Positional competition: from low-number license plates to McMansions

Abstract

Over a century ago, Thorstein Veblen explored "conspicuous consumption" to signal social status. Building on Veblen, Fred Hirch (1976) defined a "positional good" as a good that derives its value from a social consensus regarding its desirability relative to other goods in its category. My thesis examines two positional goods: Delaware low-number license plates, and luxury houses, sometimes referred to as "McMansions." First, because car-owners in Delaware actually own their license plate numbers, the state has an active secondary market in low-number tags. These are perfect positional goods: they have no functional value beyond that of an ordinary license plate, and they have explicit rank order. I used sale-price and asking-price data to estimate an empirical pricing model and the total economic surplus embodied in this small market. Next, I analyze evidence of positional competition in the US housing market, using Decennial Census and American Community Survey data to show that (1) higher concentrations of wealth in a county induce high-income households to spend proportionally more of their incomes on housing, and (2) higher levels of local income inequality appear to augment this effect. My analysis also demonstrates how choice of residential location involves a trade-off between housing costs and commuting costs. In the housing analysis, the primary dependent variable (the percent of high-income households that spend more than 30 percent of monthly income on housing) is truncated at zero, which biases the OLS estimator. Therefore I estimate Tobit models of the positionality effect and test them for heteroscedasticity and normality; I compare these with CLAD models (Powell) which are robust with respect to non-normality and heteroscedasticity.