Apples and Uber

Why Markets Don’t Always Select The Best Technologies

After 30 years he had finally done it. David Bedford and his team at the University of Minnesota had been cross breeding apples for decades before finally coming up with what is today America’s favorite apple: the HoneyCrisp. Large, sweet, and crisp (as its name implies), the HoneyCrisp is the perfect hand apple. That is, it's the perfect apple for just eating. There are lots of good apples. The McIntosh apple is a tender, sweet-tart apple that is popular for stewing, and the Braeburn’s similar taste but crisp texture makes it great for baking. But if we judge by the price each apple commands, then the HoneyCrisp is king.

The decades of deliberate effort from Bedford to develop the HoneyCrisp is a sharp contrast to the chance discovery of an apple variety conspicuously missing from my examples of good apples above. Sometime in the 1870’s, an apple grower named Jesse Hiatt noticed a volunteer sapling in his orchard, so he chopped it down. When it grew again the following year, he chopped it down again. But when it popped up a third time the next year, Hiatt said, “If thee must grow, thee may.” 20 years later, he entered the first International New Fruit Show with fruit of that tree, simply called the “delicious”. Today we know that variety as the Red Delicious.

Red delicious, also unlike the HoneyCrisp, is only half well described by its name. While some may genuinely like the Red Delicious, and it certainly does have a beautiful red skin, it is not widely considered to actually be that delicious. Yet despite that, it had been America’s “most popular apple” for almost 80 years. Why?

To answer that question, let's go back to the introduction of the Red Delicious in 1893. It was more common then than it is today to simply plant apple seeds and use those trees for cider rather than cultivate specific varieties for eating. The most common variety at the time was the Ben Davis. The Ben Davis wasn’t popular because it tasted particularly good. In fact, it was quite bland. It was popular because it was hardy.

The even less delicious Ben Davis.

The 19th century lacked today's shipping technologies and global supply chains. An apple had to survive for a long time in rough conditions if farmers were going to ship them any substantial distance. This was why the Red Delicious had such an advantage. It had a bright red color that was very appealing and hid internal blemishes and bruises. The thick skin gave the apple a longer shelf life. And the Red Delicious turns red before it is fully ripe. So growers could pick them early and thus ship them further and they would still be ripe by the time they got on shelves. All of this, while still being a lot tastier than the Ben Davis.

As the United States began its transition to industrial agriculture, these advantageous traits helped kick Red Delicious production into high gear. During the 20th century the number of growers shrank, while the number of apples grown increased. Most apples were produced by only a handful of firms, each producing 10s of millions of apples a year.

Today there a small number of producers which produce mond bogglingly large quantities of apples.

And for all the reasons mentioned above, they were selecting to mostly grow Red Delicious. By the 1940s the Red Delicious had become the most popular apple in the U.S., and by the 1980s the Red Delicious made up 75% of the entire U.S. market.

By the end of the 1980s the Red Delicious had fallen to under 50% of total production, but still by far the most numerous variety.

But it doesn’t really matter anymore that an apple ships and stores well. Technology has advanced, and we can ship apples faster, further, and more gently than when Red Delicious rose to prominence. Furthermore, technologies like ethylene inhibitors ensure that apples of essentially any variety arrive fresh to stores.

And the point of comparison is no longer the Ben Davis. Since the Red Delicious there have been numerous other cultivars. As early as the 1910s the Golden Delicious was vying with its red counterpart. I’ve already mentioned HoneyCrisp, the McIntosh, and the Brauburn. But there are numerous others. My favorite is the Empire, but there are also the Granny Smith, Gala, Pink Lady, Fuji, Cortland, and quite literally dozens of others. Yet with all this competition, and its advantages nullified by advances in technology, why was Red Delicious still the most popular variety until 2018?

It all goes back to a little concept developed by economist W. Brian Arthur called “lock-in.” The basic idea is that in situations of increasing returns, historical events can cause a runaway effect where one technology dominates its competitors and gets “locked in” to the market. That is, only one technology effectively competes in the space, and it becomes very, very difficult for competing technologies to gain a foothold, even if they are better.

Arthur describes how this happened with the QWERTY keyboard, which survived competition, all the way through the transition from typewriters to keyboards, largely because no one wanted to have to learn a new typing method and companies didn’t want to have to retrain their entire workforces. Arthur describes how essentially random factors gave VHS technology a lead over Betamax that the technology, though technically superior in many ways, wasn’t able to overcome.

The Sony Betamax *almost* looks just like the VHS that eventually won out.

Red Delicious worked much the same. By the time American consumers were widely aware of the superior alternatives, it was too late. With only a few growers producing extremely large quantities of apples the cost and difficulty of switching varieties was too much for most growers to bear. Replanting an orchard is extremely costly - ripping up mature, productive trees to replace them with new saplings is expensive. And even after that large capital expenditure, it can take decades before the new trees to bear fruit, and therefore decades before that expenditure can be recouped. The risks are very high. Will the grower be able to convince sellers to buy the new variety? What if they switch prematurely and another, even more popular variety comes along? Once growers selected the Red Delicious, and then consolidated into only a few firms, it became essentially impossible for those firms to switch to a better apple. Red Delicious became locked in.

Even now, when Red Delicious are no longer the best selling variety in the U.S., those growers haven’t switched to another variety. In Washington, where most Red Delicious were historically grown, Red Delicious was still the most grown variety, even in the same year the cultivar lost its “most popular” title nationwide. Rather than switch, even with the writing on the wall, it is easier to simply export the apples! 60-65% of Red Delicious are now shipped abroad, mostly to China. In China, the Red Delicious doesn’t have the same level of competition as it does here. Consumers there also tend to value aesthetics a bit more than American consumers. So the bright red color, which also has cultural significance, gives the Red Delicious more appeal compared to other varieties than it does in the U.S.

The lock in of Red Delicious means its easier for growers to export the apples rather than switch to a more popular variety.

From this perspective, the Red Delicious is a technology like any other, and one that, in the right conditions, remained dominant even long after it became clear there were superior alternatives. Sometimes, the Market System doesn’t select the best option.

But what if the stakes were higher than just access to the most delicious apples?

Uber became the most valuable startup in the world in 2015. A lot has been said about what Uber did and why it was successful. Some claim that Uber is a disruptive innovator. Others claim that Uber’s main innovation was screwing over its workers. Yet others might scoff at the characterization of a company that doesn’t make money as “successful.” But Uber isn’t all that unique. It is part of a trend. Not just a trend of app based services, but a longer trend of companies trying to turn Arthur’s concept of lock-in through increasing returns into a business model.

In the standard model of a market, as demand goes up, price goes up. In response to higher prices, firms produce more. But with most commodities increasing production uses more resources, thus creating a disincentive for production to exceed certain levels.

This creates a stable equilibrium of supply, demand, and price. Sometimes, like the case with apples, those diseconomies of scale that create equilibrium happen at scales of production too large to avoid lock-in. Even though it is increasingly expensive to maintain increasingly large orchards, the cost of planting a new orchard is just so much higher than maintaining one. But with digital technologies, there just effectively isn’t a level of output at which a firm reaches diseconomies of scale.

When diseconomies of scale occur at reasonable outputs, markets stay at stable equilibria. But when there is no functional limit to economies of scale, equilibrium is unstable.

Think about how Microsoft became so dominant. Nearly all of the cost for Windows is in development. Copying software is nearly free. Even in an era of physical media, like floppy disks and CDs, producing the disks was so cheap as to be nearly negligible. So there was no disincentive for increasing the scale of production. In fact, the opposite. Such systems have so-called “increasing returns.” For each new user of Windows, the next user was more likely to choose Windows as well. The more people who used Windows, the more familiar people became with it. The easier it was to get support. The more locations you could buy it. The more programmers made third party programs for it. Growth only stops when the market is saturated.

But cars are physical commodities. How can there be increasing returns for a taxi service?

Because Uber isn’t a taxi service. They aren’t a service at all. They are a social media platform. Ubers main innovation was taking a physical service, and devising a way to leach onto it that could benefit from increasing returns.

The more drivers there are for Uber, the more areas they can effectively cover and the less time customers have to wait. The more customers there are, the more likely people are to sign up as drivers, as they will have more passengers and make more money. If Uber owned the fleet of cars, or actually employed their drivers, eventually the expense of adding more cars and drivers would outweigh the benefits of increased market saturation, and there would be an equilibrium. But they externalize those costs to the drivers. All Uber has to do is run an app, which doesn’t have the same diseconomies of scale.

Social media like Facebook or Twitter (I’m sorry, I can’t bring myself to call it X) work on the same principle. The more users there are, the more likely new users are to join and the more likely users are to stay. The more users there are the more valuable advertisements are, and the more likely advertisers are to use the platform.

As Arthur himself pointed out about Uber, there are hard limits. Once an area reaches a certain level of saturation, doubling that saturation doesn’t make a noticeable improvement to service for passengers. So, much like apples, Uber didn’t perfectly lock out all competitors. Lyft comes to mind. But it certainly has locked out, or at least contributed to dramatically negative outcomes for any competing transportation services that don’t try to take advantage of increasing returns as a business model.

While competitors like Lyft have cropped up, traditional taxis have been devastated. As a result, when the value of taxi medallions dropped, and driving stopped being a pathway to the middle class, some taxi drivers even committed suicide. And compared to taxi services, drivers at Uber are far less vetted. This has resulted in a service that is decidedly less safe. And of course taxi driver was a much better gig than Uber driver. Uber drivers earn notoriously low wages, and are still responsible for the costs and risks associated with driving. So while on the surface it might seem like just a convenient app that calls a cheap taxi, in many important ways Uber’s service is worse than the taxi service it locked out.

Some have also claimed that Uber and other rideshare apps could (or should) be replacements for public transit that operate at a “loss.” And in some ways, Uber is doing just that. Although some transit advocates hoped rideshare would act as a last mile solution, instead it seems like people are using such services as transit replacements. While in the Uber vs. Transit contest Uber is often the faster option, that advantage is offset by being substantially more expensive. And this tradeoff is more than just a personal choice. Rideshare is relatively inefficient. It makes traffic worse because there are so many cars driving around looking for passengers. And to the extent that transit agencies rely on fares, that means potentially reduced service. And, because poor people rely on transit more exclusively, rideshare is certainly a worse technology for them.

So, while many modern executives try to use increasing returns in order to lock their companies and products in to a market, they forget some of the key insights of Arthur’s theory. Namely that increasing returns doesn’t mean the best technology gets locked in. Instead, it can mean that we get stuck with sub-par technologies, which kill people, suppress wages, and lock out otherwise superior modes of transit so that we can’t easily switch to better alternatives. Uber locked-out taxis, made public transit worse, and the only thing we got in exchange was a class of gig workers barely able to make ends meat.