Business groups are large collections of firms that, although legally independent, are joined by complex ownership links, in many cases controlled by a single family. Though uncommon in the United States — in part due to the introduction of intercorporate dividends in the 1930s that has made the groups less attractive here — they are found in most other parts of the world.
The complexity of business groups presents a number of challenges for anyone wanting to assess the performance of a group’s member firms, their strategic position or their valuation, among other considerations. Further, unlike a stand-alone firm, a firm that is part of a group must be judged in the context of its position in the group, says Professor Daniel Wolfenzon. “For example, lenders may view such a firm as less risky than a stand-alone firm since it can fall back on other firms in the group should it falter."
Business groups have not been the subject of much empirical research. Ownership data needed to establish the firm’s connections is hard to come by, and until recently there were few measures for assessing relationships among the firms in a group.
After a number of the largest chaebols — Korean business groups — collapsed in the wake of the Asian financial crisis of 1997–1999, the Korean government, through the Korean Fair Trade Commission, began requiring chaebols to report a great deal of data about their constituent firms. The data provided Wolfenzon, along with Heitor Almeida of the University of Illinois at Urbana-Champaign, Sang Yong Park of Yonsei University in Seoul and Marti Subrahmanyam of New York University, an opportunity to study how chaebols form and evolve over time.
To undertake this study, the researchers first established metrics that allowed them to make sense and summarize the complex ownership structure of groups [Link: Hyundai Motor Group, PDF, 438KB]. They computed a firm’s position, or how close or far a firm is to the controlling family. They also devised a new metric, centrality, that captures the degree to which a firm is used by the family to acquire other firms. Finally, they measured the family’s level of control by taking into account all possible links from the family to the firm in question. “The new metrics we devised are easy to apply to any group structure,” Wolfenzon says, “no matter how complex.”
The researchers used these metrics to look at how new firms are added to the group and, in particular, where firms were positioned in relation to the controlling family.
The poor performance of lower-positioned groups in chaebols is often attributed to agency problems arising from the separation of ownership from control — a firm placed far away from the controlling family will be subject to less direct oversight, resulting in lackluster performance. But Wolfenzon and his coresearchers wanted to see if this effect of distance was causing poor performance or if in fact these firms were selected into a lower position in a group as a result of their poor performance at the time of acquisition.
The researchers found that a firm’s low profitability prior to acquisition led to more pyramiding (in which new firms are placed under the control of another firm) in the chaebol’s overall structure, while a firm’s high profitability prior to acquisition resulted in its placement much closer to the controlling family.
Moreover, the researchers found that firm performance neither improved nor deteriorated once acquired and placed low in the chaebol. In other words, being placed low and in a pyramidal part of the chaebol did not appear to affect profitability one way or the other, bolstering the case for the researcher’s selection hypothesis.
Overpayment for acquisitions also affected placement in the chaebol. The researchers compared the book value of an individual firm to price paid for acquisition and found that firms for which the chaebol paid more relative to book value are placed near the bottom while those for which the chaebol paid less relative to book value are placed near the top. They also found that most central firms are valued at a discount, particularly those that actively acquire other firms. “This suggests that shareholders are anticipating the selection of low net present value (NPV) firms into pyramids,” says Wolfenzon, “and shareholders discount the share of these central firms accordingly.”
Why do groups engage in acquisitions that have low NPV at all? “It is possible that the family derives other benefits, not just monetary, from controlling these firms,” Wolfenzon says, “but don’t want to use family money for the acquisition. Pyramiding involves tradeoffs but can be an efficient way for the controlling family to access resources.”
Daniel Wolfenzon is the Stefan H. Robock Professor of Finance and Economics at Columbia Business School.