The Real Effects of Securities Indices
Washington University in Saint Louis, John M. Olin Business School, Students
November 2, 2015
I study how a prominent segmentation device in capital markets – the membership (or lack thereof) of a firm’s securities in a securities index- affects real economic activity. Using an exogenous shock to the constituents of a popular bond index, I find that firms removed from the index experience decreases in investment, decreases in debt financing, and increases in equity issuances. Moreover, I find that firms removed from the index experience permanent increases in both their primary and secondary market bond yields. My results suggest that shocks to secondary markets can indirectly affect real firm decisions.
The Real Effects of Securities Indices
In a neo-classical capital market in the style of Modigliani and Miller (1958), the price of a security perfectly reflects its fundamental value and firms invest in all positive NPV projects. However, in segmented capital markets, the prices of securities may diverge from fundamental values and, potentially, affect real firm decisions. In this paper, I study how a prominent segmentation device in capital markets – the membership (or lack thereof) of a firm’s securities in a widely-benchmarked securities index – affects real economic activity. A number of studies 1 have shown that institutional investors and asset managers have strong incentives to tilt their investments in favor of index securities. These incentives generate “index effects” in equilibrium – the addition (deletion) of a security from an index results in an increase (decrease) in the security’s price and institutional ownership. However, since index additions and deletions do not create any direct transfers of resources to the firms themselves, it remains unclear whether secondary market index effects translate into real effects at the firm level. Consequently, an important question remains unanswered: does index membership affect firm investment and financing decisions?
I answer this question by examining whether being removed from a widely-benchmarked fixed-income index has real economic effects. My hypothesis is simple: the removal of a firm’s securities from a bond index may increase the cost of raising future debt capital for the firm. This may then affect which projects the firm invests in and how the firm finances these projects (Fischer and Merton (1984), Stein (1996)). Using the January 2005 redefinition of the Merrill Lynch Investment Grade index (“IG index” hereafter) as a source of exogenous variation, I find that firms whose bonds were removed from the IG index experience decreases in investment, decreases in the use of debt financing, and increases in the use of equity financing. Furthermore, I find that removed firms experience increases in their future primary market yields and that the real effects are deferentially stronger for firms that are more reliant on external capital to finance new investments. Overall, my results suggest that index membership and bond market segmentation have real economic effects.
I begin my empirical analysis by documenting the secondary market effects of the IG index redefinition. On October 13, 2004, Merrill Lynch announced a significant change in the methodology used to construct their fixed-income indices. As a result, the bonds of over 200 non-financial firms would mechanically transition out of the index on January 01, 2005. Using a difference-in-differences analysis, I find that bond issues that transition out of the IG index experience permanent increases in yields. Removed bond issues experience 90-day abnormal yield increases of 27 basis points (5.6% from the unconditional mean). Moreover, I find that the increases in yields are largely permanent – bonds removed from the index experience a 20.4 basis point increase in yield for the remainder of the sample period. The observed spike in yields and subsequent downward revision is consistent with the literature on both index effects (Kaul, Mehrotra, and Morck (2000), Greenwood (2005)) and slow-moving capital and the limits of arbitrage (Duffie (2010), Duffie, Garleanu, and Pedersen (2007)).
While it is then tempting to assume that increases in secondary market yields translate directly into increases in the cost of debt , this relation is not necessarily true3 and needs to be empirically verified. Again, using a difference-in-differences analysis, I find that removed firms’ future bond issues command a higher yield after redefinition. The average offering yield for removed firms’ first new debt issuance proceeding the redefinition increases by 31 basis points. In addition, the effect appears to be persistent over time. On average, removed firms experience a 38 basis points increase in their primary market yields for the remainder of the sample period. The results suggest that the marginal cost of debt capital (e.g. the discount rate most relevant for evaluating new projects) does indeed increase after being removed from the IG index.
See full PDF below.