CEO Incentives, Wealth and Risk Aversion

This project provides a comprehensive analysis of how managerial risk aversion affects the design and incentive effect of compensation contracts. Determining whether monetary incentives mitigate conflicts of interest between owners and managers is difficult since the observed contracts depend on personal characteristics that are hard for researchers to measure. Moreover, standard models make simplifying assumptions that are unlikely to hold in reality. For example, models often assume that CEOs are unable to take private action to adjust their exposure to firm risk. The project addresses both issues, using a novel data set on CEOs? personal wealth and leverage as well as the allocation of wealth across different asset classes and individual securities. Our results suggest that wealthier CEOs allocate a larger portion of their outside wealth toward risky assets, such as equity, and away from cash. This is consistent with wealth being a valid proxy for risk aversion. If wealthier CEOs are indeed less risk-averse, we expect them to invest a larger portion of their outside wealth in risky assets. Having established a positive link between wealth and private risk-taking, we then test if incentive strength is sensitive to private wealth. Standard principal-agent theory predicts a negative relationship between risk aversion and incentive strength. Our results confirm this hypothesis - wealthier CEOs receive steeper contracts. We then test recent theoretical work, in which the CEO can diversify or hedge firm risk using her private wealth. The new data allows us to analyze how the investment behavior of the CEO changes in response to incentive changes, controlling for the level of wealth and other key inputs. CEOs with higher exposure to firm risk allocate a larger portion of their private wealth to safe assets, controlling for both the level of wealth and other confounding factors. This suggests that CEOs use their private wealth to adjust the total level of risk. Lastly, we examine the motives behind the observed allocation decisions. This is important since the results up to this point are consistent with both diversification of systematic risk, which would constitute a form of self-indexing, and hedging of idiosyncratic risk, which would undo part of the incentive effect. First, we show that the negative link between incentive strength and risky allocation decisions becomes stronger when the systematic risk component of the firm?s stock returns increases, which suggests that CEOs are mainly trying to diversify their exposure to systematic risk. Second, using instrument-level data, we show that CEOs hold fairly diversified investment portfolios and that the composition of these portfolios is not significantly affected by the provision of incentives, i.e. CEOs do not change the composition of the portfolio in response to incentives. The main decision for the CEO seems to be how much of her total wealth to allocate to the risk-free asset and how much to allocate to a diversified investment portfolio.

Despite its importance, there is limited prior evidence on CEOs' attitudes toward risk and, in particular, the tendency to trade outside the insider portfolio. Our study extends the literature on managerial incentives in several ways. First, we show that private wealth is a valid proxy for risk aversion. This result is significant and previously unreported. Since a person's actual risk aversion is inherently difficult to measure, it is difficult to test the standard model without an accurate proxy for risk aversion. Second, we document that CEOs use their private wealth to adjust the risks embedded in their incentive contracts. This result has implications for the ongoing debate on relative performance evaluation and the effectiveness of using equity-based compensation to incentivize managers. Overall, our new evidence provides essential insights for the design of the optimal compensation contracts.

This project aims to provide a comprehensive analysis of how CEO risk aversion affects the design and incentive effect of compensation contracts. Determining whether monetary incentives optimally address the principal-agent problem is difficult since the observed contracts depend on characteristics that are hard for researchers to measure, such as the risk aversion of the CEO. Moreover, standard models make simplifying assumptions that are unlikely to hold in reality. For example, models generally assume that the CEO is unable to take private action to adjust her exposure to firm risk. The project will address both issues, using a novel data set from Sweden with details on CEO wealth. We will first document how CEOs allocate their private wealth between different asset classes and examine the relation between their investment decisions and wealth. If wealthier CEOs are less risk averse, we expect them to invest a larger portion of their wealth in risky assets. Having established a link between wealth and private risk-taking, we will test if incentive contracts are sensitive to private wealth. Standard principal-agent theory predicts a negative relationship between risk aversion and incentive strength. Lastly, we will test recent theoretical work that allows the CEO to diversify or hedge firm risk using private wealth. Again, we can directly test how the investment behavior of the CEO changes in response to changes in incentives, controlling for the level of wealth and other key inputs. This project has several important implications for practitioners and policymakers: (1) it will examine if risk aversion can explain the strength of monetary incentives; (2) it will shed light on whether stock-based incentives, which are costly for shareholders to provide, are efficient in alleviating the principal-agent problem. The results could provide key insights for the design of efficient managerial compensation contracts, which ultimately lowers firms' funding costs.