Distress Anomaly and Shareholder Risk: International Evidence

• Author: Assaf Eisdorfer,Amit Goyal,Alexei Zhdanov
• DOI: http://doi.org/10.1111/fima.12203
• Published date: 01 September 2018
• Date: 01 September 2018

Distress Anomaly and Shareholder

Risk: International Evidence

Assaf Eisdorfer, Amit Goyal, and Alexei Zhdanov∗

Financially distressed stocks in the United States earn puzzlingly low returns giving rise to the

distress risk anomaly.We provide evidencethat the anomaly exists in developed countries, but not

in emerging ones. Using cross-country analyses,we explore several potential drivers of returns to

distressed stocks. The distressanomaly is stronger in countries with stronger takeover legislation,

lower barriers to arbitrage, and higher information transparency. In contrast, shareholder bar-

gaining power and expected stock return skewness in a country do not affect the anomaly. These

findings suggest that various aspects of shareholders’ risk play an important role in shaping

distressed stocks returns.

Since Fama and French (1992) suggested financial distress risk as a potential explanation

for the value premium, several academic studies have examined the performance of financially

distressed stocks (see, for example, Dichev, 1998; Griffin and Lemmon, 2002; Campbell, Hilscher,

and Szilagyi, 2008). However, contrary to the general intuition, distressed equities are typically

found to have lower returns, a finding inconsistent with risk-based theory. Furthermore, the

excess returns to most distressed stocks are negative (Campbell et al., 2008). We refer to this

phenomenon as the “distress risk puzzle.” While a few potential explanations havebeen proposed,

there is still no consensus in the literature as to what drives this anomaly.

To the extent that distress risk represents a priced risk factor, Campbell et al. (2008) find that

financially distressed firms have high market betas and high loadings on SMB (small minus

big) and HML (high minus low) factors. From a risk perspective, it is hard to reconcile this

finding with low returns to distressed equities. Garlappi, Shu, and Yan (2008) propose a model

in which distressed stocks become safer as they approach the default boundary due to violations

of the absolute priority rule. However, their model cannot account for negative excess returns to

distressed equities (and appears to contradict the finding that distressed stocks usually have high

loadings on risk factors).

The distress anomaly may result from unexpected developments during the sample period, as

suggested by Campbell et al. (2008). In support of this view, Chava and Purnanandam (2010)

argue that the association between distress risk and expected returns is actually positive, and the

negative relation between distress risk and realized returns is due to a streak of surprisingly low

realized returns on distressed stocks in the United States in the 1980s that were not anticipated

by investors. That is, the distress anomaly may be an in-sample phenomenon that is unlikely to

continue in the future and is most likely specific to the United States.

We thank Bing Han (Editor), an anonymous referee, Tarun Chordia, Norman Sch¨

urhoff, Christian Wagner, and Josef

Zechner for valuable comments and suggestions. Amit Goyalwould like to thank Rajna Gibson for her support through

her NCCR-FINRISK project.

∗Assaf Eisdorfer is an Associate Professor of Finance at the University of Connecticut in Storrs, CT. Amit Goyal is a

Professor of Finance at the Swiss Finance Institute at the University of Lausanne in Switzerland. Alexei Zhdanov an

Assistant Professor of Financeat Penn State University in University Park, PA.

Financial Management •Fall 2018 •pages 553 – 581

554 Financial Management rFall 2018

The distress anomaly might also have a misvaluation explanation. For example, investors may

not fully understand how to predict failure risk and, as such, may not discount the prices of

distressed stocks enough to offset their failure probability. Campbell et al. (2008), however, cast

doubt on this argument by showing that returns to distressed firms are not concentrated around

earnings announcements.

Some alternative explanations have also been proposed. Campbell et al. (2008) point out that

investors could exhibit preferences for positively skewed returns and, therefore, are willing to

hold distressed stocks (that are likely positively skewed) despite their low returns. It could also

be that the majority owners of distressed companies can extract private benefits, for example, by

purchasing the company’s output or assets at bargain prices.

In this paper, we go beyondthe limitations of the United States data and study the perfor mance

of distressed stocks around the globe. Wepursue two goals. First, we examine whether the distress

anomaly is specific to the United States or is present in other countries as well. This evidence

can help us evaluate some of the potential explanations for this anomaly. For example, if the poor

performance of distressed equities is caused by the divergence between expected and realized

returns due to some specific events that occurred in the United States, we are unlikely to observe a

similar pattern in most other countries. Second, an international study allows us to relate returns to

distressed stocks to various country-level characteristics (e.g., measures of shareholder protection

or the friendliness of takeover legislation, among others). By doing so, and measuring which of

these country-specific variables affect returns to distressed stocks, we can also better understand

what drives the distress anomaly in the United States.

We study returns to distressed companies in 34 countries over the period 1992–2010. Firm-

specific data are taken from WorldScope, while country-level data are obtained both from ag-

gregation of firm-specif ic data and from sources used in prior international studies. Identifying

financially distressed companies in different countries is challenging for several reasons. Econo-

metric models of financial distress are typically based on a large set of accounting variables (see,

for example, Altman, 1968; Ohlson, 1980; Campbell et al., 2008), which are missing for many

international firms. Moreover, the parameters of most of these models were estimated using US

stocks and may not be suitable for other countries (see Altman, 2005). We therefore employ the

distance-to-default measure derived from Merton’s (1974) model (also employed by Moody’s

KMV) to measure the extent of a firm’s financial distress (see also Vassalou and Xing, 2004).

Unlike most alternative models, the distance-to-default measure uses only equity value, equity

volatility,and the face value of debt, allowing us to estimate this measure on a large cross-section

of international firms.

We present two main findings. First, we find that the distress anomaly only appears to exist

in developed countries (with a magnitude comparable to that reported in the United States),

but not in emerging markets. In developed countries, value-weighted portfolios of stocks in the

bottom distress quintile underperform the portfolios of stocks in the top distress quintile by an

average of 0.17% to 0.63% a month (measured by excess return and alphas from factor models).1

In emerging markets, however, no such effect is found. In fact, in some specifications, results

for emerging markets indicate that returns to distressed stocks are higher than those on solvent

stocks. Weverify this result by running f irm levelmonthly Fama and MacBeth (1973) regressions

of excess stock return on the extent of financial distress separately for firms in developed and

emerging countries. Thus, while the distress puzzle is not US-specific, it is more pronounced in

developed markets than in emerging markets. This finding also suggests that the distress anomaly

goes beyond an in-sample phenomenon, as suggested by Chava and Purnanandam (2010).

1The distress effect using the KMV measure is weakerthan when using other distress models (see Campbell et al., 2008).