Short sales restrictions and market quality: Evidence from Korea

Abstract

We examine changes in volatility and market quality around the shorting ban during the 2008 global financial crisis in a unique policy environment in Korea. Using intraday quotes and transaction data, we find no evidence that the shorting ban increased volatility or worsened liquidity. While the ban was lifted only for nonfinancial stocks, the patterns of change in volatility and market quality were similar across nonfinancial and financial stocks, suggesting that the changes in volatility and market quality around the shorting ban and its removal were likely to have been driven by changes in overall market conditions rather than the shorting ban itself.

Introduction

Short selling refers to the practice of trading designed to profit from a price decline by selling a borrowed security with the intention of buying it back at a future date at a lower price. Short selling is widely used as a means of speculation and hedging by hedge funds and market makers. The prevailing academic view of short selling is that the presence of short sellers tends to improve market efficiency: short sellers contribute to market liquidity and price discovery, and their efforts to identify overvalued firms enhance market efficiency.¹ Empirical evidence supporting the notion of short sellers enhancing market efficiency and price discovery is provided by Jones and Lamont, 2002, Bris et al., 2007, Boehmer et al., 2008, Boehmer et al., 2013, Diether et al., 2009, Karpoff and Lou, 2010, Beber and Pagano, 2013 and Comerton-Forde et al. (2016), among others.

In contrast to the academic consensus on the role of short sellers, the media coverage of short selling is not as positive, especially when the stock market sharply declines: short sellers are often blamed for aggravating market decline. Indeed, in the midst of the 2008 global financial crisis, regulators in many countries imposed bans on short selling in one form or another with the objective of stabilizing stock prices. In Korea, the shorting ban for financial stocks lasted longer than in any other country, and the ban for nonfinancial stocks lasted longer than in any other country except for Greece. The financial authority in Korea lifted the ban on financial stocks on November 13, 2013, after more than 5 years of its imposition.

In this paper, we study changes in various measures of market quality around the shorting ban in Korea during the 2008 global financial crisis to assess whether the shorting ban was effective in mitigating increasing market volatility and worsening liquidity. While the shorting ban in the U.S. was limited to financial stocks and lasted less than a month (imposed on September 19, 2008 and lifted on October 9, 2008), the shorting ban in Korea was imposed on all stocks on October 1, 2008, and it was lifted after eight months on June 1, 2009 only for nonfinancial stocks. The focus of our empirical study is this partial removal of the ban on short selling, which helps us to address the endogeneity problem in which changes in volatility and market quality measures may be due to changes in the overall market conditions rather than changes in short selling regulations. The unique policy environment in Korea provides a quasi-natural experiment setting where we can compare changes in the volatility and market quality measures across nonfinancial stocks and financial stocks before and after the partial removal of the ban, which helps us to partially address the endogeneity problem in sorting out the impact of the short selling ban and its removal.

Using intraday quotes and transaction price data from July 1, 2008 through August 31, 2009, we compare intraday volatility and market quality measures across three sub-periods separately for nonfinancial stocks and financial stocks: the three months prior to the shorting ban (July 1 to September 30, 2008), the eight months of the short selling ban (October 1, 2008 to May 31, 2009), and the three months following the lifting of the ban on nonfinancial stocks (June 1 to August 31, 2009). We also provide additional results for the longer sample period from July 1, 2008 to February 28, 2014 using daily frequency data. Previous research on the impact of the shorting ban in Korea, such as Kim (2010) and Choe and Lee (2012) also uses intraday transaction data. Kim (2010) examines the impact of the ban on spread, and Choe and Lee (2012) study changes in spread and intraday volatility. Both papers examine the impact of the ban only on financial stocks, and neither considers the impact on volatility and market quality when the shorting ban was lifted for nonfinancial stocks in June 2009. In contrast to these studies, we examine changes in market quality of both financial and nonfinancial stocks when the shorting ban was imposed for all stocks but lifted only for nonfinancial stocks to assess whether the ban mitigated increasing volatility and worsening liquidity.

Examining changes in volatility and spread across the three periods, we find no evidence that the shorting ban reduced volatility in stock returns or improved market liquidity. In fact, following the shorting ban, volatility increased and spread widened, and following the lifting of the ban, volatility decreased and spread narrowed. During the post-ban period, the shorting ban was still in effect for financial stocks, but we find that their average volatility is lower and spread narrower during the post-ban period than during the ban period as was the case for nonfinancial stocks. Since the pattern of changes in volatility and spread is similar for nonfinancial stocks and financial stocks, our findings suggest that the changes in volatility and spread were more likely the results of changes in the overall market conditions than the shorting ban itself.

We partially address the confounding effects of changes in market conditions on volatility and liquidity by conducting a matching sample analysis comparing the short sale sample and the matching sample. However, we find the marginal impact on the volatility and spread of the short sale sample to be statistically insignificant, which suggests that the shorting ban was ineffective in mitigating increasing volatility and worsening liquidity.

We also find that the relative bid depth increased following the ban, indicating a reduction in the ask quote balance relative to the bid quote balance. Since the Korean stock market is order-driven without market makers or specialists, market orders can be interpreted as consuming liquidity, and limit orders can be interpreted as providing liquidity in transactions. Under the uptick rule in Korea, by which a stock can be shorted only at a price higher than the last transaction price, short sellers are likely to submit limit sell orders at the best ask quote, contributing to the ask quote balance. Therefore, our finding of the increased relative bid depth following the ban is likely to be due to the absence of short sellers who provide liquidity to the market.

We also examine changes in volatility and liquidity using daily frequency data over a longer sample period from July 1, 2008 to February 28, 2014. During this longer sample period, the shorting ban was imposed on all stocks twice, lifted twice for nonfinancial stocks, and lifted once for financial stocks. We find that the shorting ban is associated with increasing volatility and worsening liquidity and that the lifting of the ban is associated with decreasing volatility and improving liquidity, which is similar to the pattern found in the shorter sample measuring volatility and liquidity using intraday data.

This study differs from previous research in that we find no evidence of the shorting ban negatively affecting the market quality. Previous empirical evidence on the impact of the shorting ban in the U.S. in 2008, such as Battalio et al., 2012, Battalio and Schultz, 2011, Boehmer et al., 2013 and Lobanova et al. (2010), all show that short sale bans are detrimental to market liquidity. Empirical evidence in the U.K. (Marsh and Payne, 2012), France (Lioui, 2010b), and Spain (Arce and Mayordomo, 2016) also shows that shorting bans have a negative impact on market liquidity. In Beber and Pagano’s (2013) comprehensive study of 30 countries, shorting bans are shown to be detrimental to liquidity and price discovery.

Recent studies on market quality in the stock markets of Taiwan and China such as Cheng et al. (2012) and Hu et al. (2019) find that lifting the shorting ban improve the efficiency of price discovery. In particular, Cheng et al. (2012) show that the removal of the shorting ban for IPO stocks in Taiwan improved price efficiency. Hu et al. (2019) show that the removal of the shorting ban on a designated list of firms in China resulted in firms experiencing a lower cost of equity and enjoying higher market liquidity.

According to our study, the ban on short sales during the 2008 global financial crisis failed to prevent the deterioration of market quality as in the previous research. However, our study also investigates the period during which the ban on short sales was lifted. If the ban on short sales had worsened the market quality, the market quality would have had to improve after the lifting of the ban. However, we find no such evidence.

This paper is organized as follows. Section 2 describes a unique policy experiment in Korea and data used in this study. Section 3 provides the main hypothesis and explains the empirical methodology and the various measures of market quality. Section 4 provides the main empirical findings on changes in market quality around the shorting ban. Section 5 provides the results of several robustness checks, and Section 6 presents concluding remarks.