فراسوی بازارهای سرمایه داخلی: انتقال شوک های منفی فروش و کانال وثیقه

We study how shocks to some business segments affect investment in a firm's non-shock segments. We find that subsequent investment in the non-shock segments is significantly lower compared to segments of firms that do not experience shocks. Surprisingly, lower availability of internal funds does not account for the lower investment. We find that segment shocks propagate within the firm by decreasing the value of collateral assets and reducing the availability of external finance. Our results support the operation of an external finance collateral channel ([Kiyotaki, N., Moore, J., 1997. Credit cycles. Journal of Political Economy 105, 211–248.]) previously discussed in the literature.

Putting different industry segments under the same roof is likely to have real consequences single bond events in one segment of a firm may affect investment policies of, and resources allocated to, other segments. In a world of capital market imperfections, segment investment depends not only on investment opportunities in a particular segment, but also on the cash flows and asset values of the whole firm. Therefore, an unfavorable shock that decreases cash flows or asset values of a segment may reduce investment in the remaining segments. Shin and Stulz (1998) find that segment investment depends on the cash flow of the firm's other segments. Lamont (1997) reports that non-oil investment by oil companies declined following large decreases in the oil segments' cash flows. Lamont's (1997) pioneering work was somewhat limited by the fact that the data came only from one particular industry episode, so that he could only look at 26 firms and 40 segments. Although focusing on the oil industry gave him a very clean “controlled experiment” in that the shock was unquestionably exogenous to the non-oil segments, the paucity of data meant that he could not investigate the mechanisms through which the shocks to the oil segments were transmitted to the non-oil segments. Disentangling the effects of alternative channels of transmission by utilizing the full sample of Compustat firms is one of the main contributions of this paper. We identify multi-segment firms that experience major sales declines (sales shocks) but where the decline in sales is confined to only some of the segments (the “shock” segments). We then investigate the effect of the sales shock on the subsequent investment in the segments that did not experience a contemporaneous decline in sales (the “non-shock” segments). We are especially interested in seeing whether the mechanism of transmission goes beyond the obvious cash channel — i.e., lower availability of internal funds. We find shock firms invest less in their “non-shock” segments relative to segments of firms that do not experience sales shocks. Contrary to what might be expected, however, our results suggest that the effect of a decrease in the overall availability of internal funds is not of first-order importance in explaining the lower investment by the non-shock segments of the affected firms. While a lower availability of internal funds — ceteris paribus — is expected to result in lower investment in the non-shock segments of financially constrained firms, the shock is also associated with a loss of investment opportunities in the shock segments. The latter implies that more internal funds are available for investment in the non-shock segments, which should cause the non-shock segments to invest more. Indeed, our results show that for the shock firms, these two effects typically tend to offset each other. Importantly, even after controlling for the availability of internal funds, investment by the non-shock segments is significantly lower. It follows, therefore, that the transmission mechanism of the sales shocks goes beyond internal capital markets and the availability of internal funds, and other “non-cash” channels of transmission are at work. One potential candidate is the external financing channel. The cost of external financing could increase, and the availability of external funds could decrease, subsequent to the shock. There may be several specific channels through which this happens. For example, the debt overhang problem could become more severe as the firm becomes more likely to default, making it costly to raise new junior debt. We examine one particular channel of transmission, namely the “collateral channel”.1 Financially constrained firms that depend on the collateralizable value of their fixed assets to raise external capital are especially vulnerable to adverse shocks that impair the resale value of these assets. A drop in collateral value will impair the ability of the firm to raise external capital, which in turn will lower investment in fixed assets, further reduce the ability of the firm to raise external capital, and so on (Kiyotaki and Moore, 1997). Therefore, we expect the investment cuts in the non-shock segments to be larger for firms with more collateralizable assets under conditions that will impair the value of these assets. We find evidence consistent with a collateral channel of transmission. As a first step, we examine whether firms with a higher ratio of an estimate of fixed assets in the shock segments to total firm assets invest less in the non-shock segments (after controlling for the relative size of the shock segments). We find that non-shock segment investment is significantly lower for these firms, and the effect persists for at least 3 years after the shock. We then examine the possibility that firms may experience a drop in collateral value for different reasons, depending on whether or not the shock is primarily firm-specific, or industry-wide. We argue that the collateral value of the assets is most likely to be impaired after a shock if the firm's competitors (the potential buyers of these assets) are also experiencing negative shocks (Schleifer and Vishny, 1992). We find that a higher proportion of collateral that is impaired by an industry shock to the firm's total assets is associated with a bigger decline in investment in the non-shock segments when the firm is highly levered. We conjecture that the effect shows up only for high debt firms because managers may be more willing to sell assets when the shock is of industry origin (as such sales do not signal managerial mistakes (Boot, 1992), and this largely offsets the lower liquidation value of these assets for this type of shocks; however, for highly levered firms, the lower liquidation value has a more serious effect on the firm's ability to raise external finance, leading to more severe decline in investment in the non-shock segments. Our primary criterion for identifying shock firms is in the spirit of Lamont's (1997) classification, and similar to that in several other studies (Opler and Titman, 1994, Bertrand and Mullainathan, 2001 and Campello, 2003). We identify firms that are below-median in terms of sales growth and stock return in industries that experience adverse median sales growth and stock returns. Since the shocks in these instances are of industry origin, they are exogenous to the non-shock segments of the shock firms, i.e., those segments that experience a contemporaneous increase in sales. Our primary tests of the transmission mechanism and the collateral channel are based on this sample. However, to test hypotheses about the transmission of firm-specific versus industry-wide shocks, we need a different sample. Accordingly, we identify a sample of shock firms and non-shock segments based solely on a major drop in sales and significantly negative stock returns for the firm.2 Since the latter sample comprises of both types of shocks, it allows us to contrast the ways in which they are transmitted to the non-shock segments. The rest of the paper is organized as follows. Section 2 reviews the literature. Section 3 describes the data and our sample. Section 4 presents the empirical results. Finally, Section 5 concludes.

We document that multi-segment firms experiencing adverse sales shocks to some segments reduce investment in their remaining segments (the non-shock segments). We find that the affected firms invest less in the non-shock segments relative to other firms with segments in these same industries. Our multivariate regressions show that the lower investment by the affected firms in the non-shock segments persist for at least 3 years after the shocks, controlling for cash flows, investment opportunities, capital structure, industrial and macroeconomic conditions. This suggests that the transmission mechanism of shocks to the non-shock segments goes beyond internal capital markets and the availability of internal funds. We examine a specific external financing channel, namely, the collateral channel. We find that firms with higher ratio of shock segment fixed assets to the firm's total assets prior to the shock invest less in the non-shock segments. This suggests that a fall in the collateral value of assets contributes to the increase in financial constraints faced by these firms and the decline in investment in the non-shock segments.