Theoretical and empirical evidence of the influence of economic linkages on stock returns
University of Cambridge
Department of Land Economy
Corpus Christi College
MetadataShow full item record
Meyricke, R. (2013). Theoretical and empirical evidence of the influence of economic linkages on stock returns (doctoral thesis).
Inter-linkages between suppliers and customers are a channel by which shocks can spread between firms. When firms buy and sell intermediate goods from one another, they may rely on each other for the supply of input goods or for cash-flow from sales. This is a problem because financially distressed suppliers can pose significant risk to the economic activity of customers that rely on them for goods and services. A case in point is the heavy loss suffered by General Motors when its equipment and parts supplier Delphi went on strike in 1998. Vice-versa, distressed customers can negatively impact suppliers' business operations. Real economic activities are highly related to major stock pricing factors. The main hypothesis of this thesis is that shocks to a firm's direct and indirect suppliers and customers influence its stock price. There is a large amount of research addressing how shocks spread between international financial markets and asset classes influence stock prices during financial crises (financial contagion). Past research has identified the macroeconomic conditions and the types of linkages between markets and assets that make a country or market vulnerable to financial contagion. Little is known, however, about how shocks spread via economic linkages influence firm-level stock returns. Studies find that significant movements in a firm's stock price forecast subsequent movements in the stock price of its major suppliers. Several questions remain open, however, regarding how shocks spread via economic linkages influence stock returns, such as: how shocks spread via economic linkages influence return volatility and correlation; what characteristics of economic linkages (e.g. the degree or the concentration of linkage) are most important in the process of contagion; and whether the spread of shocks via economic linkages increases during recessions. The main objective of this thesis is to increase knowledge of how economic linkages between firms influence stock returns. My approach is to examine how a firm's economic linkages influence three dimensions of its stock returns: volatility, pairwise correlation between linked firms' returns and the cross-sectional distribution of average returns. The research questions addressed are: 1. How does the structure of a firm's economic linkages influence the volatility of its stock returns? 2. How do shocks transmitted via economic linkages increase correlation between linked firms' returns? 3. How do shocks transmitted via economic linkages affect average returns, cross-sectionally and over time? For each dimension of stock returns (volatility, pairwise correlation and average returns) I examine what characteristics of economic linkages are most influential, and whether the influence of economic linkages increases in recessions. I develop a theoretical model explaining how the spread of cash-flow shocks via economic linkages between firms influences the volatility, pairwise correlation and average level of stock returns. The reduced form of the theoretical model corresponds to a factor model of stock returns (based on Arbitrage Pricing Theory), with an additional factor added to allow for non-diversifiable risk created by economic linkages. This model describes the relationship between economic linkages and return volatility, pairwise correlation and average returns. To answer the first research question, I apply the Lindeberg-Feller theorem to derive an explicit relationship between a firm's stock return volatility and the structure of its linkages to other firms. I prove that when the distribution a firm's economic linkages is heavy-tailed (such that it has an extremely high degree of economic linkage to a few firms and a far lower degree of economic linkage to all others), shocks to the firm's key suppliers and/or customers can significantly influence its return volatility. Intuitively, shocks to the most connected suppliers and/or customers are not offset by shocks to less connected suppliers and/or customers, so they can significantly influence a firm's cash-flow and therefore stock returns. Monte Carlo simulations con firm that shocks transmitted via economic linkages are diversified away at rate much slower than the 1/(√N) rate implied by the law of large numbers in many common supply chain structures. In these 'concentrated' supply chain structures, shocks transmitted via economic linkages can create portfolio return volatility in excess of that explained by systematic risk factors, even in large portfolios. To answer the second and third research questions, I use monthly stock return data and annual accounting data on the major customers of all listed US firms between 1990 and 2010 from the CRSP/Compustat database. To investigate how shocks transmitted via economic linkages influence correlation between linked firms' returns, I test the hypothesis that an increase in the degree of linkage between two firms increases the pairwise correlation between their stock returns. First, I adapt correlation-based tests of contagion to test whether pairwise return correlation is higher when two firms are linked than when they are not linked. Second, I develop measures of the strength of pairwise linkage between firms (using principles from network theory and economic input-output modeling). I then estimate regressions of firm-pairs' return correlation against the strength of their linkage and a number of controls (such as industry-pair fixed-effects and credit usage along the supply chain). The regression results show that an increase in the economic linkage between two firms is associated with increased correlation between their stock returns. Linked firms' returns are more correlated when credit is involved in the supplier-customer relationship and in recessions, implying that it is harder to replace a supplier or customer in these situations. Finally, I test whether shocks spread via economic linkages influence average stock returns over and above other factors that have been shown to influence stock returns. My method is to develop measures of the degree and concentration of a firm's supplier and customer linkages. I include these measures in a factor model of stock returns alongside a number of other factors that have been shown to explain stock returns. Cross-sectional regressions show that, in a given time-period, firms with more concentrated supplier bases have higher average returns than firms with less concentrated supplier bases. Second, time-series regressions showed that an increase in the concentration of a firm's supplier-base lowered realized returns in the following period. These results suggest that investors demand a positive risk premium (higher expected return) for holding the stock of firms whose supplier-base is concentrated. This places downward pressure on prices following an increase in supplier-base concentration. While concentration of a firm's supplier and customer linkages has a significant influence on stock returns, the magnitude of this effect is small compared to the influence of systematic risk factors. The influence of economic linkages on stock returns, however, increases in recessions. Together the results in this thesis provide solid evidence that shocks spread via economic linkages can affect the volatility, correlation and average level of stock returns. The thesis establishes a robust framework for modeling the returns of portfolios in which the underlying securities or firms are linked via economic relationships. This is an important extension to existing models that ignore the potential impact of shocks spread via linkages between firms on stock prices. The model can be used for pricing securities with concentrated supply chain exposures or to identify stock portfolios that are susceptible to contagion.
This record's URL: http://www.dspace.cam.ac.uk/handle/1810/244390