Anything at Scale

At some point it starts to break down.

Airbnb currently has only 5,500 listings in San Francisco (the number was 10,000 but the other 4,500 listings were recently removed after failing to register with San Francisco’s Office of Short Term Rentals). With 376,000 homes for San Francisco’s 870,000 people, Airbnb represents a small percentage of overall visitor occupancy. But compared to San Francisco’s 34,000 hotel rooms, with average hotel occupancy of around 85.5%, Airbnb starts to look large (even assuming a lower occupancy for Airbnb listings).

In New York City there are 3.4 million housing units (2014 data) of which Airbnb has 50,000 listed (2018) for a population of 8.5 million. So in New York and San Francisco, there are similar ratios of the number of housing units, population, and the number of Airbnb listings (after San Francisco’s crackdown). And in both cities, Airbnb is noted as causing rents to increase and in changing the landscape of the neighborhood. Research by David Wachsmuth and his team estimate that, over the last three years, Airbnb has increased long-term rents in New York City by 1.4 percent. The median household looking for a new apartment pays $384 more per year than they would have, due to the growth of short-term rentals.

In the Airbnb situations we have what seems like more of something (more places to stay) actually leading to higher prices. This is because in some cases, more listings actually leads to restricted supply.

Restricted supply happens when some tipping point of local Airbnb properties is reached. In extreme examples there are two effects. One is that already present high property rental prices require that renters or homeowners occasionally rent their property on Airbnb in order to just be able to live in the location themselves. Without the option to generate income by property rental, fewer people could afford to live in the area and prices would drop naturally. This is a dangerous positive reinforcement effect.

The other impact emerges when a significant percentage of limited local housing is Airbnb on a regular basis. This happens most easily in smaller communities, such as small resort towns. And in those resort locations that are seasonal in nature (attracting most of their visitors in the winter or most in the summer), this effect probably happens the fastest. As the tipping point percentage of Airbnb rentals is reached, it becomes difficult for locals to afford to live in their community year-round. As the high demand for a small number of residences increases price for a specific time of year, even if prices fall in the off season, what are the local residents to do during the high season?

These effects can be felt more readily where there is a larger ratio of tourists to locals. In Reykjavik of the 50,000 housing units, 5% are listed on Airbnb. This is three to four times the ratio of San Francisco or New York City. As a result, locals find it difficult to live in Reykjavik, which has recently become a town dominated by tourism.

Airbnb violates the multiple dwelling law in many locations (laws designed to prevent rentals of less than 28 days except to those who hold a hotel license). However, the company’s market entry depended on it being able to attract listings, grow to a size where it had clout, and then pressure local municipalities to allow it to continue to operate. It’s a business that only works when done at scale. Local municipalities were not prepared to handle this new entrant. And it all happened so fast.

Another interesting effect is when wealth scales. Around the world at different times specific populations that suddenly become wealthy create chaos elsewhere depending on their preferences. But the most visible example of this today is from China, which previously economically held back its population for decades. But over the last 30 years or so, that economic suppression started to unwind, with the effect being that today, there are tens of millions of wealthy Chinese mainlanders who are now suddenly looking for other places to put their money.

In China, the relative low security of keeping all of an individual’s savings domestic exacerbates the situation. Since there are limitations on moving assets out of China, people look for other options which are not prohibited, or where it is easier to avoid the prohibition. The most visual movement of this wealth today comes in the form of real estate purchases outside of China (there are many more less visible forms). As a result, there are communities with uninhabited housing units which serve as a store of wealth and also serve to reduce inventory and increase the price for locals who wish to buy or rent.

Local municipalities did not preemptively think to limit property purchases from those who do not live in the area. This formerly wasn’t a problem that needed to be considered. But more recently there are restrictions on property purchase by foreign national where the above problem is extreme, including Hong Kong, Australia, and Canada. The most interesting to me are the penalties for property vacancies. This was one of the few attempts I found that hacks away at negative effects of these scaling examples.

Demand at a Distance

Culture and history influence demand for physical items. An existing population that rather suddenly has money to spend means that a larger percentage of its people can act on their demands.

For example, rhinoceros horn, elephant ivory, cordyceps, and wild ginger mostly grow thousand of miles from their primary modern sources of demand. That there is demand at all for these items is rooted in historical medicinal preferences. But in the above examples, including that of the unlucky rhinoceros, either because of government restrictions or long growth cycles, there are limited ways to increase supply and harvest the product. When poached, the rhinoceros (and elephant) is often killed. Why? Poachers kill the rhinos rather than only cut off their horns since it is more dangerous to sedate these large animals than to kill them outright. Additionally, once poachers spend days tracking an animal only to find that its valuable horn or tusks are already gone, they kill the animals out of spite. For high-risk, high-payoff illegal work, with competition, and without opportunities to raise captive rhinos, keeping one without horns alive just means that the poacher might waste time tracking it again later. That is why preemptive rhino horn removal (or tusk removal in elephants) does not necessarily protect the animals.

Unlike the demand for other illegal drugs such as heroin (from opium poppies) and cocaine (from coca leaves), rhino horn is an animal product and not renewable in the ways it is illegally collected today. A great report on this specific problem is here. Two black markets exist for rhino horn in China and Vietnam. One is for medicinal purposes and the other for luxury products carved from the horn. Black market rhino horn has been sold for up to US$60,000 per kilo recently. And the number of rhinos poached grows with demand from wealthy in China and Vietnam: from 60 rhinos in 2006 to 1,342 in 2015.

China also banned rhino horn as a traditional Chinese medicine ingredient starting in 1993, but to less of an effect.

It’s easier to create car inventory than housing inventory or wealth and so with rideshare’s entrance into a location, we can see a more extreme effect.

New York City’s 13,000 taxis (a fixed number) each have a medallion and medallions can be bought and sold freely. Uber entered the New York City market in 2011, after years of taxi medallion price increases. But medallion prices kept rising for another year and plateaued before starting to drop from $1 million in 2014. For years after Uber and Lyft entered New York City, rideshare only made a small impact on the number of monthly taxi rides. Related to that, the value of a taxi medallion did not change much. And historically, we should note that taxi medallions were approximated in value as annuities calculated by taking the annual desired income for taxi driving divided by the interest rate. So when interest rates were low, taxi rate medallion prices rose. That was the trend over the last decade or more. But it took years of Uber and Lyft operating in the city to start to depress the number of taxi rides and, by extension, medallion prices. The dramatic fall in New York taxi medallion prices has led to multiple medallion owner suicides, sadly.