In our annual family weekend last spring, my nephew, Paul, organized a beer tasting test. He let us taste seven Dutch beers: Heineken, Grolsch, Bavaria, and some lesser-known brands including one from a supermarket. The question was simple: which beer is from which of the seven brands? The outcome: Each family member had about one brand correct on average. One! My seven-year-old kid would not have done worse if he had listed the beers randomly (without drinking them!).
What to make of this? Was my family too drunk after the beer tasting to list the beers in any sensible way? That seems unlikely. Most of my family would not even notice seven glasses of beer, let alone seven sips. A more probable explanation is that all those beers are truly indistinguishable taste-wise.
This made me wonder: Why do all beers taste the same? Why do beer companies not differentiate their beer? “That is a great question,” nephew Paul said. “It cannot be hard for a great economist like you to come up with a convincing answer.” The pressure was on! “Let me think about it,” I said, nonchalantly, and I withdrew to my room. The first thing I did was to find out what Paul might have meant by “a great economist.” I found the answer in the work of John Maynard Keynes:
“Economics is the science of thinking in terms of models joined to the art of choosing models which are relevant to the contemporary world.”
So, a great economist is someone who has mastered the art of choosing the model that has the greatest bite in a real-world setting.
The first model that came to mind was the Hotelling model, published by statistician and economic theorist Harold Hotelling in 1929. Here is a simple version of the model. Imagine a beach full of sunlovers, uniformly spread out over the beach. Two ice cream vendors, Ben and Jerry, sell identical ice cream at the very same price. At the start of the day, they decide where to locate their ice cream stands on the beach.
Where did I see a link with beer? I’d argue that a location on the beach is analogous to the taste of beer, let’s say on a scale from not bitter at all to very bitter. Consumers vary in terms of preferred bitterness, which corresponds to their locations on the beach. When picking a beer, they make ‘travel costs’ depending on how far the beer’s bitterness deviates from their preferred bitterness.
Back to Ben and Jerry. It is clear that both will locate in the middle. Their prices are fixed so their profits depend only on the number of customers. If Ben locates in the middle, the best thing Jerry can do is to stand next to him. He will get half the consumers. In any other spot, fewer consumers will come because they want to minimize their travel costs. I knew many applications of this model: In two-party elections, parties tend to avoid extreme programs (‘the median voter theorem’). The Economics & Business bachelor programs offered by the UvA and the VU are like two peas in a pot. In London, Brick Lane is famous for a cluster of Indian curry restaurants.
However, I soon decided that the Hotelling model failed to offer a convincing explanation of why all Dutch beers taste the same. For one thing, the number of companies is greater than two. If Ben and Jerry settle in the middle, Ola, a third ice cream vender, will be better off locating somewhere else. In fact, in the case of three or more firms, some degree of product differentiation will be observed according to Hotelling’s theory. Another critical assumption is that the prices are fixed. Of course, beer companies can change their price once they have settled on the taste of their beer. In that case, they have a strong incentive to avoid price competition and differentiate themselves as I tried to articulate in my previous column.
It was almost morning when I realized that with a small twist, the Hotelling model can be a reasonable description of the Dutch beer market. Imagine that Ben is ill one day so that Jerry is the only ice cream vender on the beach. Of course, Jerry will locate in the middle of the beach and sell his ice cream at a pleasing monopoly price. The next day Ben returns to the beach. Before he has the chance to open his ice cream stand, Jerry comes over and says: “Listen, Ben. We can make much higher profits when only one of us locates in the middle of the beach charging the monopoly price. Why don’t we make the following agreement: We alternate the days that we offer ice cream on the beach. You today, me tomorrow, you the day after, and so forth.” Ben immediately realized that this was a great plan. He only had to work once every two days and he would still earn more money than before.
Indeed, I thought, the Dutch beer market looks divided, too: Heineken serves the west of the Netherlands, Grolsch the east and the north, and Bavaria the south. The result is regional monopolies that, in line with the model, brew barely distinguishable beer. I also knew that a couple of years ago, Dutch beer brewers Heineken, Grolsch, and Bavaria were fined by the European Commission for cartel formation.
The next day, nephew Paul congratulated me on my reasoning. He added: “In line with your model, we would expect much greater product variety in the UK which has many more beer brewers than the Netherlands.” “Why?” I asked. “Well, it is much harder to sustain a cartel agreement in such a fragmented market. Indeed, there is little evidence of cartel pricing in the UK, as Margaret Slade shows. Moreover, there is a much greater variety of beer on offer: stout, brown ale, golden ale, pale ale, Indian pale ale, session bitter, best bitter, premium bitter, to name a few.”
It was a fun family weekend. Cheers!