Content area
Abstract
This dissertation makes two related arguments: one about markups and one about the structure of consumption and preferences.
The main argument regards markups. There have been significant changes to markups since the 1980s: an increase in the average resulting from an increase in the right tail of the markup distribution, with the median relatively unchanged (De Loecker et al. 2020). Whereas most previous explanations rely on changes on the supply side -- changing levels of competition or market power -- I suggest that it is at least partly the result of changes on the demand side brought about by increasing income inequality. Because of declining marginal utility, rich consumers have lower price elasticities than poor consumers. Thus, luxury products, which cater more to the rich, will tend to have higher markups. Moreover, when the composition of consumers change because of an increase in income inequality, the composition of demand changes resulting in markups which change differently for different products. Luxury, high-markup products will tend to see an increase in rich consumers and fewer median income consumers, leading to lower average elasticities and higher markups. The number of poor consumers factors very little into the decision of these firms, as they make up such a small portion of their market shares. Conversely, basic, low-markup products, which tend to have equal shares of rich and poor consumers, will see much smaller changes in their markups.
The secondary argument -- although it comes first in the order of the dissertation -- regards the structure of demand. The argument about markups relies heavily on certain assumptions -- that price sensitivity declines with income, and that rich and poor consumers purchase different bundles of goods -- while the numerical importance of this channel relies heavily on the structure of demand -- how much and in what ways do consumption bundles vary across incomes? Investigating these features of demand leads me to conclude that preferences are satiable, creating a hierarchical demand structure. It is this satiability which lends its name to the dissertation's title.
The structure of the dissertation then is as follows. In Chapter 2, I examine the empirical features of consumption, income, and markups. Using a dataset of retail markups based on the Nielsen Homescan Database, I show that rich consumers tend to pay higher markups, suggesting that rich consumers may be less price sensitive than poor consumers. Next, I show some facts which lend credence to preferences being satiable: (i) inferior products are ubiquitous in the dataset -- a fact difficult to reconcile without satiability, and (ii) for a given category of goods, increases in expenditure are mainly the result of increasing average prices paid, rather than increasing physical quantities.
In Chapter 3, I present a model of satiable preferences which is able to explain the facts presented in Chapter 2. I show that, once these preferences are aggregated, they resemble a discrete choice model. However, whereas a discrete choice model is formulated for a single sector, the model presented here must take into account the macroeconomic characteristics of the model, represented by an endogenous marginal utility of money.
Chapter 4 then does the heavy lifting regarding our question of the effects of income inequality on the distribution of markups. Here, I calibrate the model presented in Chapter 3 to match the data for 2016. Then, I shock the model by changing the distribution of income to that in 1983, median-adjusted. Moving from the 1983 equilibrium to the 2016 equilibrium, the average markup is higher, and this is brought about by an increase in the right tail of the markup distribution, with the left tail relatively unchanged. The model is able to generate about 25% of the change in the average markup which we see empirically. I also explore the important welfare implications of the model. As well as the usual aggregate welfare costs of income inequality which come from a concave utility function, this model suggests that endogenous changes in markups may cause additional costs. The welfare analysis finds that the additional costs of changing markups is almost equally important to reducing welfare as the effects coming from a concave utility function. However, these costs would be even greater if all markups increase equally; the fact that they increase most for the rich lessens the increase in real income inequality to some extent.
