In the post-WWII period, U.S. stock returns have averaged 9%, whereas returns on treasury bills have been around 1%. Standard theories attribute such systematic differences in expected returns among different assets to risk. According to the Consumption-based Capital Asset Pricing Model (CCAPM) of Lucas (1978) and Breeden (1979), investors are willing to accept low returns in exchange for insurance against consumption risk. However, the risk of the stock market as measured by the covariance of stock returns with aggregate consumption growth is insufficient to justify the large risk premium observed in historical stock return data (Mehra and Prescott (1985)).
In an effort to reconcile the observed high equity premium and low consumption risk within the CCAPM framework, extant literature has explored modifications in investor preferences (Sundaresan (1989), Constantinides (1990), Epstein and Zin (1991), Campbell and Cochrane (1999), Bansal and Yaron (2004), Hansen, Heaton, and Li (2008), and Malloy, Moskowitz, and Vissing-Jorgensen (2009)), implications of incomplete markets (Constantinides and Duffie (1996)), market imperfections (Mankiw and Zeldes (1991), Constantinides, Donaldson, and Mehra (2002)) and alternative ways of measuring aggregate consumption risk (Ait-Sahalia, Parker, and Yogo (2004), Parker and Julliard (2005), Jagannathan and Wang (2007), and Jagannathan, Takehara, and Wang (2007)). Recently, Savov (2010) proposed an alternative approach using annual garbage generation as a measure of consumption, which is twice as volatile as the standard NIPA consumption measure (real per capita personal expenditure on nondurable goods and services) and highly correlated with market returns. As a result, this measure yields a more modest implied relative risk aversion coefficient than those obtained using conventional NIPA consumption expenditures.
In this paper, we propose electricity usage as a new measure of consumption. In modern life, most consumption activities involve the use of electricity. For example, when preparing a meal, food may have been stored in a freezer, defrosted in a microwave oven, cooked in an electric oven, and eaten while watching a sports game aired on TV. Shopping malls may use an extra amount of electricity during the winter holiday season to be open longer hours and to power ornamental lights. We expect this additional electricity consumption to correlate well with the increased general consumption during that time.
While we focus on total electricity usage since it is widely available for many countries, we confirm that in the U.S., excluding industrial electricity usage, which measures production more than consumption, hardly alters our main results. In contrast, the garbage generation, which is estimated with the help of production data, is highly correlated with industrial electricity usage (72.8%) and less correlated with residential and commercial electricity usage (12.0% and 22.3%, respectively).
Compared to existing consumption measures including the garbage generation, electricity consumption has at least two advantages, which make it potentially a better candidate for testing the CCAPM. First, due to technological limitations, electricity cannot be stored: once produced, it has to be either consumed or wasted. Hence, we expect electricity consumption to reflect investors’ consumption activities in a real time manner. In contrast, the popular NIPA consumption data is smoothed with the X-11 seasonal adjustment. As forcefully argued by Ferson and Harvey (1992), “Obviously, one does not purchase goods at seasonally adjusted prices… seasonal addition, consumers derive utility from consumer products not at the time of purchase, but rather continuously over the life of the product. For example, when you buy a new TV set, electricity consumption correctly measures consumption over the life of the TV, while the garbage generation would treat the purchase as a one-time consumption, registering it only when the new owner throws out the box.
