.gov .

Learn more about efforts supporting Puerto Rico's goal of 100% renewable energy by 2050 .

In Case You Missed It: GRIP Program Training Webinars Now Available On-Demand

Recordings of three critical training webinars from the  Grid Resilience and Innovation Partnerships (GRIP) Program are now available on demand. Available for second round funding applicants and those who may choose to apply for funding in the future, these webinars provide important information about eligibility and legal requirements, supply chain challenges, application components, and next steps.

• The  GRIP Program Community Benefits Plan (CBP) Training Webinar shared insights and lessons learned from first round GRIP funding applications, examples of strong Community Benefits Plans, an overview of the Community Benefits Plan requirements as outlined in the FOA, and best practices. The video recording and presentation are now available  online .
• The  Concept Paper Trends & Application Updates Webinar provided updates on the second-round funding opportunity including concept paper submission trends and next steps. The video recording and presentation are now available   online .
• The  Critical Compliance Information Webinar provided information about navigating supply chain challenges, non-domestic content, and other compliance requirements mandated by the Davis-Bacon Act (DBA) and the Build America Buy America (BABA) Act. The video recording and presentation are now available  online .

Learn more about the GRIP Program, and access webinar materials and other informational resources .

DOE's Northern Lights - Nick Palso

The Arctic Energy Office is hosting a series of interviews to highlight the lives and achievements of some of DOE’s outstanding employees with Arctic and Alaska connections. Meet Nick Palso. He went to college in Fairbanks and spent two summers doing his doctoral dissertation research in Alaska. Nick is now a senior project manager heading up a hydroelectric incentives program at GDO.

In this profile, you will learn about Nick’s experiences in the arctic, his role and interests, and more .

Open Opportunities

Upcoming events, recent meeting materials.

• Open access
• Published: 15 April 2024

Life expectancy, long-term care demand and dynamic financing mechanism simulation: an empirical study of Zhejiang Pilot, China

• Xueying Xu 1 ,
• Yichao Li 2 &
• Hong Mi 2  

BMC Health Services Research volume  24 , Article number:  469 ( 2024 ) Cite this article

Metrics details

China has piloted Long-Term Care Insurance (LTCI) to address increasing care demand. However, many cities neglected adjusting LTCI premiums since the pilot, risking the long-term sustainability of LTCI. Therefore, using Zhejiang Province as a case, this study simulated mortality-adjusted long-term care demand and the balance of LTCI funds through dynamic financing mechanism under diverse life expectancy and disability scenarios.

Three-parameter log-quadratic model was used to estimate the mortality from 1990 to 2020. Mortality with predicted interval from 2020 to 2080 was projected by Lee-Carter method extended with rotation. Cohort-component projection model was used to simulate the number of older population with different degrees of disability. Disability data of the older people is sourced from China Health and Retirement Longitudinal Study 2018. The balance of LTCI fund was simulated by dynamic financing actuarial model.

Life expectancy of Zhejiang for male (female) is from 80.46 (84.66) years in 2020 to 89.39 [86.61, 91.74] (91.24 [88.90, 93.25]) years in 2080. The number of long-term care demand with severe disability in Zhejiang demonstrates an increasing trend from 285 [276, 295] thousand in 2023 to 1027 [634, 1657] thousand in 2080 under predicted mean of life expectancy. LTCI fund in Zhejiang will become accumulated surplus from 2024 to 2080 when annual premium growth rate is 5.25% [4.20%, 6.25%] under various disability scenarios, which is much higher than the annual growth of unit cost of long-term care services (2.25%). The accumulated balance of LTCI fund is sensitive with life expectancy.

Conclusions

Dynamic growth of LTCI premium is essential in dealing with current deficit around 2050 and realizing Zhejiang’s LTCI sustainability in the long-run. The importance of dynamic monitoring disability and mortality information is emphasized to respond immediately to the increase of premiums. LTCI should strike a balance between expanding coverage and controlling financing scale. This study provides implications for developing countries to establish or pilot LTCI schemes.

Peer Review reports

The lack of sufficient long-term care (LTC) for older individuals has become a pressing concern in both developed and developing countries with global population aging and increased longevity [ 1 ]. Although healthy life expectancy generally increased over last decades [ 2 ], the episode of disability in older people could have catastrophic impact on their household welfare [ 3 ]. Several developed countries, such as the Netherlands, Germany, and Japan, have established social long-term care insurance (LTCI) to address LTC demands of households with disabled older individuals. This approach proves more efficient in pooling disability risks than private LTCI [ 4 , 5 ]. Nonetheless, many developed countries had to reform their LTCI systems to deal with increasing aging population with LTC demands, often by raising premiums. Even though, these adjustments usually had time lags which affected the long-term sustainability of LTCI schemes. However, establishing social LTCI in developing countries proves more challenging than in developed countries because the lower income of residents restricts the financing capacity of LTCI. In addition, the lack of high-quality death registration and health survey data hinders optimizing LTCI systems design according to changing LTC demands, particularly in developing countries or small areas [ 6 ].

Massive evidence shows that there will be a steady and slow increase in life expectancy [ 7 , 8 , 9 ]. Evidence from developed countries shows that the long-term care needs increasing rapidly because of the increasing life expectancy [ 10 , 11 ]. The trend of the gap between life expectancy and healthy life expectancy is still inconclusive [ 12 ], which also affects the identification of LTC needs [ 13 ]. There is still mixed conclusion of disability and LTC demands trend in the future based on the three different assumptions of health transitions [ 14 , 15 , 16 ]. Whereas, there is less evidence regarding the assessment of LTC needs under different mortality scenarios. Zeng, et al. [ 17 ] calculated long-term care needs under different life expectancy scenarios, but the setting of life expectancy was relatively subjective. Besides, many studies in country-level controlled the impact of underreported mortality on the LTCI system by using modified mortality data [ 18 , 19 ], but few studies in the provincial level took that into consideration.

Most countries such as Germany and the Netherlands adopt a fixed percentage of income model to collect social LTCI premiums from individuals [ 20 ], and a few countries such as Singapore adopt a fixed amount premium model [ 21 ]. The premium of Germany LTCI has been 3.05% of gross income or 3.40% if individuals aged 23 and above without children since 2020 [ 22 ]. The Netherlands also has a tax-funded LTCI with the compulsory contribution of 9.65% of taxable income since 2017 [ 20 ]. In Singapore, fixed amount premium of LTCI is determined by the age of starting contribution and sex. The premium for a 30-year-old male (female) is around 200 (250) Singapore Dollars in 2020 [ 21 ], with an increase of 2% per year from 2020 to 2025 [ 23 ]. Financing parameters from both models should be adjusted regularly to ensure sustainability [ 24 , 25 ]. In China, both models are adopted in different LTCI pilot areas [ 26 ], but the areas that adopt the fixed amount of premium have not increased the premium level since the pilot, which affects long-term sustainability.

OECD countries will face high pressure of LTCI financing because of increasing average public LTC expenditures to 2.3% of GDP in 2040 for the future financing level of LTCI [ 27 ]. Therefore, an adjustment factor is suggested incorporated to simulate LTCI fund to reduce future financing pressure [ 22 ], but a higher short-term financing will bring greater resistance to reforms. Most simulation studies on China’s LTCI, based on fixed percentage of income model, demonstrated that LTCI financing will increase rapidly based on different disability scenarios [ 28 , 29 , 30 , 31 ]. Some studies also simulated LTCI financing based on fixed amount of premium model [ 32 , 33 ], but they did not consider its variation under different mortality scenarios. Only one study modified the mortality in a pilot city by using national mortality data when simulating the dynamic financing burden [ 34 ]. However, it only simulated to 2040 which did not cover plateau period of China’s aging.

China, as a developing country, pioneered social LTCI schemes in 2016. Local governments were granted significant autonomy, resulting in fragmented LTCI structures due to regional disparities in the pilot cities [ 35 ]. Thus it has become crucial to ensure the sustainability of China’s LTCI pilot areas. Zhejiang Province stands as a representative case among these pilot areas and its five cities (Tonglu, Ningbo, Jiaxing, Yiwu and Whenzhou) have piloted LTCI since 2017. Zhejiang has standardized disability assessments, coverage groups, benefit levels, and financing amounts of LTCI in province-level by 2022 [ 36 ]. It faces rapid aging ahead with high life expectancy in China. Notably, Zhejiang, one of the areas with fixed amount of premium of LTCI in China, has never increased its fixed premium since the pilot’s inception [ 36 ]. This lack of financing adjustment coupled with inflationary pressures strains Zhejiang’s LTCI fund. Zhejiang has capacities to facilitate LTCI operations through modified financing mechanism as the demonstration zone for the Initiative of Common Prosperity in China. Therefore, it can serve as a practical model for other developing countries establishing LTCI schemes to evaluate life expectancy and LTC demand parameters and guide its LTCI financing.

In summary, massive studies predict the LTC needs in developed countries and China. However, most of the studies on LTCI financing in China pilots overlook the potential death underreporting in census and uncertainty of mortality in projection period, which may misestimate the future LTC needs and financing pressure. In addition, current studies on the sustainability of China’s LTCI rarely involve the dynamic financing adjustment of fixed amount of premium model, and most studies do not cover the plateau period of China’s aging in the future, which may underestimate the financing level to achieve sustainable LTCI. Therefore, drawing from the Zhejiang Province case in China, this study proposes a dynamic financing mechanism to achieve a balance between sustainability and efficiency in social LTCI schemes, utilizing a simulation model with limited mortality and disability information. Our aim is to offer insights for developing countries to establish or pilot LTCI schemes. Three research questions will be addressed:

What is the long-term trend of life expectancy in Zhejiang from 1990 to 2080?

What extent of LTC demand will be reached among older people in Zhejiang from 2023 to 2080, with aging process?

What level of LTCI dynamic financing standards will achieve an actuarial equilibrium of the LTCI fund in Zhejiang, with rising life expectancy and LTC demand?

Data sources

For demographic data, the age-specific mortality and the population number by gender are from population census of Zhejiang Province in 1990, 2000, 2010 and 2020. The population census, which has been conducted once every 10 years since 1990, is a complete account of the entire population, mortality and fertility by age and sex in each census year and has the province-level representativeness of Zhejiang. Child mortality data is from Chinese Center for Disease Control and Prevention (CDC) in 1990–2013 [ 37 ], and official annual data of Zhejiang reported u p to 2020 [ 38 ]. Chinese CDC sorted and estimated under-5 mortality rates in China before 2013 with county-level and province-level representativeness, including data in Zhejiang. Data on the prevalence rate of disability of the older people is sourced from China Health and Retirement Longitudinal Study (CHARLS) in 2018. CHARLS is a national representative survey which covers a wide range of topics related to the adults aged 45 and above, including demographic information and health status. The national prevalence rate of disability by age and sex from CHARLS is used as a proxy for Zhejiang referring to existing research, due to lack of latest representative disability data in Zhejiang [ 39 ]. Older people are defined as those aged 60 and above based the statistical standards from World Health Organization [ 40 ], whose age groups are covered by CHARLS. The benefit criteria and financing criteria data is from the LTCI official regulations of pilot cities in Zhejiang [ 41 , 42 , 43 , 44 , 45 ]. Healthcare Consumer Price Index (CPI) from 2010 to 2020 in Zhejiang is from National Bureau of Statistics of China, covering the socio-economic indicators at province-level [ 46 ]. The change rate of total fertility of China from 2020 to 2080 is from World Population Prospects 2022 which forecasted fertility in country-level around the world [ 47 ].

Estimation of mortality pattern with three-parameter model life table approach

Model life tables methods are widely used in simulation of mortality for their effectiveness and accessibility to overcome the limited mortality information in developing countries [ 48 , 49 ]. Two-parameters log-quadratic model considering the child and adult mortality overcomes the shortage of Coale-Demeny and UN model life tables, among those model life tables methods [ 50 ]. Three-parameter log-quadratic model is designed on this to calculate the life table considering extra old-age mortality parameter with an adjustment of intercept with real census information [ 51 ]. It is so-called developing countries mortality database (DCMD) model which was adopted in the World Population Prospects 2019 since the three-parameter log-quadratic model life table was initially used in those developing countries without the high-quality mortality data [ 52 ]. The basic function of DCMD model is showed below:

This study used adjusted DCMD model to estimate the mortality in Zhejiang from 1990 to 2020 to make it usable for open population conditions. Child mortality ( \({\,}_{5}{q_0}\) ) is the first parameter of DCMD model, and adult mortality ( \({\,}_{{45}}{q_{15}}\) ) is the second parameter to be compared with estimated adult mortality ( \({\,}_{{45}}{\hat {q}_{15}}\) ) from two-parameter log-quadratic model with adjustment factor \(k\) . Specifically, child mortality by gender in consecutive years is estimated by sex ratio of child mortality in China [ 53 ]. Adult mortality in census years is calculated from census life table directly as the register completeness of adults’ death is higher in China [ 53 ]. Moreover, we averaged old-age mortality estimated from two-parameter log-quadratic model and from survival model for midpoint of old-age mortality between censuses (1995, 2005 and 2015) [ 51 ]. We averaged old-age mortality from two-parameter log-quadratic model and from census life table calculations for census years (1990, 2000, 2010 and 2020). The adjusted DCMD model was constructed on the incorporated old-age mortality. After that, the cubic hermite polynomial interpolation approach (pchip package in R) was adopted to estimate adult and old-age mortality from 1990 to 2020 [ 54 ]. The life table for consecutive years was estimated with DCMD model.

After that, Lee-Carter method extended with rotation (LC_ER) (mortcast package in R) was used to forecast the mortality up to 2080 [ 55 ], which provides critical death parameters to assess the LTCI demands in our case area. Since in low mortality countries, mortality decline is decelerating at younger ages and accelerating at older ages [ 56 ], the static assumption of mortality decline of traditional Lee-Carter model would be anomalous in long-term projection. LC_ER is a time-varying Lee-Carter model considering the changes of mortality decline between different age groups when modeling, which was widely recognized and adopted by World Population Prospects 2022 [ 57 , 58 ]. Therefore, potential LTCI demands change caused by changes in old-age mortality decline in long-term projections could be captured by LC_ER. The predicted mean of life expectancy would be set as the medium life expectancy scenario, and the lower and upper 95% predicted interval would be set as the low and high life expectancy scenarios.

Number of severe disabled older adults

LTCI beneficiaries refer to the severe disabled population according to the rules of LTCI in Zhejiang [ 36 ]. The study used the cohort-component projection (CCP) method to forecast the number of older population of Zhejiang from 2020 to 2080 [ 59 ]. The number of age-specific population by sex from Zhejiang population census 2020 was used as the base population of CCP model. Furthermore, the age-specific prevalence rate of disability from CHARLS 2018 was calculated. After that, the number of severe disabled older adults as the LTCI beneficiaries was calculated by multiplying age-specific older population and prevalence rate of disability. The basic project method is as follows:

\({\,}_{{x+1}}P_{x}^{{t+1}}\) represents the population of single age groups with the age of x to x  + 1 at the t  + 1 time. \(\left[ {{L^t}(x+1)/{L^t}(x)} \right]\) represents the survival ratio of age x to x  + 1 at t time. \(N{I^*}\) represents the net migration numbers in the corresponding age group from the t to t +  1 period, from other regions to Zhejiang.

Our estimated mortality will be used in CCP model. Since the total fertility of Zhejiang is lower than that of China, this study assumed that the total fertility of Zhejiang would start at 1.04 in 2020 based on Zhejiang population census [ 60 ]. Then, the future trend of Zhejiang’s total fertility would follow the United Nations’ estimated change rate of total fertility of China from 2020 to 2080 [ 47 ]. For net migration, The Census Survival Ratio Method was used to estimate the migration pattern based on the census data [ 61 ]. As one of the highest net in-migration provinces since 2010, Zhejiang will face the lower net in-migration intensity and be close to migration equilibrium in 2040 [ 62 ]. Based on this, it is assumed that the net migration rate in Zhejiang will experience a linear decrease and realize migration equilibrium by 2045.

Disability is defined as a difficulty in performing at least one of six Activities in Daily Living (ADL) [ 63 ], including bathing, dressing, eating, getting in/out of bed, using the toilet, and controlling urination and defecation in CHARLS. Then, mild disability is defined as having difficulty in 1–2 items of ADL, moderate disability as having difficulty in 3–4 items of ADL, and severe disability as having difficulty in at least 5 items of ADL [ 64 , 65 ]. Based on the discussion on the Disease Expansion, Disease Compression and Dynamic Equilibrium Theory [ 66 ], three different scenarios in changing disability were calculated [ 16 ]: a 0.8% annual decrease for age-specific prevalence rate of disability as the low disability scenario, the constant age-specific prevalence rate of disability as the middle disability scenario, and a 0.8% annual increase for age-specific prevalence rate of disability as the high disability scenario.

Dynamic financing actuarial model of social LTCI schemes

The study built a macro simulation model to further simulate the expenditure, financing and fund balance of LTCI based on the projection of severe disabled older population ( \(DisOP\) ) aged 60 and above and contribution population ( \(CP\) ) of LTCI aged 20 and above. The macro model is showed below:

In Formula (4), \(LTCE\) means LTC expenditures, \(HbdcCost\) , \(IcCost\) , \(HbdmcCost\) and \(NhcCost\) represent the unit cost of home-based daily living care (HBDC), institutional care, home-based daily living & medical care (HBDMC) and nursing hospital care per person per year, respectively. Among them, HBDC means that beneficiaries only receive formal daily living care services at home but without any medical care. HBDMC means that beneficiaries receive both formal daily living care services and professional medical care services at home. The difference of institutional care and nursing hospital care lies in that the former focuses more on daily living care, while the latter specializes in medical care. From 2023 to 2080, the unit cost of each type of LTC services is given an increase of 2.25% annually based on the average increase of healthcare CPI from 2010 to 2020. \(\alpha \) means the percentage of different types of LTC services utilization. Formula (5) describes the dynamic financing model and current balance of LTCI every year. \(premiu{m_{{t_0}}}\) is the fixed amount of premiums of LTCI in our base period. \(\lambda \) is annual growth of the amount of LTCI premiums. Formula (6) shows the accumulated balance of LTC fund which is determined by the current balance and the accumulated balance in previous period. \(\gamma \) is the interest rate of LTCI fund which represents the time value of the LTCI fund. Taking the inflation rate (2.25%) as a reference in the simulation process, we test the minimum value of \(\lambda \) that ensures a consistently positive accumulated balance in the LTCI fund up to 2080 across various disability scenarios.

Parameters of LTCI schemes in Zhejiang Province, China

The policies of LTCI schemes in five pilot cities in Zhejiang are sorted in Additional Table 1  (see Additional file 1 ) [ 41 ]. The LTCI schemes in Jiaxing City are representative among five pilot cities of LTCI in Zhejiang. Firstly, Jiaxing is the first city covering all employees and urban and rural residents equitably with the same benefits and premium since the adoption of LTCI (in 2017), which has navigated the reform of LTCI in Zhejiang. Secondly, LTCI benefits in Jiaxing are at the middle level among the five pilot cities, which is representative of average level in Zhejiang. The maximum benefits of HBDC in Jiaxing are lower than those in Yiwu and Wenzhou, and equal to those in Tonglu and Ningbo. Besides, the maximum benefits of institutional care are also lower than those in Yiwu, but higher than those in Tonglu and Ningbo. Overall, Jiaxing’s LTCI benefits stay average in Zhejiang. Thirdly, LTCI financing criteria in Jiaxing align with Ningbo and Tonglu (90 Chinese Yuan (CNY)/person/year), reflecting the standards across five cities. Therefore, this study adopted Jiaxing’s LTCI criteria as the parameters of LTCI simulations in Zhejiang. The unit costs of HBDC, institutional care, HBDMC and nursing hospital care are set at 1200 CNY/month, 2100 CNY/month, 1680 CNY/month and 1680 CNY/month in 2024 according to LTCI maximum benefits in Jiaxing (see Additional Table 1 , Additional file 1 ) [ 41 ]. The contributory group of LTCI is the group participating in social health insurance, whether retired or not. The LTCI financing parameter \(premiu{m_{{t_0}}}\) is based on a fixed amount of premiums in Jiaxing, of which the standard is 90 CNY/person/year [ 41 ].

Chinese government proposed a model of elderly care named “9073” model: 90% of older people receive home-based care, 7% receive community care and 3% receive institutional care [ 67 ]. “9073” model represents the prospects of China’s elderly care and is therefore suitable for the long-term simulation in this study [ 29 , 62 ]. Specifically, proportion of HBDC ( \({\alpha _{\text{1}}}\) ), institutional care ( \({\alpha _2}\) ), and combination of HBDMC and nursing hospital care ( \({\alpha _3}\) + \({\alpha _4}\) ) are set at 90%, 3% and 7%, respectively. Disabled older people can choose to receive HBDMC at home or receive nursing hospital care at medical institutions when facing medical care needs. It is free to choose the locations for these two LTC services, and it is quite similar to receiving community care in nature, as it also allows the option of receiving services at home or at community centers. Additionally, the LTCI benefits of these two LTC services in Jiaxing are equal. Therefore, we grouped them together when determining the beneficiaries’ choice of LTC services type ( \({\alpha _3}\) + \({\alpha _4}\) ). We set the interest rate of LTCI fund at 2.5% based on current interest rate of 5-year time deposit in China’s banks [ 68 ]. The sources of each parameter for simulation framework of the study are demonstrated in Additional Fig.  1 (see Additional File 1 ).

The mortality pattern and life expectancy of Zhejiang

The estimated mortality of Zhejiang from 1990 to 2020 is demonstrated in Fig.  1 based on adjusted DCMD model. Overall, the mortality for male and female presents a declining trend. Specially, the child mortality had a continued decline during the estimation period, but the adult mortality and old-age mortality had a slight increase between 1990 and 2000, then with a sharp decline between 2000 and 2020 afterwards.

We further predict the life expectancy at birth with 95% confidential interval under the LC_ER model from 2020 to 2080. The estimated and predicted life expectancy is demonstrated in Fig.  2 . Life expectancy of female had a stable increase from 1990 to 2020. While there was a slight decline of life expectancy of male from 1990 to 2000, then there was a rapid increase until 2020. The model results based on historical information show that life expectancy of both female and male will have an upward trend from 2020 to 2080. Besides, the gender difference in life expectancy will remain relatively stable in the future. In 2020, life expectancy was 80.46 years for male, 84.66 years for female. In 2080, the life expectancy will reach 89.39 [86.61, 91.74] years for male, 91.24 [88.90, 93.25] years for female. Besides, the age-specific rates of mortality decline of Zhejiang from 2021 to 2080 estimated by LC_ER are illustrated in Additional Fig.  2 (see Additional File 1 ).

Mortality pattern of Zhejiang in 1990–2020 based on adjusted DCMD model

Estimated and predicted life expectancy of Zhejiang in 1990–2080

The simulation of long-term care demand and expenditures in Zhejiang

Based on CCP method, the study has projected the number of older people and the number of severely disabled older people with different scenarios of disability in Zhejiang from 2020 to 2080 (shown in Table  1 ). It is illustrated that the population aged 60 and above in Zhejiang will firstly expand to around 2060 and then shrink until 2080. The number of older people with disabilities, especially those with severe disability, reflects the long-term care demand from a demographic perspective. We found that the number of older people with severe disability will continue to increase to 2080 under both medium and high disability scenarios. However, the number of older people with different degrees of disability will increase before 2060, and then decline slightly in the following 20 years under the low disability scenario. We also found that the number of severely and moderately disabled older people will be of little difference before 2050, which means that severe and moderate LTC demand is roughly equal.

Besides, the results of LTC demand under the high and low life expectancy scenarios are illustrated in Additional Table  2 and Additional Table  3 (see Additional file 1 ). It can be seen that Zhejiang Province will have a higher LTC demand under the scenario of higher life expectancy. The number of older people with severe disability under 95% upper interval of life expectancy in 2080 is 154 thousands higher than that under the predicted mean of life expectancy. And the number of older people with severe disability under 95% lower interval of life expectancy in 2080 is 169 thousands lower than that under the predicted mean of life expectancy. This result demonstrates the importance of mortality level prediction for assessing LTC demand.

Our study further calculated the LTCI expenditure paid by insurance fund every year from 2020 to 2080 to analyze the future long-term care demand in our case area from a financial perspective. The expenditure from LTCI illustrates an upward trend from 2023 to 2080 (see Fig. 3 ), with the higher price of long-term care services and increasing number of severe disabled older people. The LTCI expenditure is still increasing although there will be a slight decline in severe disabled older people under low disability scenario.

Projection of Long-term care insurance expenditure in Zhejiang, 2024–2080. Notes Results are based on the predicted mean of life expectancy

The simulation of LTCI fund under diverse disability and financing scenarios

The accumulated balance of LTCI fund from 2022 to 2080 is simulated on different dynamic financing growth scenarios in order to test how to make LTCI achieve actuarial balance in the long run. The accumulated balance and current balance of LTCI fund in Zhejiang are shown in Figs.  4 and 5 . When we set the annual premium growth rate at 2.25% which is equal to the average increase of healthcare CPI, there will be a deficit of current balance before 2028. As a result, the accumulated balance will become negative in 2032 under medium disability scenario, under high disability scenario in 2030 and under low disability scenario in 2036. This result shows that LTCI fund can only be sustainable within 12 years if the financing level grows at a low pace from 2024.

Accumulated balance of LTCI fund under different financing and disability scenarios. Notes Results are based on the predicted mean of life expectancy

Current balance of LTCI fund under different financing and disability scenarios. Notes Results are based on the predicted mean of life expectancy

The minimum annual premium growth is further tested to achieve the positive accumulated balance of LTCI fund under various scenarios from 2022 to 2080. We found that when the annual premium growth rate equals to 4.20%, LTCI fund will realize the long-term sustainable under low disability scenario, which means that the 4.20% financing growth standard is effective to make LTCI sustainable at a relatively low premium level under low disability scenario; however, it will still face the risk of the shortage of financing with 4.20% annual premium growth under the medium and high disability scenarios after 2039 and 2033.

Furthermore, the accumulate balance of LTCI fund remains at a moderate surplus and will not face a shortage until 2080 under the medium disability scenario when the annual premium growth rate equals 5.25%. Although the current balance of LTC fund will be negative in 2043 to 2058 under 5.25% annual premium growth (see Fig.  5 ), the accumulated surplus before 2042 and continuous interest will still realize the accumulated surplus of LTCI fund (5.83 billion CNY) in 2058. Overall, the annual premium growth rate at 5.25% is the best parameter choice if the age-specific prevalence rate of disability in Zhejiang Province is projected to remain stable. Finally, LTCI will be sustainable under all disability scenarios when the premium increases by 6.25% per year. However, this level will put a heavy payment burden on the residents, and there will be a large amount of fund redundancy if the disability does not continue to increase.

The simulation of LTCI fund under diverse life expectancy and financing scenarios

The impact of different life expectancy trend on the sustainability of LTCI schemes is further discussed. The simulation results of accumulated balance of LTCI fund under predicted mean, 95% upper confidential interval and lower confidential interval of life expectancy scenarios are demonstrated in Fig.  6 . It is learned that the sustainability of the LTCI fund will face a completely different situation in the long-term because of the difference trends in life expectancy even under the same disability level and financing level. Under the 5.25% annual premium growth rate and medium disability scenario, LTCI fund will become accumulated deficit under 95% upper interval of life expectancy after 2045. However, the LTCI fund will always remain in surplus before 2080 with the predicted mean or lower 95% interval of life expectancy. Therefore, the balance of LTCI fund is sensitive to life expectancy. In addition to affecting LTC expenditures when other conditions are the same, life expectancy is also related the total amount of financing by the number of contributors, thereby influencing the sustainability of LTCI fund.

Current and accumulated balance of LTCI under different life expectancy scenarios. Notes Results are based on the 5.25% annual premium growth rate scenario and medium disability scenario

This study shows two novel contributions to the existing literature. The first contribution is that we have found an important but often overlooked point that LTCI financing is sensitive to the variability of life expectancy in the long-term. In 2080, the 95% upper interval of the life expectancy in Zhejiang Province will be 2.01 years for female (2.35 years for male) higher than the predicted mean, and its cumulative impact will make LTCI unsustainable 35 years in advance. This finding shows that the accurate estimation of life expectancy is critical for assessing the sustainability of social insurance schemes like LTCI [ 69 , 70 ], and also reveals the significance of life expectancy analysis in this study, because health factors can be dynamically monitored through the evaluation and reimbursement records within the LTCI system [ 34 , 71 ], but life expectancy estimation will become difficult due to the lack of timely statistical data. Besides, the study also finds that LTCI financing is also sensitive to the variability of prevalence rate of disability in the long-term. Only 4.20% annual growth of premium can make Zhejiang’s LTCI sustainable under a disability compression assumption. However, the 6.25% annual growth of premium is necessary for Zhejiang’s LTCI sustainability under disability expansion assumption. The results are consistent with some existing research with various disability scenarios [ 28 , 72 ]. The overall incidence of disability will face a growing trend with population aging [ 17 ]. Therefore, proposing health promotion and postponing disability actions to reduce the incidence and duration of severe disability among older people will mitigate the pressure of LTCI funding [ 73 ].

The second contribution is that Zhejiang’s LTCI financing needs to grow at a relative high speed annually (5.25% under the medium scenario) to achieve sustainability in the long-term. It should be noticed that the LTCI financing parameters to achieve short-term and long-term fund equilibrium are different, and it is clear that long-term fund balance is a necessary condition to ensure the sustainability of the system [ 22 , 29 ]. If the accumulated surplus of the LTCI fund in Zhejiang Province before 2050 is used as a criterion for determining sustainability, as many studies have done [ 19 , 74 ], our results indicate that Zhejiang LTCI fund is projected to experience an accumulating deficit for over 20 years after 2050. Like Zhejiang, there are also several pilot cities in China that have adopted the fixed amount of premium model without premium adjustment [ 32 ]. LTCI funds in these regions will run the risk of accumulating deficits in the short term [ 43 ]. China and other countries adopting social LTCI need to adjust the scale of premium in a timely and dynamic manner to cope with the long-term LTCI financing pressure since China’s aging plateau will continue after 2060 [ 47 ].

Our simulation results can also be used as a reference for countries and regions that adopt a fixed percentage of income model of LTCI financing although we focus on the fixed amount model of LTCI financing. The study finds that LTCI premium in Zhejiang needs to increase by 5.25% per year to ensure sustainability to 2080 under the assumption of disability with dynamic equilibrium. However, the growth rate may exceed the income growth rate of some countries in the context of declining global economic growth [ 75 ]. Therefore, even those countries based on a fixed percentage of income model need adjust financing parameters dynamically [ 1 ]. In LTCI fund management, China and other countries can learn from Germany’s experience to deal with the long-term impact of population aging, which has established a demographic reserve fund which saves 0.1% of premium every year for payment in the future [ 25 ].

Reasonable coverage and benefits are also important factors to achieve sustainable LTCI. Like developed countries, the LTCI pilot cities in Zhejiang Province cover all urban and rural residents. However, most of the LTCI pilot cities in China only cover urban employees [ 35 ]. Therefore, the analysis of LTCI in Zhejiang Province in this paper provides implications for other LTCI pilot cities in China to expand the coverage and promote the equity of receiving LTC. Besides, it should be noted that this study only considers the older adults with severe disabilities according to the rules when estimating LTC needs in Zhejiang Province [ 36 ]. Whereas, it is not only the families of severely disabled groups that face the burden of long-term care [ 17 ]. Moderately disabled people in some developed countries and pilot cities in China are also covered by LTCI [ 76 , 77 ]. Even considering only severe disability, our simulation results show that only a high premium growth rate can make the system sustainable in the long run. Therefore, LTCI policymakers need to comprehensively weigh residents’ payment pressure and long-term care benefits, and make a balance between expanding coverage and increasing financing with the aim of protecting the most vulnerable groups.

This study has explored and built a long-term care insurance system that can be a reference for China and other developing countries to provide LTC services for the disabled older adults in the future. The strength of this study is that a more accurate life expectancy estimation based on the DCMD model is adopted when estimating dynamic financing of LTCI. However, this paper still has some limitations. Firstly, the paper only considers the activities of daily living when estimating the prevalence rate of disability of older people in Zhejiang Province, but does not consider cognitive function, perception and communication function due to the lack of data. Secondly, this study only considers the expenditure cost of LTC in the simulation analysis, but does not consider the operating cost of the LTCI system. Thirdly, this study only considers the total amount of financing for LTCI, but does not discuss the financing structure including individual contributions, government subsidies, and pooling funds. Finally, this study focuses only on the case in Zhejiang, but does not simulate the LTCI financing standard for actuarial equilibrium in other LTCI pilot areas in China.

In summary, this study estimates and predicts the mortality rate in Zhejiang Province from 1990 to 2080 through the DCMD model and LC model, and further evaluates the increasing LTC need in Zhejiang Province in the future. The LTCI dynamic financing in Zhejiang Province under different disability scenarios and life expectancy scenarios is simulated on the LTCI expenditure forecast results, and it is found that only by maintaining a relatively high level (5.25% under medium scenario) of premium growth can Zhejiang’s LTCI be sustainable in the long run. Our empirical case in Zhejiang offers implications for developing countries and LTCI pilot areas that lack high-quality mortality information to establish and dynamically optimize LTCI financing. Therefore, policy makers are called upon to assess the sustainability of LTCI from a long-term perspective, and regularly monitor changes in residents’ health and life expectancy to ensure that LTCI fund can meet LTCI expenditure and control the financing burden.

Data availability

In this study, all the data sources are publicly available. The data calculated in this study is available upon request to the corresponding author.

Abbreviations

Long-term care

• Long-term care insurance

China Health and Retirement Longitudinal Study

Center for Disease Control and Prevention

Consumer Price Index

Developing Country Mortality Database

Lee-Carter method extended with rotation

Cohort-component projection

Chinese Yuan

Home-based daily living care

Home-based daily living & medical care

Chen L, Zhang L, Xu X. Review of evolution of the public long-term care insurance (LTCI) system in different countries: influence and challenge. BMC Health Serv Res. 2020;20:1–21.

Article   CAS   Google Scholar  

Collaborators GBD, Ärnlöv J. Global age-sex-specific fertility, mortality, healthy life expectancy (HALE), and population estimates in 204 countries and territories, 1950–2019: a comprehensive demographic analysis for the global burden of Disease Study 2019. Lancet. 2020;396(10258):1160–203.

Article   Google Scholar  

Cullinan J, Gannon B, O’Shea E. The welfare implications of disability for older people in Ireland. Eur J Health Econ. 2013;14:171–83.

Article   PubMed   Google Scholar  

Karagiannidou M, Wittenberg R. Social Insurance for Long-Term Care. J Popul Ageing. 2022;15(2):557–75.

Article   PubMed   PubMed Central   Google Scholar  

Brown JR, Finkelstein A. The private market for Long-Term Care Insurance in the United States: a review of the evidence. J Risk Insur. 2009;76(1):5–29.

Cardoso T, Oliveira MD, Barbosa-Póvoa A, Nickel S. Modeling the demand for long-term care services under uncertain information. Health Care Manag Sc. 2012;15:385–412.

Liou L, Joe W, Kumar A, Subramanian SV. Inequalities in life expectancy: an analysis of 201 countries, 1950–2015. Soc Sci Med. 2020;253:112964.

Leon DA, Jdanov DA, Shkolnikov VM. Trends in life expectancy and age-specific mortality in England and Wales, 1970–2016, in comparison with a set of 22 high-income countries: an analysis of vital statistics data. Lancet Public Health. 2019;4(11):e575–82.

Vollset SE, Goren E, Yuan C-W, Cao J, Smith AE, Hsiao T, Bisignano C, Azhar GS, Castro E, Chalek J. Fertility, mortality, migration, and population scenarios for 195 countries and territories from 2017 to 2100: a forecasting analysis for the global burden of Disease Study. Lancet. 2020;396(10258):1285–306.

Spillman BC, Lubitz J. The effect of longevity on spending for acute and long-term care. New Engl J Med. 2000;342(19):1409–15.

Article   CAS   PubMed   Google Scholar  

Schulz E, Leidl R, König H-H. The impact of ageing on hospital care and long-term care—the example of Germany. Health Policy. 2004;67(1):57–74.

Cao X, Hou Y, Zhang X, Xu C, Jia P, Sun X, Sun L, Gao Y, Yang H, Cui Z. A comparative, correlate analysis and projection of global and regional life expectancy, healthy life expectancy, and their GAP: 1995–2025. J Glob Health. 2020;10(2).

Rickayzen BD, Walsh DE. A multi-state model of disability for the United Kingdom: implications for future need for long-term care for the elderly. Brit Actuar J. 2002;8(2):341–93.

Crimmins EM, Beltrán-Sánchez H. Mortality and morbidity trends: is there compression of morbidity? J Gerontol B-Psychol. 2011;66(1):75–86.

Chatterji S, Byles J, Cutler D, Seeman T, Verdes E. Health, functioning, and disability in older adults—present status and future implications. Lancet. 2015;385(9967):563–75.

Bai C, Chen H, Wang Z, Zeng Y. Regional Comparative Research of Home-based care needs and costs Predictionfor Disabled elders in China. China Econ Quart. 2022;22(02):653–76.

Google Scholar  

Zeng Y, Chen H, Wang Z, Land KC. Implications of changes in households and living arrangements for future home-based care needs and costs for disabled elders in China. J Aging Health. 2015;27(3):519–50.

Vanella P, Heß M, Wilke CB. A probabilistic projection of beneficiaries of long-term care insurance in Germany by severity of disability. Qual Quant. 2020;54(3):943–74.

Zhang L, Fu S, Fang Y. Prediction the contribution rate of long-term care insurance for the aged in China based on the balance of supply and demand. Sustainability-Basel. 2020;12(8):3144.

Villalobos Dintrans P. Designing long-term care systems: elements to consider. J Aging Soc Policy. 2020;32(1):83–99.

Fong JH, Borowski A. Long-term care insurance reform in Singapore. J Aging Soc Policy. 2022;34(1):73–90.

Bahnsen L, Wimmesberger FM. Curbing the demographic drifting dune in long-term care insurance financing: the case of Germany. Int Soc Secur Rev. 2023;76(1):35–56.

Central Provident Fund Board of Singapore. CareShield Life premiums and subsidies. 2020. https://www.cpf.gov.sg/member/healthcare-financing/careshield-life/careshield-premiums-and-subsidies . Accessed 01 February 2024.

Kim H, Kwon S. A decade of public long-term care insurance in South Korea: policy lessons for aging countries. Health Policy. 2021;125(1):22–6.

Nadash P, Doty P, von Schwanenflügel M. The German long-term Care Insurance Program: evolution and recent developments. Gerontologist. 2018;58(3):588–97.

Feng Z, Glinskaya E, Chen H, Gong S, Qiu Y, Xu J, Yip W. Long-term care system for older adults in China: policy landscape, challenges, and future prospects. Lancet. 2020;396(10259):1362–72.

Biase Pd, Dougherty S. From local to national: delivering and financing effective long-term care. Paris: OECD Publishing; 2023. https://doi.org/10.1787/578b296f-en .

Book   Google Scholar  

Zhang L, Fu S, Wang Y, Fang Y. Research on the optimization of financing scheme of long-term care insurance in China based on system dynamics simulation. Front Public Health. 2022;10:908864.

Guo P, Li Y. Innovation and financial sustainability in China’s long-term care insurance: an empirical analysis of the Nantong pilot. Front Public Health. 2022;10:1024491.

Zhang L, Fu S, Fang Y. Research on financing mechanism of long-term care insurance in Xiamen, China: a system dynamics simulation. Front Public Health. 2021;9:714044.

Mi H, Ji M, Liu W. Long-term Care Insurance Blue Book 2018: research on long-term Care Insurance in Qingdao, China. Beijing, China: China Human Resources Social Security Publishing Group Co, Ltd; 2019.

Chen H, Zhao S, Cui B. An evaluation study on the risk of financial deficit for the long-term care insurance pilots: based on analysis of the operational scheme in the 15 first-batch pilots. Chin J Health Policy. 2021;14(12):42–50.

Yang H, Fan W, Jiang W. Estimation of the financing burden of long-term care insurance funds:comparative analysis based on 8 LTC insurance pilot regions. Chin J Health Policy. 2023;16(06):1–9.

Chen H, Zhao S. Financial sustainability of the long-term Care Insurance-based on Microsimulation and Claims Data. Insur Stud. 2021;(10): 64–78.

Zhou W, Dai W. Shifting from fragmentation to integration: a systematic analysis of long-term care insurance policies in China. Int J Integr Care. 2021; 21(3).

Zhejiang Healthcare Security Administration. Guiding Opinions on Deepening the Pilot Program of the Long-Term Care Insurance System. 2022. http://ybj.zj.gov.cn/art/2022/6/28/art_1229113757_2409463.html . Accessed 14 April 2023.

Wang Y, Li X, Zhou M, Luo S, Liang J, Liddell CA, Coates MM, Gao Y, Wang L, He C. Under-5 mortality in 2851 Chinese counties, 1996–2012: a subnational assessment of achieving MDG 4 goals in China. Lancet. 2016;387(10015):273–83.

Zhejiang Provincial Statistics Bureau. Statistical monitoring report of children’s development planning in Zhejiang Province. 2020. http://tjj.zj.gov.cn/art/2020/9/30/art_1229129214_3637001.html . Accessed 14 April 2023.

Liu H. Research on disability grading based on ICF functional framework: empirical evidence from Zhejiang Province, China. Front Public Health. 2021;9:616180.

WHO. World report on ageing and health. World Health Organization; 2015.

Jiaxing Municipal Bureau of Healthcare Security. Long-term care insurance pilot implementation opinion for standard creation and efficiency improvement in Jiaxing city. 2023. https://ybj.jiaxing.gov.cn/art/2023/12/22/art_1229562038_2504155.html . Accessed 24 January 2024.

Wenzhou Healthcare Security Administration. Wenzhou City’s Implementation Plan for Deepening the Pilot of Long-term Care Insurance System (2023–2025). 2023. https://ybj.wenzhou.gov.cn/art/2023/5/24/art_1229135971_2022182.html . Accessed 02 February 2024.

Ningbo Municipal People’s Government. Guiding Opinions of the General Office of the Ningbo Municipal People’s Government on Deepening the Pilot of the Long-term Care Insurance System. 2022. https://www.ningbo.gov.cn/art/2022/9/13/art_1229533176_1752149.html . Accessed 02 February 2024.

Yiwu Healthcare Security Administration. Notice on adjusting some provisions of the. Yiwu City Long-term Care Insurance System Implementation Opinions (Trial). 2023. https://www.yw.gov.cn/art/2023/11/22/art_1229355060_52809.html . Accessed 02 February 2024.

Tonglu Healthcare Security Administration. Notice from Tonglu Medical Security Bureau on adjusting some policies of long-term care insurance. 2022. http://www.tonglu.gov.cn/art/2022/11/22/art_1229740453_1827580.html . Accessed 02 February 2024.

National Bureau of Statistics of China. National Data. 2023. https://data.stats.gov.cn/index.htm . Accessed 2 June 2023.

United Nations. World Population Prospects 2022. 2022. https://population.un.org/wpp/ . Accessed 27 July 2023.

Liu Z, Zheng H, Wu Y, Wang S, Liu Y, Hu S. Self-rated healthy life expectancy changes in Jiangxi Province of China by gender and urban–rural differences, 2013–2018. Front Public Health. 2021;8:596249.

Pandey SP, Adair T. Estimation of national and subnational all-cause mortality indicators in Nepal, 2017. BMC Public Health. 2022;22(1):1–11.

Wilmoth J, Zureick S, Canudas-Romo V, Inoue M, Sawyer C. A flexible two-dimensional mortality model for use in indirect estimation. Pop Stud-J Demog. 2012;66(1):1–28.

Li N, Mi H, Gerland P. Using child, adult, and old-age mortality to establish a Developing Countries Mortality Database (DCMD). Demography and Health Issues: Population Aging, Mortality and Data Analysis. 2018:51–62.

United Nations. World Population Prospects 2019: Methodology of the United Nations population estimates and projections. 2019. https://population.un.org/wpp2019/Publications/Files/WPP2019_Methodology.pdf . Accessed 27 July 2023.

Li C, Mi H. Assessments of provincial mortality in China’s 2010 population census based on the developing Countries Mortality Database model life table. J Popul Res. 2021;38(2):169–96.

WHO. WHO methods and data sources for life Tables 1990–2016. 2018. https://cdn.who.int/media/docs/default-source/gho-documents/global-health-estimates/lt_method_2016.pdf . Accessed 27 July 2023.

Li N, Lee R, Gerland P. Extending the Lee-Carter method to model the rotation of age patterns of mortality decline for long-term projections. Demography. 2013;50(6):2037–51.

Chen Z, Shi Y, Shu A. Managing Mortality and Aging Risks with a Time-Varying Lee–Carter Model. In: Healthcare-Basel : 2023: MDPI; 2023: 743.

Basellini U, Camarda CG, Booth H. Thirty years on: a review of the Lee–Carter method for forecasting mortality. Int J Forecast. 2023;39(3):1033–49.

United Nations. World Population Prospects 2022: Methodology of the United Nations Population Estimates and Projections. 2023. https://population.un.org/wpp/Publications/Files/WPP2022_Methodology.pdf . Accessed 01 February 2024.

Preston SH, Heuveline P, Guillot M. Demography: measuring and modeling population processes. Blackwell; 2000.

Zhejiang Provincial Bureau of Statistics. Part One of the Series Analysis of the Seventh Census of Zhejiang Province: Total Number and Distribution. 2022. https://tjj.zj.gov.cn/art/2022/7/22/art_1229129214_4955981.html . Accessed 01 Feb 2024.

Zachariah KC. A note on the census survival ratio method of estimating net migration. J Am Stat Assoc. 1962;57(297):175–83.

Deng Y, Liu S, Cai J, Lu X, Nielsen CP. Spatial pattern and its evolution of Chinese provincial population: methods and empirical study. J Geogr Sci. 2015;25:1507–20.

Jette AM. How measurement techniques influence estimates of disability in older populations. Soc Sci Med. 1994;38(7):937–42.

Manton KG, Gu X, Lamb VL. Change in chronic disability from 1982 to 2004/2005 as measured by long-term changes in function and health in the US elderly population. P Natl Acad Sci Usa. 2006;103(48):18374–9.

Chen L, Shi X. Forecast of China’s Long-Term Care Insurance Fund demand. Chin J Popul Sci. 2021;06:54–67.

Cai L, Lubitz J. Was there compression of disability for older americans from 1992 to 2003? Demography. 2007;44(3):479–95.

Hu B, Li B, Wang J, Shi C. Home and community care for older people in urban China: receipt of services and sources of payment. Health Soc Care Comm. 2020;28(1):225–35.

Zhu M, He M, Ma Z. Improving long-term Care Insurance: Model Selection and System Design. Chin J Popul Sci. 2023;01:3–20.

Lamnisos D, Giannakou K, Jakovljevic M. Demographic forecasting of population aging in Greece and Cyprus: one big challenge for the Mediterranean health and social system long-term sustainability. Health Res Policy Sy. 2021;19(1):21.

Bai R, Liu Y, Zhang L, Dong W, Bai Z, Zhou M. Projections of future life expectancy in China up to 2035: a modelling study. Lancet Public Health. 2023.

Liu C, Shu R, Liang H, Liang Y. Multimorbidity patterns and the disablement process among Public Long-Term Care Insurance Claimants in the City of Yiwu (Zhejiang Province, China). Int J Env Res Pub He. 2022;19(2):645.

Schut FT, Van Den Berg B. Sustainability of Comprehensive Universal Long-Term Care Insurance in the Netherlands. Soc Policy Admin. 2010;44(4):411–35.

Campbell JC, Ikegami N, Gibson MJ. Lessons from public long-term care insurance in Germany and Japan. Health Affair. 2010;29(1):87–95.

Kim N. Long-term care services expenditure projection in South Korea from 2015 to 2050. Int J Health Plan M. 2015;30(1):45–56.

International Monetary Fund. World Economic Outlook Update. 2023. https://www.imf.org/en/Publications/WEO/Issues/2023/07/10/world-economic-outlook-update-july-2023 . Accessed 15 August 2023.

Zhu Y, Österle A. China’s policy experimentation on long-term care insurance: implications for access. Int J Health Plan M. 2019;34(4):e1661–74.

Rhee JC, Done N, Anderson GF. Considering long-term care insurance for middle-income countries: comparing South Korea with Japan and Germany. Health Policy. 2015;119(10):1319–29.

Download references

Acknowledgements

We would like to thank Professor Xiangming Fang from Zhejiang University, Professor Guangdi Chen from Zhejiang University and Chengxu Long from King’s College London for their constructive advice during the research process of the paper. We would also like to appreciate any comments from the 34th REVES meeting.

This work was supported by the Major Project of Zhejiang Provincial Natural Science Foundation of China (LD21G030001).

Author information

Authors and affiliations.

School of International Studies, Zhejiang University, Hangzhou, China

School of Public Affairs, Zhejiang University, Hangzhou, China

Yichao Li & Hong Mi

You can also search for this author in PubMed   Google Scholar

Contributions

All authors contributed to the conception and design of the study, as well as the planning of analyses. XX and YL drafted the manuscript and performed the simulation analyses. YL & HM planned and prepared the simulation analyses. All authors took part in the revision of the manuscript for important intellectual content.

Corresponding author

Correspondence to Yichao Li .

Ethics declarations

Ethics approval and consent to participate.

Not applicable.

Consent for publication

Competing interests.

The authors declare no competing interests.

Additional information

Publisher’s note.

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Electronic supplementary material

Below is the link to the electronic supplementary material.

Supplementary Material 1

Rights and permissions.

Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/ . The Creative Commons Public Domain Dedication waiver ( http://creativecommons.org/publicdomain/zero/1.0/ ) applies to the data made available in this article, unless otherwise stated in a credit line to the data.

Reprints and permissions

About this article

Cite this article.

Xu, X., Li, Y. & Mi, H. Life expectancy, long-term care demand and dynamic financing mechanism simulation: an empirical study of Zhejiang Pilot, China. BMC Health Serv Res 24 , 469 (2024). https://doi.org/10.1186/s12913-024-10875-7

Download citation

Received : 26 August 2023

Accepted : 18 March 2024

Published : 15 April 2024

DOI : https://doi.org/10.1186/s12913-024-10875-7

Share this article

Anyone you share the following link with will be able to read this content:

Sorry, a shareable link is not currently available for this article.

Provided by the Springer Nature SharedIt content-sharing initiative

• Life expectancy
• Dynamic financing model
• Simulation analysis
• Zhejiang province

BMC Health Services Research

ISSN: 1472-6963

• General enquiries: [email protected]

Special Report

Automotive Bankruptcy Enterprise Values and Creditor Recoveries (2024 Fitch Case Studies)

Mon 15 Apr, 2024 - 8:56 AM ET

Little Auto Stress and Few Bankruptcies The easing of supply chain pressures along with ongoing pent-up demand due to industry under-production over the past several years is supporting a healthy auto sales market resulting in a low degree of default stress in the global auto sector. Most issuers in the sector have stable outlooks underpinned by solid balance sheets and sufficient financial flexibility that should leave ratings intact through a moderate stress scenario. Fitch has observed only one default from within the auto sector over the past 12 months, with Wheel Pros Inc. executing a distressed debt exchange (DDE) in September 2023. No auto issuer on Fitch’s Market Concern Lists has filed for bankruptcy since Garrett Motion sought Chapter 11 protection in September 2020 with $417.6 million in leveraged loan debt. In this edition, we include a case summary of IEH Auto Parts Holding LLC (known as Auto Plus), which filed for bankruptcy in January 2023 and emerged after executing a court-supervised sale of all assets in October 2023. (Fitch only includes issuers with syndicated term loans in its Market Concern Loan lists.) The company cited poor inventory management and an oversized retail footprint as factors contributing to the filing. The company’s bilateral loan lender had a near par cash recovery, with the sale consideration sufficient to pay off all major funded debt.