«Investment, Idiosyncratic Risk, and Ownership ∗ Vasia Panousi Dimitris Papanikolaou October 21, 2009 Abstract We ﬁnd a signiﬁcant negative ...»
JOINT USC FBE FINANCE SEMINAR &
MACROECON & INT'L FINANCE WORKSHOP
presented by Dimitris Papanikolaou
FRIDAY,Oct. 23, 2009
10:30 am – 12:00 pm, Room: ACC-201
Investment, Idiosyncratic Risk, and Ownership
Vasia Panousi Dimitris Papanikolaou
October 21, 2009
Abstract We ﬁnd a signiﬁcant negative eﬀect of idiosyncratic stock-return volatility on investment. We address the endogeneity problem of stock return volatility by instrumenting for volatility with a measure of a ﬁrm’s customer base concentration. We propose that the negative eﬀect of idiosyncratic risk on investment is partly due to managerial risk aversion, and ﬁnd that the negative relationship between idiosyncratic uncertainty and investment is stronger for ﬁrms with high levels of insider ownership. Several mechanisms can mitigate this eﬀect namely the use of option-based compensation and shareholder monitoring. We ﬁnd that the investment-idiosyncratic relationship is weaker for ﬁrms that make use of option-based compensation, and insider ownership does not matter for ﬁrms primarily held by institutional investors.
∗ Federal Reserve Board, firstname.lastname@example.org, and Kellogg School of Management, d-papanikolaou @kellogg.northwestern.edu. We would like to thank Snehal Banerjee, Gadi Barlevy, Luca Benzoni, Paco Buera, Jeﬀ Campbell, Andrea Eisfeldt, Jiro Kondo, Guido Lorenzoni, Byron Lutz, Jae Sim, Harald Uhlig, Toni Whited, Egon Zakrajsek, and the seminar participants at the Federal Reserve Board, the Chicago Fed, the Kellogg School of Management, Michigan State University, and MIT for very useful comments and discussions. The views presented are solely those of the authors and do not necessarily represent those of the Board of Governors of the Federal Reserve System or its staﬀ members. Dimitris Papanikolaou would like to thank the Zell Center for ﬁnancial support.
Introduction We ﬁnd a signiﬁcant negative relationship between the volatility of idiosyncratic risk and the investment of publicly traded ﬁrms in the United States. We provide evidence in support of a causal relationship by instrumenting for a ﬁrm’s idiosyncratic volatility with the concentration of the ﬁrm’s customer base. The relationship is stronger for ﬁrms with high levels of insider ownership andweaker for ﬁrms with a convex executive compensation schedule. We interpret this negative relationship as the result of managerial risk aversion, in that managers may be reluctant to undertake new projects when idiosyncratic risk is high.
Neoclassical ﬁnance theory predicts that only the systematic component of risk is relevant for investment decisions, since ﬁrm owners are diversiﬁed and managers maximize shareholder value. Nonetheless, and consistent with empirical evidence, agency theory suggests that the managers who undertake investment decisions usually hold a substantial stake in the ﬁrm for incentive purposes. Consequently, managers may underinvest, since high idiosyncratic risk projects increase the volatility of their consumption stream. If this is the case, then the component of risk which would disappear under portfolio diversiﬁcation will be relevant for investment decisions. In fact, proponents of option-based compensation have used this argument to justify providing executives with some measure of downside protection, as a compromise between supplying incentives and mitigating risk-averse behavior. Alternatively, strong monitoring may dampen this wedge between managers’ and shareholders’ valuation of investment projects.
Our ﬁrst main result concerns the negative relationship between idiosyncratic risk and the investment of public ﬁrms in the US. We decompose stock-return volatility into a systematic and a ﬁrm-speciﬁc component, and we use the latter as our measure of idiosyncratic risk.
We ﬁnd that higher idiosyncratic volatility is associated with lower investment. Moreover, higher idiosyncratic volatility is associated with lower payout and increased cash holdings, evidence consistent with a precautionary savings motive.
Two concerns prevent us from interpreting the negative correlation between investment and idiosyncratic volatility as a causal relationship. First, idiosyncratic volatility could be a proxy for a ﬁrm’s growth opportunities, inducing omitted variable bias, since Tobin’s Q is measured with error. Second, investment decisions may aﬀect stock return volatility, since they alter the mix of growth options and assets in place, and the former are more volatile than the latter. We address the ﬁrst concern by considering two alternative measures of growth opportunities: the ﬁrst is constructed from analyst forecasts based on Bond and Cummins (2004), and the second is constructed from stock return correlations based on Kogan and Papanikolaou (2009). As a separate exercise, we also follow Erickson and Whited (2000) and deal with measurement error in Tobin’s Q directly in the estimation, We address the endogeneity of stock-return volatility by instrumenting for it with a measure of how concentrated a ﬁrm’s customer base is. Firms that sell to few customers cannot diversify demand shocks for their product across their customers, and will thus be riskier. To the extent that some of these demand shocks are idiosyncratic, the ﬁrm will face higher idiosyncratic risk. Our identiﬁcation assumption is that, controlling for the level of sales, investment decisions do not aﬀect how concentrated a ﬁrm’s customer base is. We ﬁnd that, in both cases, idiosyncratic volatility remains a statistically signiﬁcant predictor of investment and thus conclude that the relationship is likely to be causal.
After concluding that idiosyncratic volatility does not proxy for unobserved growth opportunities, we proceed to explore the implications of agency theory. If the eﬀect is due to managerial risk aversion, we expect it to be stronger in ﬁrms with higher levels of insider ownership and weaker in ﬁrms that make more use of options in their compensation schemes.
Indeed, we ﬁnd that the negative relationship between investment and idiosyncratic risk is strongest for ﬁrms where insiders hold a higher fraction of the shares. This eﬀect is partially mitigated by convex compensation contracts: controlling for the level of insider ownership, ﬁrms which provide compensation contracts that are more convex, and therefore increase in value with volatility, display investment behavior that is not sensitive to idiosyncratic risk.
In addition, we explore a related prediction, which is related to costs of external ﬁnance.
Froot, Scharfstein and Stein (1993), among others, show that ﬁrms who face convex costs of external ﬁnance may behave in a risk-averse fashion, even if managerial preferences are risk-neutral. We use two variables to measure the likelihood of a ﬁrm being ﬁnancially constrained: the Whited and Wu (2006) index and whether a ﬁrm is assigned a credit rating by Standard and Poor’s. We ﬁnd that the eﬀect is indeed stronger for ﬁrms that are more likely to be constrained.
If the discount rate managers use to value a project depends on its idiosyncratic risk, absent any other frictions, it will lead to ineﬃcient investment decisions from the shareholders’ perspective. An additional way that shareholders can prevent this destruction in value is through increased monitoring. Furthermore, monitoring may be easier or more eﬀective when institutions rather than households own the majority of the ﬁrm. The former have more expertise and since they typically hold larger shares, suﬀer less from the free-rider problem. Thus, we expect the eﬀect to be stronger for ﬁrms with low levels of institutional ownership. Indeed, the level of insider ownership matters for the sensitivity of investment to idiosyncratic risk only when institutional ownership is low.
Finally, we are concerned that insider ownership is endogenous as it arises from an optimal contract. Thus, it is possible that the eﬀect we are identifying comes not from insider ownership per se, but some other ﬁrm characteristic that insider ownership responds to.
One such candidate is the degree of industry competition. A competitive product market will exercise higher discipline upon the manager to exert eﬀort, and thus there may be less need to provide incentives through ownership. Moreover, as Caballero (1991) shows, the degree of product market competition can aﬀect the relationship between investment and uncertainty. Given the degree investment irreversibility facing ﬁrms, the relationship between investment and risk should be more negative for less competitive ﬁrms. We do not ﬁnd any evidence that this is the case, however. Using three diﬀerent proxies for the degree of investment irreversibility, ﬁrms that operate in more ore less concentrated product markets display similar levels of sensitivity of investment to idiosyncratic volatility.
The rest of the paper is organized as follows. Section 1 reviews the related literature;
Section 3 shows the negative relationship between idiosyncratic risk and investment; Section 4 addresses concerns about omitted variables and reverse causality; Section 5 shows how the eﬀect varies with levels of insider ownership, convexity of executive compensation schemes and likelihood of ﬁnancial constraints; Section 6 examines the interaction with institutional ownership; Section 7 examines an alternative explanation where product market competition is the underlying mechanism; Section 8 concludes.
1 Related literature On the theoretical front, the sign of the relationship between investment and total uncertainty facing a ﬁrm has been examined mostly in the real options literature. The conclusions are rather ambiguous, and the the sign of the eﬀect of uncertainty on investment depends, among other things, on assumptions about the production function, the market structure, the shape of adjustment costs, the importance of investment lags and irreversibilities. An incomplete list includes Oi (1961), Hartman (1972), Abel (1983), Caballero (1991), Abel, Dixit, Eberly and Pindyck (1996), and Abel and Eberly (1999). While the previous papers focused on the ﬁrm’s partial equilibrium problem, Angeletos (2007), Bloom (2009), Nakamura (1999) and Saltari and Ticchi (2007), among others, investigate the general equilibrium eﬀects of an increase in uncertainty on investment and how this depends on the coeﬃcients of risk aversion and intertemporal substitution. In most of these papers, no distinction is made between idiosyncratic and systematic uncertainty. More recently, a number of papers explore the eﬀect of managerial risk aversion and idiosyncratic risk on the timing of investment decisions in real option models [Hugonnier and Morellec, 2007b; Hugonnier and Morellec, 2007a; Henderson, 2007; Miao and Wang, 2007; Chen, Miao and Wang, 2009].
On the empirical front, a number of studies use the volatility of stock returns as a measure of uncertainty, and explore its eﬀect on investment. An incomplete list includes Brainard, Shoven, Weiss, Cagan and Hall (1980), Leahy and Whited (1996), Henley, Carruth and Dickerson (2003), Bond and Cummins (2004), Cummins, Hassett and Oliner (2006), Bulan (2005), Bloom, Bond and Reenen (2007), and Baum, Caglayan and Talavera (2008).
The conclusions of this literature is mixed both as to sign and as to the signiﬁcance of the investment-uncertainty relationship, and they appear to be somewhat sensitive to the estimation method. Moreover, in most of the papers above, no theoretical distinction is made between idiosyncratic and systematic uncertainty. We complement this literature by addressing the issue that the volatility of stock returns in endogenous, which in principle depends on the ﬁrm’s investment policy.
Furthermore, our paper complements the ﬁndings of Himmelberg, Hubbard and Love (2002), who document high insider ownership shares in publicly traded ﬁrms worldwide.
They argue that, absent investor protection legislation, ﬁrm insiders are given endogenous ownership as a commitment device to not steal from shareholders. However, these incentives are costly, because they force insiders to bear idiosyncratic risk. They test their structural model using cross-country data, and monthly country stock returns for their measure of market-based idiosyncratic volatility. They ﬁnd that countries like the US, where investor protection is high, are characterized by a lower level of insider ownership, a smaller idiosyncratic risk premium, and a steady-state capital stock closer to the no-frictions level.
In this section, we propose a simple two-period model that demonstrates how idiosyncratic risk can aﬀect capital investment in an eﬃcient capital market, even in the absence of adjustment costs or other investment frictions. We
from the latter not because
we think that they are not important, but because we are interested in a diﬀerent channel:
investment decisions are taken by managers who hold undiversiﬁed stakes, and thus are reluctant to undertake high idiosyncratic ﬁrm projects, even though the shareholders are perfectly diversiﬁed.
A ﬁrm starts with cash C at t = 0 and produces output at t = 1 according to
where e is managerial eﬀort, K is installed capital and X ∼ N (µ, σ 2 ) is a shock that is speciﬁc to the ﬁrm.
The manager owns a fraction λ of the ﬁrm, and the remaining 1 − λ shares are held by shareholders who are risk averse but hold the market portfolio. We assume that, for incentives to have any bite, the manager cannot diversify his stake in the ﬁrm. The manager
derives utility from consumption (c0, c1 ) and disutility from eﬀort (e):
The manager’s contract consists of a choice of ownership λ and an initial transfer T.
Given the contract, the manager will choose how much to invest in capital K, how much eﬀort to provide e and how much to save in the riskless asset, B to maximize (2), subject to (1) and the two budget constraints