The Delisting Bias in CRSP's Nasdaq Data and Its Implications for the Size Effect

We investigate the bias in CRSP's Nasdaq data due to missing returns for delisted stocks. We find that the missing returns are large and negative on average, and that delisted stocks experience a substantial decrease in liquidity. We estimate that using a corrected return of −55 percent for missing performance-related delisting returns corrects the bias. We revisit previous work which finds a size effect among Nasdaq stocks. After correcting for the delisting bias, there is no evidence that there ever was a size effect on Nasdaq. Our results are inconsistent with most risk-based explanations of the size effect.     

The leverage effect puzzle: Disentangling sources of bias at high frequency

The leverage effect refers to the generally negative correlation between an asset return and its changes of volatility. A natural estimate consists in using the empirical correlation between the daily returns and the changes of daily volatility estimated from high frequency data. The puzzle lies in the fact that such an intuitively natural estimate yields nearly zero correlation for most assets tested, despite the many economic reasons for expecting the estimated correlation to be negative. To better understand the sources of the puzzle, we analyze the different asymptotic biases that are involved in high frequency estimation of the leverage effect, including biases due to discretization errors, to smoothing errors in estimating spot volatilities, to estimation error, and to market microstructure noise. This decomposition enables us to propose novel bias correction methods for estimating the leverage effect.     

Size effect,book-to-market effect,and survival

Previous studies find that small stocks have higher average returns than large stocks, and the difference between the returns can not be accounted for by the systematic risk, β. In my analysis of Compustat and CRSP data from 1976 to 1995, and simulation experiments based on the data, I find the size effect can be largely explained by data truncation that is caused by survival. Small stocks’ returns are more volatile, and small stocks are more likely to go bankrupt and less likely to meet the stock exchanges’ minimum capitalization requirements for listing. As a result, they are more likely to drop out of the sample. Including small stocks that do well and excluding those that do poorly, ex post, gives rise to higher returns for small-size portfolios. I conclude that the size effect is largely a spurious statistical inference resulting from survival bias, not an asset pricing ‘anomaly’.     

Cross-listing and subsequent delisting in foreign markets

Employing a sample of stocks cross-listed and subsequently delisted from foreign markets, we examine the consequences of delisting to investors in terms of price, risk, and liquidity. We also provide a direct comparison between the firm's performance after a foreign cross-listing and after its subsequent delisting. We find a positive cross-listing and negative delisting effect on stock price, both of which dissipate in the long run. No significant changes in the market risk are found for either event. Foreign cross-listing and delisting are associated with increasing and decreasing long term trading volume respectively. Further analysis reveals that firms delist in response to low host market return and low firm trading volume in the host market. The changes in liquidity and market risk from delisting relate those from cross-listing. Finally, our results show that the bonding hypothesis fails to explain the listing premium and the delisting loss.     

Dissecting the Long-Term Performance of the Chinese Stock Market

Domestically listed Chinese (A-share) firms have lower stock returns than externally listed Chinese, developed, and emerging country firms during 2000 to 2018. They also have lower net cash flows than matched unlisted Chinese firms. The underperformance of both stock and accounting returns is more pronounced for large A-share firms, while small firms show no underperformance along either dimension. Investor sentiment explains low stock returns in the cross-country and within-A-share samples. Institutional deficiencies in listing and delisting processes and weak corporate governance in terms of shareholder value creation are consistent with the underperformance in stock returns and net cash flows.     

Carry Trades,Order Flow,and the Forward Bias Puzzle

We investigate the relation between foreign exchange (FX) order flow and the forward bias. We outline a decomposition of the forward bias according to which a negative correlation between interest rate differentials and order flow creates a time‐varying risk premium consistent with that bias. Using 10 years of data on FX order flow, we find that more than half of the forward bias is accounted for by order flow—with the rest being explained by expectational errors. We also find that carry trading increases currency‐crash risk in that order flow generates negative skewness in FX returns.     

Equity issues when in distress

We investigate the role of financial distress in the seasoned equity market. We find that distressed firms comprise about 40% of SEOs and these distressed issuers have worse abnormal announcement returns than non‐distressed issuers. Stock return volatility is an important determinant for announcement returns for non‐distressed SEO issuers but not for distressed SEO issuers. Signals of firm quality are associated with better announcement returns, larger issues, increased investment, improved operating performance, and lower likelihood of delisting for distressed SEO firms as compared to non‐distressed firms. Our findings suggest equity finance is valuable for financially distressed firms with strong growth prospects.     

Trading volume and stock returns: A meta-analysis

I examine 468 estimates on the relationship between trading volume and stock returns reported in 44 studies. I study publication bias together with Bayesian and frequentist model averaging to explain the heterogeneity in the estimates. The results yield three key conclusions. First, publication bias distorts the findings of the primary studies. Second, the predictability of stock returns varies with different markets and stock types. Third, different data characteristics, structural variations and methodologies used drive the heterogeneity in the results of the primary articles. In particular, one should be cautious when using monthly data or VAR models.     

Endogenous Events and Long-Run Returns

We analyze event abnormal returns when returns predict events.In fixed samples, we show that the expected abnormal returnis negative and becomes more negative as the holding periodincreases. Asymptotically, abnormal returns converge to zeroprovided that the process of the number of events is stationary.Nonstationarity in the process of the number of events is neededto generate a large negative bias. We present theory and simulationsfor the specific case of a lognormal model to characterize themagnitude of the small-sample bias. We illustrate the theoryby analyzing long-term returns after initial public offerings(IPOs) and seasoned equity offerings (SEOs).     

INTANGIBLE ASSETS,BOOK‐TO‐MARKET,AND COMMON STOCK RETURNS

I examine two anomalies where the Fama and French three‐factor model fails to adequately explain monthly industry and index returns. Both anomalies are consistent with a bad model problem where the book‐to‐market factor introduces a negative bias in the intercepts. I propose the intangibles model as an alternative where the three‐factor model is known to have difficulty. This alternative model, which replaces the book‐to‐market factor with zero investment portfolio returns based on prior investments in intangible assets, is well specified in random samples, has comparable power, and fully explains both anomalies.     

Prospect theory and the risk-return paradox: some recent evidence

There is extensive evidence indicating a negative risk–return relation when a firm’s performance is measured based on accounting measures such as return on asset (ROA) and return on equity (ROE). Previous studies show that the risk-return paradox can be explained by the prospect theory, which predicts that managers’ risk attitudes are different for firms of different performances. However, those studies mostly use earlier data from the COMPUSTAT database, which suffers from a survivorship bias. Failure to account for delisting firms may understate the risk–return relation. We reexamine the mixture of risk-seeking and risk-averse behaviors based on an updated 20-year sample period that is free from the survivorship problem. Interestingly, our results show stronger and robust evidence supporting the prospect theory during the period from 1984 to 2003.     

New evidence on the announcement effect of convertible and exchangeable bonds

This study investigates the announcement and issuance effects of offering convertible bonds and exchangeable bonds using data for the Swiss and German markets during January 1996 and May 2003. The analysis suggests that announcement effects of convertible bonds and exchangeable bonds are associated with significantly negative abnormal returns. German firms exhibit a stronger reaction than Swiss firms, possibly for institutional reasons. We also investigate the effect of the market return of the announcement effect and find that the negative abnormal returns are significantly more pronounced when previous market returns have been negative. Furthermore, we analyze the relation between the announcement effects and equity components by controlling for the equity signal sent to the market. We find the size of the equity component of an issue to have a strong influence on the announcement effect for convertible but not for exchangeable securities and offer an explanation for this difference.     

An Investigation of Market Microstructure Impacts on Event Study Returns

We investigate the importance of bid-ask spread-induced biases on event date returns as exemplified by seasoned equity offerings by NYSE listed firms. We document significant negative return biases on the offering day which explain a large portion of the negative event date return documented in the literature. Buy-sell order flow imbalance is prominent around the offering and induces a relatively large spread bias. If order imbalances are suspected, the researcher can use returns calculated from the midpoint of the closing bid and ask quotes instead of returns calculated from closing transaction prices to avoid this return bias.     

Cycles in the IPO market

We develop a model in which time-varying real investment opportunities lead to time-varying adverse selection in the market for IPOs. The model is consistent with several stylized facts known about the IPO market: economic expansions are associated with a dramatic increase in the number of firms going public, which is in turn positively correlated with underpricing. Adverse selection is procyclical in the sense that dispersion in unobservable quality across firms should be more pronounced during booms. Taking the premise that uncertainty is resolved (and thus private information revealed) over time, we test this hypothesis by looking at long-run abnormal returns and delisting rates. Consistent with the model, we find (a) greater cross-sectional return variance, and (b) higher incidence of delisting for hot-market IPOs.     

Why do companies delist voluntarily from the stock market?

We analyse the motives and market valuation of voluntarily delisting from the London Stock Exchange. We show that firms that delist voluntarily are likely to have come to the market to rebalance their leverage rather than to finance their growth opportunities. During their quotation life, their leverage and insider ownership remained very high, they did not raise equity capital, and their profitability, growth opportunities, and trading volume declined substantially. They also generate negative pre-event and announcement date excess returns. These results hold even after controlling for agency, asymmetric information, and liquidity effects, and suggest that firms delist voluntarily when they fail to benefit from listing. Overall, these firms destroyed shareholder value and they should not have come to the market.     

Trading Returns for the Weekend Effect Using Intraday Data

The existence of the weekend effect has been documented as early as 1885. This paper examines whether the serial dependence in returns around weekends and the magnitude of negative Friday returns can be used to produce superior trading returns. We find some success for this endeavor after accounting for transaction costs (including the bid/ask spread), especially when trading is confined to weekends for which there are large negative Friday returns and to positions opened on Friday afternoons. The effect of stocks trading ex-dividend on Mondays does not appear to bias our results.     

Stock market returns and annuitization

I investigate the strong negative relation between recent stock returns and the annuitization of retirement savings using a novel data set with over 100,000 actual payout decisions. After controlling for several standard explanations (e.g., wealth effects), I present evidence supporting naïve beliefs and extrapolation from past returns. The effect of recent returns on annuitization dramatically increases with age, confirming that the elderly rely most heavily on recent information. My results provide insights into how beliefs are formed in old age and have implications for the design of public policies seeking to promote annuitization.     

Option Delisting of Stocks That Continue Trading: An Examination of Welfare Effects

This paper addresses the resulting effects from the delisting of options on underlying stocks that continue to trade. The evidence generally supports the argument that options in this sample are delisted as a result of financial difficulty and/or a lack of interest. An insignificant average abnormal return and a small significantly negative average standardized abnormal return are observed around the delisting date, but the average price effect is determined to be considerably less than the normal discrete trading intervals at which stocks trade on exchanges. There are no effects found on volatility measures directly attributable to option delisting.     

Are real estate markets becoming more efficient?

This paper compares real rates of return to housing capital with returns to other forms of fixed reproducible capital. All data are from the national income and capital accounts and are gross of taxes and depreciation. Returns to housing capital have been consistently lower than those to other capital during the fifty-five year period for which data are available. Since 1950, the disparity between returns to housing and other capital has narrowed steadily and substantially. The paper presents speculations as to the reasons for the improvement in capital market efficiency.I am indebted to Robert Godrick and Daniel Siegel for comments on an earlier draft.