روحیه سرمایه داری در میان طبقات درآمد

This paper tests the consumption-based capital asset model within the context of the spirit of capitalism. The spirit of capitalism asserts that consumers gain utility not just from consumption of goods and services, but also from the social status obtained from wealth. We examine two asset pricing models developed by Bakshi and Chen (1996) that employ wealth in the utility function, for households sorted by income quintiles. In the first model, households obtain utility from both consumption and the social status that comes from their own wealth. In the second model, households gain utility from both consumption and the social status obtained from their own wealth relative to the wealth of other peer households. Our results indicate that both models are inconsistent with the data regardless of income. However, using cointegration methods as a diagnostic tool, we find that the data are “loosely” consistent with the spirit of capitalism, at least for the upper income quintiles.

The notion that wealth enters the household's utility function is a popular idea among financial economists wishing to explain consumption patterns. In many asset pricing models, wealth represents future consumption and may be used by households to reduce their consumption volatility. In the “spirit of capitalism” literature, however, wealth brings additional utility in the form of social status. The literature on the spirit of capitalism was introduced by Max Weber in the early 1900s as he argued that capitalism and the accumulation of wealth were consistent with Christianity.2 In more recent times, Bakshi and Chen (1996) further develop this idea by deriving several asset pricing models that are designed to empirically measure the spirit of capitalism. Their work, like most other empirical analyses of the consumption-based asset pricing model, uses U.S. aggregate consumption and wealth data within the context of a single representative agent. The representative agent model employed with aggregate data, however, ignores important differences in wealth among households. Such differences in wealth are crucial to an empirical analysis of the spirit of capitalism. Our paper contributes to the spirit of capitalism literature by examining the economic behavior of households with different levels of wealth. Specifically, our analysis employs five representative households (agents) based on their income quintiles. Empirical evidence indicates that on average households sorted into income quintiles are a good proxy for households sorted by their total wealth, defined as income, housing value, and financial assets. Both the Survey of Consumer Finance, published by the Federal Reserve, and the Consumer Expenditure Survey, published by the Bureau of Labor Statistics, support this finding. In this heterogeneous environment, consumption and wealth decisions for households seeking to gain social status critically depend on the wealth levels of peers, which are typically households in neighboring social strata. According to Duesenberry (1962, p. 30), people will often make comparisons to others with similar income levels. These comparisons will result in either favorable or unfavorable outcomes. Duesenberry conjectures that unfavorable comparisons bring about changes in household consumption of goods and services that may help to eliminate this disparity. For example, high-income households, say in the fourth income quintile, will manage their consumption and wealth based on the perceived wealth of households in the same quintile, or the nearby third or fifth quintiles. While the wealth of acquaintances is typically not known, it can be reasonably deduced. Individuals are quite perceptive, for instance, in determining the relative value of primary residences by their size, style, and location. Therefore, it is important to include housing values in the definition of wealth. Our test results, across all income quintiles, fail to support the spirit of capitalism models developed by Bakshi and Chen (1996). The crux of the problem lies in the parameter estimates, λ, for the spirit of capitalism, which are frequently negative and statistically insignificant. Such estimates imply that economic agents do not view wealth as a means of gaining social status. Furthermore, we find internal inconsistencies with other parameter estimates for these models. There are several possible explanations for the empirical failure of the models. First, the models may be too tightly parameterized, meaning that there are narrow acceptable ranges for each of the parameters estimated. Second, they depend on households operating with a specific utility function — power utility. Lastly, the models require the transformation of variables into growth rates, thereby losing some of the information found in the levels of the data. We relax these restrictions and test the data within the framework of cointegrating techniques. Cointegration techniques can assist in the task of determining whether the data are consistent with the spirit of capitalism. Our cointegration results are illuminating. We find that the consumption of higher income households depends in part on the wealth of other households, but not on their own wealth. In other words, the wealth of upper income households is not significant in explaining their consumption. It turns out that the wealth of others, which we refer to as “social wealth” and an index of the broad stock market, is significant. This finding suggests that higher income household consumption, social wealth, and the stock market are in a long-run equilibrium. Our findings also show that short-run deviations from this equilibrium lead to predictable changes in both future household consumption as well as own-wealth, as upper income households strive to “keep up with-” or “catch-up to the Joneses” as discussed in Roussanov (2010) and Abel (1990). This paper is organized as follows. Section 2 discusses the spirit of capitalism and related literature. Section 3 describes the data set used to test the spirit of capitalism models and establishes the economic heterogeneity among households that motivates us to steer away from U.S. aggregate data and towards data categorized by income quintiles. In Section 4 we employ two asset pricing models designed by Bakshi and Chen (1996) to test the spirit of capitalism. In Section 5 we examine household economic data sorted by income quintiles within a cointegrating framework. Cointegration techniques help diagnose whether the household data are consistent with the spirit of capitalism. Section 6 concludes the paper.

The spirit of capitalism is about households using wealth not just as a store for future consumption, but also for status purposes. Leveraging off of the asset pricing models developed by Bakshi and Chen (1996), we implement two models and test whether the consumption and wealth of households across income quintiles, combined with asset returns, are consistent with the spirit of capitalism. Our models distinguish between two measures of wealth for determining status. Absolute wealth is the first measure. With absolute wealth households obtain utility by simply holding wealth, independent of the wealth of other households. With the second measure of wealth, households gain utility from relative wealth. This is a form of bench-marking where households are assumed to measure their well-being by comparing the level of their wealth to the wealth of others. We refer to the wealth of others as social wealth. The objective of this analysis, using standard asset pricing models, is to determine whether households gain utility from both consumption and wealth. Our test results indicate that both models are inconsistent with the data. Our parameter estimates are internally inconsistent and economically implausible across all income quintiles under a variety of specifications. These results strongly suggest that the models might be too tightly parameterized due, in part, to the econometric nature of the data (nonstationarity) and the power utility function. With these results in mind, we implement cointegration tests with consumption, own-wealth, social wealth, and the market index. Unlike the asset pricing models discussed above, these tests are independent of a utility function and employ the household data in its original form, in levels. The ultimate goal of this analysis is to determine whether households alter their consumption and wealth decisions based on the wealth of peers. Our cointegration tests, employing absolute wealth, indicate that the data for households in the upper two income quintiles are the most consistent with the spirit of capitalism. However, when we add social wealth to the cointegrating space of these two upper-income quintiles, own-wealth becomes statistically insignificant. Thus, with social wealth in the cointegrating space, the data are consistent with the spirit of capitalism, but, we are left with a puzzling result: why is own-wealth not cointegrated with consumption for upper-income households? We leave this question for future research.