Business Groups and Firm-Specific Stock Returns

Mara Faccio is the Hanna Chair in Entrepreneurship & Professor of Finance at Purdue University Krannert School of Management; Randall Morck is the Stephen A. Jarislowsky Distinguished Chair in Finance at the University of Alberta; and M. Deniz Yavuz is Associate Professor at Purdue University Krannert School of Management. This post is based on a recent paper authored by Professor Faccio, Professor Morck, and Professor Yavuz.

Measures of general and financial development tend to correlate positively with measures of firm-specific stock return volatility at the economy level (Morck et al. 2000). Lower firm-specific stock return volatility, in turn, is associated with less efficient capital allocation (Wurgler, 2000; Durnev et al., 2004; Morck et al., 2013). Business groups, collections of separately listed firms under common control through equity blocks, are also more prevalent in lower income economies (La Porta et al., 1999; Morck, Wolfenzon, and Yeung, 2005; Khanna and Yafeh, 2007), and thought to be a second best suboptimal solution to allocatively inefficient financial markets (Khanna and Palepu, 2000; Almeida and Wolfenzon, 2006; Khanna and Yafeh, 2007). This study connects these two lines of research by showing that business group affiliated firms’ stock returns exhibit less firm-specific volatility than do the returns of unaffiliated firms in similar conditions.

Firm specific (idiosyncratic) shocks are more frequent in some firms and economies than in others due more active risk taking, entrepreneurship and creative destruction (Chun et al., 2008, 2011; Bartram, Brown and Stulz, 2012) or a more competitive environment (Irvine and Pontiff, 2009). Differences in the magnitudes of idiosyncratic shocks observed by market participants may also reflect differences in transparency (Jin and Myers, 2006), analyst incentives to produce firm specific information (Veldkamp, 2006) and media coverage (Kim, Yu and Zhang, 2016). Numerous studies, surveyed in Morck et al. (2013), support all of these explanations with the common theme being higher firm-specific returns volatility reflecting better functioning financial systems in developed economies.

The prevalence of business groups also correlates with economic development. Morck et al. (2005), Khanna and Yafeh (2007) and others argue that this is due to business groups being second best suboptimal capital and risk allocation mechanisms where financial systems functions poorly. Business groups can shift earnings from firms with excess free cash flow to firms with unfinanced profitable investments (Almeida and Wolfenzon 2006). Business groups may shift earnings to fund private benefits of controlling shareholders (Johnson et al. 1999), prop up ill-run affiliates (Morck and Nakamura 1999) and share or obscure risks (Hoshi, Kashyap and Scharfstein 1990, 1991; Friedman, Johnson and Mitton 2003; Gopalan, Nanda, and Seru 2007). Investors, expecting risk sharing or income shifting between business group affiliates, may rationally expect idiosyncratic shocks to have less impact on the share prices of a group affiliate than on the share prices of an otherwise comparable unaffiliated firm, all else equal.

To understand whether group affiliated firms incorporate firm-specific shocks differently from unaffiliated firms and avoid explanations contaminated by differences in the frequency or observability of shocks one would ideally like to observe the responses of group affiliated and unaffiliated firms to identical shocks observed by all market participants. This is precisely what we attempt to do in this paper. We use shocks to global commodity prices, which affect all commodity price-sensitive firms, whether group affiliates or independent firms, simultaneously; and which all market participants observe simultaneously.

This approach requires exogenously identifying industries that are sensitive to shocks to the price of each commodity. We employ two alternative methods of matching industries to commodities. The first method uses out-of-sample US data to estimate commodity price sensitivity across industries. The second approach uses Bureau of Economic Analysis (BEA) input-output tables to match industries to commodities.

In a global sample of 43 economies we find that a given commodity price shock moves the stocks of unaffiliated firms significantly more than the stocks of group affiliates after controlling for time-varying economy-industry-level latent factors. These results are not driven by firm-level hedging, diversification, leverage, size or R&D activity. Further identification follows from difference-in-difference tests exploiting control block sales and purchases, which show the stocks of previously unaffiliated firms that become group affiliates simultaneously becoming less sensitive to such commodity price shocks. These results ensue whether the control group is matched firms whose group affiliation status is unchanged or matched firms subject to control block bids that failed for defensibly exogenous reasons.

These firm-level tests are consistent with business group affiliation attenuating firm-specific stock return volatility associated with commodity shocks. If group affiliation attenuates commodity shocks, it may well attenuate other firm-specific shocks. A final set of tests show that individual business group affiliates’ attenuated firm-specific stock return variation aggregates into an economically and statistically significant explanation of cross-economy differences in stock return co-movement. These tests reveal business group prevalence to be a complementary explanation to others advanced in the literature (Morck et al. 2013).

These findings suggest that a greater prevalence of business groups more actively shifting income between member firms explains, at least partially, why stock returns move more synchronously in some economies than in others. Thus investors’ expectations of how risks are shared across business groups is a fundamental driver of differences in idiosyncratic volatility of stock returns.

The complete paper is available for download here.

Both comments and trackbacks are currently closed.