Capital Utilization, Market Power, and the Pricing of Investment Shocks∗ Lorenzo Garlappi† Zhongzhi Song‡ University of British Columbia Cheung Kong GSB November 20, 2016 Journal of Financial Economics, forthcoming ∗ We are grateful to Toni Whited (the editor) and an anonymous referee for extensive comments and constructive suggestions. We thank Paul Beaudry, Jules van Binsbergen, Murray Carlson, Jack Favilukis, Adlai Fisher, Ron Giammarino, Burton Hollifield, Nengjiu Ju, Jiri Knesl, Xiaoji Lin, Stavros Panageas, Dimitris Papanikolaou, Carolin Pflueger, Loris Rubini, Elena Simintzi, Raman Uppal, Stijn Van Nieuwer- burgh, Amir Yaron, Harold Zhang, and seminar participants at the Central Bank of Mexico, Shanghai Advanced Institute of Finance, Texas A&M University, Tsinghua University, the University of British Columbia, the University of Lausanne, the University of Maryland, the University of Rochester, and con- ference participants at the CAPR Workshop on Production-based Asset Pricing at BI Norwegian Business School, China International Conference in Finance, Finance UC 6th International Conference at the Uni- versidad Católica de Chile, the 2nd Meeting of the Macro Finance Society at NYU, and Summer Institute of Finance for valuable comments. We are particularly grateful to Haibo Jiang for excellent research assistance. † Department of Finance, Sauder School of Business, University of British Columbia, 2053 Main Mall, Vancouver, BC V6T 1Z2, Canada, E-mail: lorenzo.
‡ Cheung Kong Graduate School of Business, 3F, Tower E3, Oriental Plaza, 1 East Chang An Avenue, Beijing 100738, China. E-mail: zzsong@ckgsb. Capital Utilization, Market Power, and the Pricing of Investment Shocks Abstract Capital utilization and market power crucially affect asset prices in an economy exposed to shocks that improve real investment opportunities through capital-embodied technological innovations. We embed these two mechanisms in a standard general equilibrium model and show that (i) the price of risk for investment shocks is negative under fixed capital utilization, but positive under sufficiently flexible capital utilization, and (ii) the equity return exposure to investment shocks is negative under perfect competition, but positive under high market power.
We further show that, high market power, persistent components in technology growth, and a strong preference for early resolution of uncertainty are jointly important to quantitatively match the observed equity risk premium. JEL Classification Codes: E22; G12; O30 Keywords: Investment shocks; Capital utilization; Market power; Risk premium 1 1 Introduction In this paper we argue that two fundamental economic mechanisms: capital utilization and mar- ket power, have important implications for the pricing of assets in financial markets. Intuitively, since flexibility in capital utilization affects how firms adjust output in response to technology changes, and the degree of market power affects how firms benefit from technology improvements, these mechanisms should also affect firms’ market values. Both mechanisms have been widely studied and featured prominently in many macroeconomic models.
For example, the theory of business cycles relies on variable capital utilization for understanding comovement across macroe- conomic aggregates.1 The endogenous growth literature relies on market power and monopoly rents from innovation for understanding aggregate economic growth.2 Surprisingly, the vast ma- jority of production-based asset pricing models in finance ignore these mechanisms and assume instead that capital utilization is fixed and firms are fully competitive.3 We fill this gap and show that these economically motivated mechanisms can have both qualitative and quantitative effects on asset prices. To highlight the importance of capital utilization and market power for asset pricing we focus our analysis on the pricing of a specific form of shocks to the economy, commonly referred to as investment-specific technology (IST) shocks or, in short, investment shocks. These shocks are widely used in economic models as important determinants of growth and business-cycle fluctu- ations. Unlike neutral total factor productivity (TFP) shocks that directly affect consumption, investment shocks are embodied in new capital and therefore affect consumption only through investment.
As we show in the paper, this distinction between IST and TFP shocks turns out to be critical for understanding the effect of capital utilization and market power on asset prices. In particular, these mechanisms have both qualitative and quantitative effects on the pricing of IST shocks, but they only have a quantitative impact on the pricing of TFP shocks. While we 1 See, among many others, Lucas (1970), Greenwood, Hercowitz, and Huffman (1988), Kydland and Prescott (1988), and Jaimovich and Rebelo (2009). The Federal Reserve estimates large procyclical variations in capacity utilization for the U.
industrial sector (see the Federal Reserve’s G. The typical variation in capacity utilization over business cycles is around 10%. During the recent 2008-09 Great Recession, the capacity utilization rate drops from around 80% in January of 2008 to 67% in June of 2009, and then bounces back to 79% in June of 2014. Similar fluctuations are observed over business cycles in other periods.
Aghion and Howitt (1998) and Acemoglu (2010) provide excellent surveys of the endogenous growth literature. The existence of monopoly power in the U. economy is a well-established fact. In a seminal study of the relation between market structure and macroeconomic fluctuations, Hall (1986, p.
286) concludes that: “These findings support the view of the monopolistic competition originally proposed by Edward Chamberlin. Through product differentiation or geographical separation, firms have market power in their own market.” 3 Monopolistic competition is featured in recent models that study the asset pricing implications of endogenous growth, see, e., Kung and Schmid (2015), Kung (2015), and Bena, Garlappi, and Grüning (2016). To the best of our knowledge, we are the first to highlight explicitly the importance of market power on the pricing of technology shocks. Moreover, while these existing models focus on total factor productivity, our focus is on the pricing of investment shocks.
2 analyze both types of shocks, our discussion in the paper will emphasize the novel aspects that we can learn from the study of IST shocks. The key insights from our analysis are: (i) the flexibility in capital utilization affects mainly the price of risk for IST shocks, and (ii) the degree of market power affects mainly the exposure of equity returns to IST shocks. Specifically, the price of risk for IST shocks is positive under flexible capital utilization, but negative under fixed capital utilization. Similarly, the equity return exposure to IST shocks is positive under high market power, but negative under perfect competition.
Our quantitative analysis further shows that, high market power, “long-run risks” in technology growth, and a strong preference for early resolution of uncertainty are jointly important for quantitatively matching key macroeconomic and asset pricing moments. Our model builds on a standard two-sector real business cycle model with investment shocks in which we allow for: (i) flexible capital utilization, which we model as variable capital utilization rates that affect both the output and the depreciation cost of equipment (see Jaimovich and Rebelo (2009)), and (ii) market power, which we model as monopolistic competition among intermediate goods producers (see Dixit and Stiglitz (1977)). Flexible capital utilization allows us to study the effect of IST shocks not only on the accumulation of new capital, but also on the utilization of existing old capital. Because we focus on the pricing of financial assets, the household’s preferences crucially affect our results.
Therefore, we study the effects of households’ attitude towards the distribution of consumption over time and across states separately, by assuming that preferences are recursive (see Kreps and Porteus (1978) and Epstein and Zin (1989)). We solve for an equilibrium allocation in this economy and derive implications for the price of risk for technology shocks and the risk premium of the aggregate market portfolio. Our first result is that flexibility in capital utilization affects mainly the price of risk for IST shocks. To illustrate the main intuition, consider the simpler case of time-separable constant relative risk aversion (CRRA) preferences, where households’ marginal utility depends only on current consumption and not on future utility.
In this case, if the consumption smoothing desire is not strong, fixed capital utilization implies a negative market price of risk for IST shocks. A positive IST shock, by increasing the productivity of the investment sector, diverts labor from the consumption to the investment sector. The drop in labor in the consumption sector induces a drop in current consumption and an increase in the household’s marginal utility, leading to a negative price of risk for IST shocks. In contrast, when capital utilization is flexible and endogenously determined in equilibrium, the market price of risk for IST shocks can be positive.
With variable capital utilization, a positive IST shock makes capital cheaper to replace and hence increases the utilization of existing capital at the expense of faster capital depreciation. The increase in capital utilization counterbalances the decline in labor supply. When capital utilization is sufficiently responsive to IST shocks, the capital utilization effect dominates, causing 3 a net increase in consumption, a decline in marginal utility, and hence a positive price of risk for IST shocks. As we discuss below, a strong consumption smoothing desire has a similar effect as flexible capital utilization on the price of risk for IST shocks.
Our second result is that the degree of market power affects mainly the exposure of equity returns to IST shocks, or, in short, market IST beta. Specifically, when firms are perfectly competitive, the market IST beta is negative. A positive IST shock implies a drop in the price of capital. Since in perfectly competitive markets a firm’s value is determined by the replacement cost of its capital stock (see Hayashi (1982)), a drop in the price of capital good leads to a drop in the firm value.
In contrast, when firms retain some degree of market power, firms’ value includes also monopoly rents from markups. Following a positive IST shock, the increase in rents originating from lower investment cost can more than compensate the decline in value of installed capital. That is, when firms have market power, the market IST beta can be positive. Our third result is that the asset pricing implications of capital utilization and market power discussed above are crucially shaped by households’ preferences.
In our model, households have Epstein-Zin preferences and therefore their marginal utility depends on both current consumption and future utility. If households have preferences for early resolution of uncertainty, a positive IST shock leads to a higher future utility and a lower marginal utility, thereby resulting in a positive price of risk for IST shocks. The opposite is true if households have preference for late resolution of uncertainty. Our model allows us to study the interaction of this preference effect with the effects of capital utilization and market power discussed above.
In particular, we show that capital utilization flexibility can ‘undo’ some of the effects of preferences on IST pricing. For example, under preferences for late resolution of uncertainty, flexible capital utilization can change the price of risk of IST shocks from negative to positive. Moreover, market power leads to positive market IST betas only when the elasticity of intertemporal substitution (EIS) is sufficiently high. With low EIS, the strong wealth effect leads to a decline in labor supply and an increase in firms’ labor cost following a positive IST shock.
This in turn leads to a drop in firm value and hence a negative market IST beta. Our final result is that the effects of capital utilization, market power, and preferences are also quantitatively important. We first document the existence of a small and persistent component in IST shocks, complementing the evidence on the existence of a similar “long-run risk” component in TFP shocks (e. We then show that long-run risks in technology growth, high market power, along with preference for early resolution of uncertainty (high EIS), can quantitatively match the key macroeconomic and asset pricing moments observed in the U.
In particular, our calibrated benchmark model generates an annual log risk-free rate of 0.50% and an annual log equity risk premium of 5. Without flexible utilization and market power, the same parametrization delivers a counterfactual negative equity risk premium.