TY - BOOK

T1 - Mispricing, momentum, and market timing

T2 - Essays on stock market puzzles and capital structure decisions

AU - Krupski, Jan

N1 - Doctoral thesis

PY - 2023

Y1 - 2023

N2 - This doctoral thesis comprises one essay on the risk management of momentum strategies and three essays on the implications of skewness preferences on financial markets. Chapter 1 provides an extensive summary and links all projects within the framework of behavioral finance. In Chapter 2 (co-authored with Maik Dierkes), we investigate momentum in stock returns and propose a novel approach to manage the downside risk of momentum strategies. Across markets, momentum is one of the most prominent anomalies and leads to high risk-adjusted returns. However, these returns come at the cost of substantial tail risk as there are short but persistent periods of highly negative returns. Momentum crashes occur in rebounding bear markets, when the momentum portfolio exhibits a negative beta and momentum volatility is high. Based on ex-ante estimates of these risk measures, we construct a crash indicator that effectively isolates momentum crashes. Subsequently, we propose an implementable trading strategy that combines both systematic and momentum-specific risk and more than doubles the Sharpe ratio of the original momentum strategy. Moreover, it outperforms existing risk management approaches over the 1928-2020 period, in sub-samples, and internationally. In Chapter 3 (co-authored with Maik Dierkes and Sebastian Schroen), we address the effects of time-varying skewness preference, referred to as lottery demand, on first-day returns and the long-term performance of initial public offerings (IPOs). Following the identification approach of Dierkes (2013), we measure lottery demand in terms of option-implied probability weighting functions and find a significantly positive impact on first-day returns, tantamount to higher IPO underpricing and more money left on the table. Furthermore, disentangling the effects of lottery demand and cross-sectional expected skewness reveals that IPO returns are particularly driven by the interaction of market-wide lottery demand and asset-specific lottery characteristics. In the long run, firms that went public during periods of high lottery demand perform poorly for up to five years after the IPO. In Chapter 4 (co-authored with Maik Dierkes, Sebastian Schroen, and Philipp Sibbertsen), we perform a simulation-based approach to estimate volatility-dependent probability weighting functions and investigate the impact of probability weighting on the pricing kernel puzzle. We first obtain risk neutral and physical densities from the Pan (2002) stochastic volatility and jumps model and then estimate probability weighting functions according to the identification strategy presented in Chapter 3. Across volatilities, we find pronounced inverse S-shapes. Hence, small (large) probabilities are overweighted (underweighted), and probability weighting almost monotonically increases in volatility, suggesting higher skewness preferences in volatile markets. Moreover, by estimating probabilistic risk attitudes, equivalent to the share of risk aversion related to probability weighting, we shed further light on the pricing kernel puzzle. While pricing kernels estimated from the Pan (2002) model display the typical U-shape documented in the literature, adjusted pricing kernels are monotonically decreasing and thus in line with economic theory. As a result, risk aversion functions are positive throughout wealth levels. Finally, in Chapter 5 (co-authored with Maik Dierkes), we employ idiosyncratic skewness as a proxy for firm-specific mispricing and investigate the impact of market timing on capital structure decisions. Consistent with the market timing theory, idiosyncratic skewness is significantly positively related to equity issues, while the impact on debt issues is negative and less important. Moreover, we find equity issues to be accompanied by debt retirement programs. Challenging the market timing theory, effects are not persistent and vanish after about three years. In line with Alti (2006), our results are therefore consistent with a modified version of the trade-off theory, including market timing as a short-term factor.

AB - This doctoral thesis comprises one essay on the risk management of momentum strategies and three essays on the implications of skewness preferences on financial markets. Chapter 1 provides an extensive summary and links all projects within the framework of behavioral finance. In Chapter 2 (co-authored with Maik Dierkes), we investigate momentum in stock returns and propose a novel approach to manage the downside risk of momentum strategies. Across markets, momentum is one of the most prominent anomalies and leads to high risk-adjusted returns. However, these returns come at the cost of substantial tail risk as there are short but persistent periods of highly negative returns. Momentum crashes occur in rebounding bear markets, when the momentum portfolio exhibits a negative beta and momentum volatility is high. Based on ex-ante estimates of these risk measures, we construct a crash indicator that effectively isolates momentum crashes. Subsequently, we propose an implementable trading strategy that combines both systematic and momentum-specific risk and more than doubles the Sharpe ratio of the original momentum strategy. Moreover, it outperforms existing risk management approaches over the 1928-2020 period, in sub-samples, and internationally. In Chapter 3 (co-authored with Maik Dierkes and Sebastian Schroen), we address the effects of time-varying skewness preference, referred to as lottery demand, on first-day returns and the long-term performance of initial public offerings (IPOs). Following the identification approach of Dierkes (2013), we measure lottery demand in terms of option-implied probability weighting functions and find a significantly positive impact on first-day returns, tantamount to higher IPO underpricing and more money left on the table. Furthermore, disentangling the effects of lottery demand and cross-sectional expected skewness reveals that IPO returns are particularly driven by the interaction of market-wide lottery demand and asset-specific lottery characteristics. In the long run, firms that went public during periods of high lottery demand perform poorly for up to five years after the IPO. In Chapter 4 (co-authored with Maik Dierkes, Sebastian Schroen, and Philipp Sibbertsen), we perform a simulation-based approach to estimate volatility-dependent probability weighting functions and investigate the impact of probability weighting on the pricing kernel puzzle. We first obtain risk neutral and physical densities from the Pan (2002) stochastic volatility and jumps model and then estimate probability weighting functions according to the identification strategy presented in Chapter 3. Across volatilities, we find pronounced inverse S-shapes. Hence, small (large) probabilities are overweighted (underweighted), and probability weighting almost monotonically increases in volatility, suggesting higher skewness preferences in volatile markets. Moreover, by estimating probabilistic risk attitudes, equivalent to the share of risk aversion related to probability weighting, we shed further light on the pricing kernel puzzle. While pricing kernels estimated from the Pan (2002) model display the typical U-shape documented in the literature, adjusted pricing kernels are monotonically decreasing and thus in line with economic theory. As a result, risk aversion functions are positive throughout wealth levels. Finally, in Chapter 5 (co-authored with Maik Dierkes), we employ idiosyncratic skewness as a proxy for firm-specific mispricing and investigate the impact of market timing on capital structure decisions. Consistent with the market timing theory, idiosyncratic skewness is significantly positively related to equity issues, while the impact on debt issues is negative and less important. Moreover, we find equity issues to be accompanied by debt retirement programs. Challenging the market timing theory, effects are not persistent and vanish after about three years. In line with Alti (2006), our results are therefore consistent with a modified version of the trade-off theory, including market timing as a short-term factor.

U2 - 10.15488/13374

DO - 10.15488/13374

M3 - Doctoral thesis

CY - Hannover

ER -