The Illusory Nature of Long-Horizon Post-Offer Negative Abnormal

The Illusory Nature of Long-Horizon Post-Offer Negative Abnormal Performance: Evidence from UK Seasoned Equity Offerings Abhay Abhyankar Keng-Yu Ho Warwick Business School University of Warwick 24 May, 2002 NTU International Conference on Finance

Presentation Outline • Introduction • Literature Review and Motivation – What do previous empirical findings say? – What are the problems in measuring and testing the significance of abnormal return? Risk Adjustment and Return Measurement. – What is the motivation of our study? • Data and Methodology – UK Seasoned Equity Offerings (Rights Issues and Placings) – Buy-and-Hold Abnormal Return (Matching Firm/Portfolio Approach) – Calendar-Time Approach (Parametric Asset Pricing Model) • Empirical Results – Pre- and Post-Event Buy-and-Hold Abnormal Performance – Calendar-Time Abnormal Performance using Fama-French 3 -factor and Carhart 4 -factor Model • Conclusion

Introduction • Standard Event Studies (Fama, Fisher, Jensen and Roll, 1969) – Use a two-day event window to study stock price relative to corporate events. • Long-Run Event Studies – Evaluate stock price performance 3 -5 years after a corporate event. – Previous empirical studies show slow incorporation of information in stock prices and therefore present arbitrage opportunities that should not remain in an efficient market. – Anamalous results may be due to Problems with Risk Adjustment (Asset Pricing) and Measurement (Econometric) issues.

Quick Review of Related Studies

Key Methodology Issues • How to measure abnormal return? – Buy-and-Hold Abnormal Returns (BHARs) – Cumulated Abnormal Returns (CARs) – Problems: New Listing, Rebalancing, Skewness and Overlapping Biases • How to correct for risk or What asset pricing models to use? – Matching by firms in similar size, book-to-market or industry. – Matching by size/book-to-market portfolios. – Calendar-time approach using parametric asset pricing models. (CAPM, Fama-French 3 -factor model)

Motivations • Good opportunity to revisit the issue by using UK data. – Less Data-Snooping Bias – Rights Issues v. Placings • Pre- and Post-Offer Long-Horizon Abnormal Performance – Complementing recent studies on announcement wealth effect. (Slovin, Sushka and Lai (2000) and Suzuki (2000)) • Using various metrics to mitigate the asset pricing and econometric problems in measuring long-run abnormal performance. – Buy-and-Hold Abnormal Returns – Calendar-Time Abnormal Returns

Data Selection • UK Seasoned Equity Offerings from 1989 -1997 – Right Issues: an offer to existing shareholders of securities to subscribe or purchase further securities in proportion to their holdings. – Placings: an offer to new shareholders on a non-pro rata basis. – UK institutional features provide a good opportunity to analysis the long-horizon abnormal performance following two different issuing methods. • Sample Selection Criteria – Common Equity offered in Pound Sterling – Non-Financial and Non-Utility Firms – Listed in London Stock Exchange Main Market • Return/Price and Market Capitalisation data from London Share Price Database (LSPD) • Book-to-Market data from Worldscope

Data Descriptive Statistics • Number and Amount of Issues (in £ million)

Methodology (I) • Buy-and-Hold Abnormal Returns • Matching Firm/Portfolio Approach – Size/Industry and Book-to-Market/Industry Matched Firm • No New Listing and Rebalancing Biases • Reduce Skewness Bias • Use Traditinal t-test – Size/Book-to-Market Matched Portfolio • 25 size/book-to-market portfolios, rebalancing each year. • Use skewness-adjusted t-test due to more skewness bias.

Methodology (II) • Calendar-Time Abnormal Returns – Eliminate Overlapping Bias – For each month, constructing equal- and value-weighted portfolio including all firms that have an event occur in the previous 3 years. – Time Series Regression using Fama-French 3 -Factor Model: – Time Series Regression using Carhart 4 -Factor Model:

Empirical Results (I) • Pre-Offer Abnormal Performance – Positive 3 -year buy-and-hold abnormal return for Rights Issues using size/industry, book-to-market/industry matched firm and size/book-tomarket matched portfolio approach. – Positive 3 -year buy-and-hold abnormal return for Placings using size/industry, book-to-market/industry matched firm and size/book-tomarket matched portfolio approach. – The results show stock price run-up before the SEOs as reported by the US studies. • Placings have more positive abnormal performance than Rights Issues.

Empirical Results (II) • Post-Offer Abnormal Performance – Negative 3 -year buy-and-hold abnormal return for Rights Issues using size/industry, book-to-market/industry matched firm and size/book-tomarket matched portfolio approach. – Negative 3 -year buy-and-hold abnormal return for Placings using size/industry, book-to-market/industry matched firm and size/book-tomarket matched portfolio approach. . – No evidence of more underperformance for smaller firms. • Placings have less negative abnormal performance than Rights Issues. – Placings are usually placed to new institutional holders. Shleifer and Vishney (1986) suggest that the blockholders can increase the value of the issuer because they are more likely to monitor the management actively. – These results are not consistent with the adverse selection literature (Eckbo and Masulis (1992)).

Empirical Results (III) • Calendar-Time Abnormal Return on Fama-French 3 -Factor Model – No evidence is found for the underperformance on both Rights Issues and Placings. • Calendar-Time Abnormal Return on Carhart 4 -Factor Model – No evidence is found for the underperformance on Rights Issues. – Underperformance on Placings is similar to that found by using buy-andhold abnormal return approach. • In general, the long-horizon underperformance after SEOs fades away when calendar-time approach is used. The results suggest that the long-run abnormal return is sensitive to the metrics used.

Conclusion • Positive Pre-Offer abnormal performance and Negative Post. Offer abnormal performance for both Rights Issues and Placings. • Rights Issues have more significant underperformance than Placings following the offers. This can be explained by the result of new institutional holders and ownership dispersion following placings, which can increase the value of the issuer because blockholders are more likely to monitor the management actively. • When calendar-time approach is used, the abnormal return following both Rights Issues and Placings fades away. • We conclude that the anomalous underperformance found for UK SEOs may be an artefact of methodology used. Our study underlines the need to adjust correctly for risk in measuring longhorizon abnormal returns.