fernando.r.anjos - Fernando R. Anjos Instituto Superior de Ciências do Trabalho e da EmpresaISCTE Business School Li

Teaching

Working papers

"Resource Complementarity, Alliances, and Merger Waves", Oct 2008 (MLcode.zip)

The fact that mergers cluster in time is an important puzzle in finance. Recent explanations rely on over-valuation arguments or the existence of an economy-wide component in merger transaction costs. This paper proposes an alternative theory for merger waves. I develop a static game-theoretic model of network formation where firms combine complementary non-tradable resources, either by establishing an alliance or by merging. In a dynamic extension, I use the results from the static model and solve an option exercise game where distinct sets of firms may choose to merge. In both the static and dynamic cases, the existence of inter-industry alliances may propagate merger activity across sectors. The model is consistent with time-series data on the aggregate number of alliance and merger deals. In particular, the data seems to indicate that merger waves are preceded by an increase in alliance activity. The model also has implications beyond the topic of merger waves: (i) merger excess returns and/or merger frequency between firms in different industries should (a) display an inverted-U relationship with respect to complementarity, controlling for the level of coordination problems in alliances; and (b) controlling for complementarity, be higher in industries where alliances exhibit starker coordination problems; (ii) the model offers a rationale for the diversification discount, namely that it is optimal for diversifying acquirers to choose inexpensive targets, if the objective of the diversifying firm is to obtain only a subset of the target's resources; (iii) spin-offs take place optimally when the level of complementarity within the conglomerate is lower and/or the acquirer of the spun off division has a high firm-specific productivity, relative to the conglomerate; (iv) mergers may be socially inefficient, for reasons distinct from market power.

"Costly Refocusing and the Diversification Discount", Oct 2008

I develop a stationary real options model with corporate restructuring costs that endogenously generates a diversification discount. This result requires that restructuring costs associated with spin-offs - i.e. refocusing moves - be significantly larger than those associated with acquisitions - i.e. diversifying moves. The discount is due to the fact that diversified firms performing poorly will still delay refocusing, given the high cost of implementing this strategy. Moreover, the model implies that a diversification discount is consistent with a cross-sectional distribution of firms where a significant portion of single-segment business units is diversified. Using data on US firms for the period 1984-2005, I analyze the differences in profitability and Tobin's Q between diversifed and focused firms. The empirical findings are in line with previous literature, namely that diversified firms trade at a sizable discount. The data also reveals that diversified firms' Return on Assets (ROA) is on average 6% higher than that of focused firms, after controlling for industry effects. I calibrate the model to this sample and obtain a good quantitative fit to the data, namely the coexistence of a diversification discount and a higher profitability of diversified firms relative to focused firms.

"Social Ties and Economic Development", with Jose Anchorena, April 2008

This paper develops a general equilibrium model where the set of goods includes ties between economic agents (e.g., friendships or acquaintances). We refer broadly to these as social ties. A tie between any two agents is produced according to a technology that uses time from both parties. The model also assumes that social ties contribute toward social capital, which economizes on transaction costs between members of the same community. Our theoretical approach yields the existence of multiple equilibria, which can be interpreted as rational outcomes in societies with different cultural beliefs, in the sense of Greif (1994). We calibrate this model to data on social ties and income per capita for a cross section of 27 countries. Our main quantitative findings are the following: (i) heterogeneity in the average number of social ties can account for a significant portion of the heterogeneity in income per capita across countries; (ii) the model can account for between 1/5 and 1/2 of the changes in use of time in the United States between 1900 and 2000; (iii) according to one measure, the calibrated model implies that without social capital countries would be between 1/2 and 3/4 their actual size in terms of income per capita. Theoretically we have the following additional results: (i) a preference for social ties may significantly mitigate an otherwise large underprovision of social capital; (ii) social capital is in some instances an important source of economic efficiency, very much complementary to labor and/or human capital; (iii) the elasticity of substitution between standard goods and social ties is an important determinant of the observed relationship between social capital and economic development; (iv) in some instances, an increase in productivity causes a decrease in welfare, via amplification of coordination failure in the production of social ties. Finally, in light of this model and our calibration, we conclude that social capital is mainly an externality of the consumption of social ties, and does not result from the agents’ motivation to economize transaction costs.

Work in progress

"Is Growth Risky?", April 2006

The objective of this paper is to investigate how the book-to-market ratio is endogenously related to firm risk, where the latter is proxied by expected returns. Using an exogenous pricing kernel and a simple profit function I obtain closed-form solutions for firm value, optimal investment and conditional expected returns. I find that (i) growth is risky in the time series, i.e. for a given firm; while (ii) the opposite holds in the cross section. The second effect dominates in a simulation environment, which rationalizes the value premium. The ambivalence of the growth-risk relation in the model may explain the prevailing ambiguity in the literature as to whether growth is riskier than value. Also, after taking into account size and book-to-market, the firm’s primitive source of risk (correlation of productivity with the pricing kernel) has small additional explanatory power of the firm’s expected returns. This last result helps rationalizing the empirical finding that beta has low explanatory power in the cross section of expected returns after double-sorting portfolios on size and book-to-market.