The impact of shareholder intervention on overinvestment of free cash flow by overconfident CEOs
Sewon Kwon (email@example.com)
Division of Business Administration
Prof. Sewon Kwon’s recent research published in International Review of Financial Analysis, entitled “The impact of shareholder intervention on overinvestment of free cash flow by overconfident CEOs,” is about the impact of shareholder intervention on investment distortions, which is measured by overinvestment of free cash flow by overconfident CEOs
While positive investment-cash flow sensitivity (hereafter, ICFS) is often considered to represent investment distortions (Jensen, 1986), Malmendier and Tate (2005) argue that the ICFS of firms managed by overconfident CEOs is more likely to constitute investment distortions. Malmendier and Tate (2005) note the potential that positive ICFS may also be observed even when incentives are perfectly aligned, and in such cases, overinvestment of FCF does not necessarily capture investment distortions. These authors argue instead that overconfident CEOs systematically overestimate investment projects, and are thus more likely to distort investments using surplus internal funds.
Despite the large volume of literature on shareholder voices and firm investment, to the best of our knowledge, virtually no literature has specifically explored Malmendier and Tate’s (2005) ICFS type. To address our research question, using this unique definition of investment distortions and the overinvestment measure proposed by Richardson (2006), we divide our sample into four distinct subsamples with different levels of CEO overconfidence and FCF: 1) overconfident CEOpositive FCF, 2) non-overconfident CEOpositive FCF, 3) overconfident CEOnegative FCF, and 4) non-overconfident CEOnegative FCF. We call subsample (1) the distorted investment subsample.
We construct our empirical predictions based on the efficient market and disciplinary threat hypotheses. These theories assume that shareholders intervene effectively to reduce suboptimal investment by processing all relevant information, and that managers consider shareholder voices as a credible threat of engaging the firm in a costly proxy contest. We thus hypothesize that if all else remains equal, a larger voting premium is negatively associated with investment distortions, and that investment distortions will decrease with the arrival of hedge fund activists. However, we do not predict similar effects of shareholder intervention when overinvestment is not necessarily distorted, because in this case, CEOs will not feel threatened by shareholder voices.
To test for these effects, we use two samples. The first is comprised of the merged file of Compustat, Execucomp, CRSP, and OptionMetrics, which results in 8,346 firm-year observations for the 1996-2014 period after excluding missing values (hereafter, the voting premium sample). We choose this sample to estimate the voting premium by using derivative trading data from OptionMetrics. The second is the merged file of Compustat, Execucomp, CRSP, and 13D filings. In this sample, we do not use the voting premium variable, and so do not require that firms be included in OptionMetrics because this requirement reduces our sample size. Instead, to reduce any systematic differences in observable firm characteristics between activist targets and non-targets, we select 354 matched pairs (2,173 firm-year observations) for 1996-2014 based on the ex ante probability of 13D filings (hereafter, the hedge fund activism sample). In all analyses, we capture overinvestment, FCF, and CEO overconfidence by adopting the frameworks proposed by Richardson (2006) and Malmendier and Tate (2005).
Using the voting premium sample, we first find that the positive ICFS becomes weaker and insignificant as the voting premium increases, signaling shareholders’ deterrent role in curbing overinvestment. Furthermore, the voting premium is negatively associated with overinvestment in the distorted investment subsample. This implies that the deterrent effect on ICFS arises because shareholder control rights mitigate overinvestment, particularly in firms with positive FCF and overconfident CEOs simultaneously. Interestingly, and consistent with the disciplinary threat hypothesis, this deterrent effect persists even after we explicitly exclude the effects of actual voting events, control contests, and shareholder litigation. These results are also robust to the adoption of an instrumental variables (IV) approach.
Using the hedge fund activism sample, we also find that the positive ICFS becomes weaker and insignificant when hedge fund activists enter target firms. Further analysis using a difference-in-differences (DID) matching estimator reveals that overinvestment drops significantly during the one- and five-year windows subsequent to hedge fund activism, but only in the distorted investment subsample. These results affirm that the deterrent effect of hedge fund activism is driven by the dynamic reduction in overinvestment of firms that hold positive FCF and are run by overconfident CEOs.
Next, we seek to attenuate concerns over our assumption that the ICFS of firms managed by overconfident CEOs captures potential investment distortions. If firms in the distorted investment subsample generate good ex post financial performance (e.g., earnings on assets), shareholders will not need to curb such overinvestments, and we may not be able to define this subsample as potential investment distortions. However, using a system generalized method of moments (GMM) approach, we find that Richardson’s (2006) overinvestment is negatively associated with firm performance over both the short and long run, mostly when CEOs are overconfident and firms hold surplus cash flow simultaneously. In contrast, overinvestments in other subsamples do not show consistently poor performance. These findings lend support to our classification of potential investment distortions.
Overall, we interpret our findings as implying that shareholder intervention does not uniformly deter all overinvestments, but is particularly effective at alleviating the investment distortions, which we capture using overinvestment of FCF by overconfident CEOs. The effects of shareholder intervention acting as a credible threat to CEOs are significantly incremental to those of actual voting events, control contests, and shareholder litigation.
Our study makes several contributions to the existing literature. First, it builds on research exploring the causes and mitigating factors of ICFS. As noted earlier, traditional ICFS does not necessarily capture investment distortions, but that of firms managed by overconfident CEOs is more likely to do so. However, since Malmendier and Tate (2005), virtually no studies have focused on this unique ICFS type. To the best of our knowledge, this paper is the first to explicitly explore the mitigating factors of Malmendier and Tate’s (2005) investment distortions. In particular, while Richardson (2006) provides some evidence that activist shareholders mitigate overinvestment using the traditional ICFS, this study complements his finding by documenting that shareholder intervention is most effective at deterring overinvestment caused by overconfident CEOs.
Second, this paper adds to the literature that explores the role of shareholder intervention in curbing investment distortions. Empirical evidence on this topic and its interpretation have been mixed at least partly because of ambiguity over the definition of investment distortions. For example, one strand of research suggests that informed shareholder monitoring is effective at mitigating the problems of suboptimal investment. Others argue the deterrent role of shareholders is ineffective due to their short-term investment horizon. In this paper, we show that shareholder intervention does not uniformly deter all overinvestments. We believe that the adoption of Malmendier and Tate’s (2005) approach to capture potential investment distortions may help resolve the mixed results and interpretations in the literature.
Jensen, M. C., & Meckling, W. H. (1976). Theory of the firm: Managerial behavior, agency costs and ownership structure. Journal of Financial Economics, 3(4), 305–360.
Malmendier, U., & Tate, G. (2005). CEO overconfidence and corporate investment. The Journal of Finance, 60(6), 2661–2700.
Richardson, S. (2006). Over-investment of free cash flow. Review of Accounting Studies, 11(2-3), 159–189.
* Related Article
Sewon Kwon, Jae Hwan Ahn, Gi H. Kim, The impact of shareholder intervention on overinvestment of free cash flow by overconfident CEOs, International Review of Financial Analysis, Volume 75, May 2021, 101751