Big If True

One assumption people sometimes implicitly make when discussing optimizing social welfare is that when someone is more productive and thereby earns an additional dollar, it helps them without hurting anyone else. To put it in economic terms, income boosts your utility without any externalities.

However, it seems plausible that at least some of the utility you gain from consumption is positional - that is its utility you gain not by spending, but by spending more than your peers.

This hypothesis is, in my amateur opinion, one of the most under-appreciated ideas in the field of economics. If true, the idea would render large swarths of economic analysis useless. This isn't to say economists aren't aware of this issue - just that the vast majority of them ignore it in their analyses even when its clearly relevant.

We'll get to some concrete implications at the end. First, let's take a look at the evidence for this theory.

Suicides and Income

Utility can be hard to measure, but one proxy we can use is suicide rates in different groups since, presumably, people with lower utility are more likely to commit suicide. For instance, if income boosts utility, we'd expect people with higher incomes to be less likely to commit suicide. This prediction happens to be true.

Likewise, more income in an area generally makes an area nice to live in because it tends to mean better maintained infrastructure, higher-quality schools, and other benefits. And, indeed, suicides are less likely in counties with higher incomes.

However, the naive economic model that income is always good claims that when we try to predict suicide rates with both household income and county-wide income, we'll find both decrease suicides. This is not the case. Instead, we find that household income predicts lower suicide while county-income predicts higher suicide Relative status and well-being: Evidence from US suicide deaths. The researchers find, for instance, that you earn 10% more and the rest of the county earns 10% more, your odds of suicide actually increase by 3%. While other estimates in the study, find it would decrease suicide by a bit. While merely correlational, this is far more consistent with the theory that most utility from income is zero-sum than with the default theory that income just simply boosts individual and social welfare, alike.

The natural alternative hypothesis is that some third variable causes both higher incomes and higher suicide reates. The researchers investigated the most obvious confounder: cost of living. After controlling for housing prices, they find (1) high home prices actually negatively correlate with suicides and (2) after controlling for housing costs the effect of county income on suicides actually increases.

Being merely correlational, this still isn't a slam-dunk argument but while I can think of lots of reason why higher neighbor incomes should reduce suicides (better social services, lower crime, etc), I have a really hard time coming up with plausible reasons for the reverse with the exception of the positional-spending hypothesis. For this reason, I'm tempted to say that the slopes found here actually bound the causal effect size, which would suggest pretty much all utility derived from income among the middle and upper classes is positional.

Lottery Losers

More evidence from lottery losers.

The Dutch, for instance, have an interesting lottery system, where they partition the country into "neighborhoods" of roughly 20 households each. Then, each week, one neighborhood is selected as the winner and each ticket-buying households wins ~$15,000.

Conventional economic wisdom predicts winners will mostly buy consumer durables, and this prediction is borne out, with consumers particularly favoring cars - though, they also spent more in other categories.

The interesting conclusion is that the neighbors of winners were also significantly more likely to buy a car but not more likely to spend more on other things.

For instance, winners were 9% more likely to buy a new car since the lottery date while neighbors were 36% more likely. Neighbors increased the number of cars they owned by 10% while winners didn't increase the number at all. On the other hand, winners tended to buy cars 1.1 years newer, while neighbors bought cars ~0.3 years newer Kuhn.

This all perfectly aligns with the theory that cars purchases are largely regulated by status-seeking (positional) behavior rather than naive utility maximization.

Predictably, then, researchers hav also found that neighbors of lottery winners are more likely to go bankrupt Agarwal and that this effect increases with the size of the lottery winnings.

They also look into these bankruptcies and find that the larger the lottery winning the more value the filer had in "conspicuous" goods (houses, cars, etc) as opposed to "inconspicuous" goods (cash, financial assets, etc), where they find no effect. In particular, they find that a 1% increase in lottery winnings leads to 0.27% and 0.21% increases in the house and car values of their neighbor's bankruptcy reports.

Debt & Savings

Survey evidence finds similar results: people who perceive their social circle has generally higher income than themselves take on more debt Georgarakos. This is consistent with the idea that what is considered "adequate" is a function of your peers' consumption - not an absolute.

Likewise, positional consumption theory predicts higher income inequality will cause lower savings. While causation is difficult to tease out, it is certainly true that the last few decades have seen inequality increase and and savings decrease.

Lack of Labor Change

John Maynard Keynes famously predicted that his grandchildren would have a 15 hour work-week since technological improvements would allow that small amount of work to support a 1940s lifestyle.

Alas, this did not come to pass, and we work roughly as many hours now as we did then. As far as I can tell, there are really only two explanations for this: (1) the elasticity of utility to income is ~1 and (2) consumption for Americans is almost entirely positional.

The Easterlin Paradox

The Easterlin paradox Easterlin paradox is the observation that (a) self-reported happiness is higher for rich people and rich nations but (b) self-reported happiness doesn't increase as a nation gets richer.

If we believe that higher incomes boost happiness in a non-positional way, this is, indeed, paradoxical. However, if we believe higher incomes boost happiness in a almost entirely positional way, this "paradox" is completely resolved.

Other Evidence

To be clear, I'm hardly the first person to come up with any of this. Robert Frank wrote an interesting paper on the subject Frank and I'm going to list the evidence he gave (that I haven't already touched on):

  • Most people would prefer to live in a 3,000-sqft house while living amongst 2,000-sqft houses than live in a 4,000-sqft houses surrounded by 6,000-sqft houses.
  • Perception of status causes significant chemical changes in the brain.
  • Average self-reported well-being in country tend to be stable even during periods of high economic growth, while self-reported well-being within a country depends a great deal on income.
  • A study found that "a woman whose sister’s husband earned more than her own husband was 16–25 percent more likely than others to seek paid employment."
  • Pay scales are compressed within companies, which only makes sense if some of the compensation via rank. For instance, a medium FANG company engineer would probably be one of the best engineers at a different company but make less since part of their compensation at the other company would be via status (big fish in small pound).
  • Counties with higher income inequality tend to have higher median house prices, personal bankruptcy rates, and commute times - all suggesting people trade off conspicuous consumption (housing) with inconspicuous consumption (saving, reduced commute).
  • "Total hours worked, both across countries and over time within countries, are also positively associated with higher earnings inequality."
  • Given the same income level, black families save more than white ones, which makes sense if blacks compare themselves to other blacks.

Quick Summary of Evidence

Although I'm still fairly uncertain how much consumption is positional, the above evidence suggest the answer is probably "most of it".

For instance, the suicide income finds that higher incomes harm collective neighbor welfare about as much as it improves the earner's welfare, suggesting nearly 100% of consumption is positional.

Likewise, the study on car purchases found that when comparing lottery winners to their neighbors, there was mixed evidence regarding who actually increased their car consumption more! This, again, suggests the consumption is almost entirely positional.

To the extent you don't think utility-income elasticity is ~1, the fact that we're working as many hours no as we used to also suggests almost all our consumption-derived utility is positional.

Finally, the Easterlin paradox also suggests that pretty much all consumption in rich nations is positional.

Implications

Let's suppose for sake of argument that the above is all correct - that pretty much all consumption in rich nations is entirely positional in nature. What does that imply about optimal policy?

Quite a bit. This theory turns much of the "laissez faire" Econ 101 perspective on labor markets on its head.

For instance, conventional wisdom sees progressive taxes and welfare as being a balancing act of (1) giving resources to those who need them most while (2) not reducing the incentive to work too much. However the positional-consumption hypothesis suggests that disincentivizing labor is a feature of progressive taxation - not a bug - since labor has has negative externalities. In fact, the hypothesis even suggests that top earners should be taxed above the revenue maximizing rate!

Likewise, the hypothesis suggests that policies like minimum sick/vacation days, a 40 hour workweek, and limits on working hours are all chronically underused since (again) people work far more than is socially optimal.

In short, if we buy this hypothesis, then what economic policies we support shifts dramatically to the left.

Daly, M. C., Wilson, D. J., & Johnson, N. J. (2013). Relative status and well-being: Evidence from US suicide deaths. Review of Economics and Statistics, 95(5), 1480-1500. https://doi.org/10.1162/REST_a_00355 Table 1.1.3. Real Gross Domestic Product, Quantity Indexes. The Bureau of Economic Analysis. https://apps.bea.gov/iTable/index_nipa.cfm. Kuhn, P., Kooreman, P., Soetevent, A., & Kapteyn, A. (2011). The effects of lottery prizes on winners and their neighbors: Evidence from the Dutch postcode lottery. American Economic Review, 101(5), 2226-47. https://doi.org/10.1257/aer.101.5.2226 Agarwal, S., Mikhed, V., & Scholnick, B. (2016). Does inequality cause financial distress? Evidence from lottery winners and neighboring bankruptcies. https://papers.ssrn.com/sol3/papers.cfm?abstract_id=2731562 Georgarakos, D., Haliassos, M., & Pasini, G. (2014). Household debt and social interactions. The Review of Financial Studies, 27(5), 1404-1433. https://doi.org/10.1093/rfs/hhu014 Christen, M., & Morgan, R. M. (2005). Keeping up with the Joneses: Analyzing the effect of income inequality on consumer borrowing. Quantitative Marketing and Economics, 3(2), 145-173. https://doi.org/10.1007/s11129-005-0351-1 Frank, R. H. (2005). Positional externalities cause large and preventable welfare losses. American economic review, 95(2), 137-141. https://pubs.aeaweb.org/doi/pdf/10.1257/000282805774670392 Wikipedia contributors. (2020, October 31). Easterlin paradox. In Wikipedia, The Free Encyclopedia. Retrieved 11:25, November 1, 2020, from https://en.wikipedia.org/w/index.php?title=Easterlin_paradox&oldid=986420126