Classical approaches to estimation and decisions requiring estimation often are at odds. When values critical to the decision are convex or concave functions of unknown parameters, the statistician's estimation error adjustments are the opposite of what is appropriate for the decision. We illustrate the conflict by studying multi-period investment problems. The proper application of Jensen's inequality to the decision turns finance intuition on its head: Multi-period investments with negative risk premia can be profitable, risk-averse investors can have infinite demand for risky securities, settings exist in which risk-averse investors should not diversify, and demand for mutual funds with negative alphas may be rational.

[When agents can learn about their abilities as active investors, they rationally "trade to learn" even if they expect to lose from active investing. The model used to develop this insight draws conclusions that are consistent with empirical study of household trading behavior: Households' portfolios underperform passive investments; their trading intensity depends on past performance; and they begin by trading small sums of money. Using household data from Finland, the article estimates a structural model of learning and trading. The estimated model shows that investors trade to learn even if they are pessimistic about their abilities as traders. It also demonstrates that realized returns are significantly downward-biased measures of investors' true abilities.]

Using geographic differences in the availability of payday loans, I estimate the real effects of credit access among low-income households. Payday loans are small, high interest rate loans that constitute the marginal source of credit for many high risk borrowers. I find no evidence that payday loans alleviate economic hardship. To the contrary, loan access leads to increased difficulty paying mortgage, rent and utilities bills. The empirical design isolates variation in loan access that is uninfluenced by lenders' location decisions and state regulatory decisions, two factors that might otherwise correlate with economic hardship measures. Further analysis of differences in loan availability—over time and across income groups—rules out a number of alternative explanations for the estimated effects. Counter to the view that improving credit access facilitates important expenditures, the results suggest that for some low-income households the debt service burden imposed by borrowing inhibits their ability to pay important bills.

We study individual portfolio choice in a laboratory experiment and find strong evidence for heuristic behavior. The subjects tend to focus on the marginal distribution of an asset, while largely ignoring its diversification benefits. They follow a conditional 1/ n diversification heuristic as they exclude the assets with an “unattractive” marginal distribution and divide the available funds equally between the remaining “attractive” assets. This strategy is applied even if it leads to allocations that are dominated in terms of first-order stochastic dominance and is clearly irrational. In line with these findings, we find that framing and problem presentation have substantial influence on portfolio decisions.

[We document the cross-sectional properties of corporate dividend-smoothing policies and relate them to extant theories. We find that younger, smaller firms, firms with low dividend yields and more volatile earnings and returns, and firms with fewer and more disperse analyst forecasts smooth less. Firms that are cash cows, with low growth prospects, weaker governance, and greater institutional holdings, smooth more. We also document that dividend smoothing has steadily increased over the past 80 years, even before firms began using share repurchases in the mid-1980s. Taken together, our results suggest that dividend smoothing is most common among firms that are not financially constrained, face low levels of asymmetric information, and are most susceptible to agency conflicts. These findings provide challenges and guidance for the developing theoretical literature.]

This article offers a unified approach for studying political violence whether it emerges as repression or civil war. We formulate a model where an incumbent or opposition can use violence to maintain or acquire power to study which political and economic factors drive one-sided or two-sided violence (repression or civil war). The model predicts a hierarchy of violence states from peace via repression to civil war; and suggests a natural empirical approach. Exploiting only within-country variation in the data, we show that violence is associated with shocks that can affect wages and aid. As in the theory, these effects are only present where political institutions are noncohesive.

We build a game-theoretic model where aggression can be triggered by domestic political concerns as well as the fear of being attacked. In the model, leaders of full and limited democracies risk losing power if they do not stand up to threats from abroad. In addition, the leader of a fully democratic country loses the support of the median voter if he attacks a non-hostile country. The result is a non-monotonic relationship between democracy and peace. Using Polity data, we classify countries as full democracies, limited democracies, and dictatorships. For the period 1816–2000, Correlates of War data suggest that limited democracies are more aggressive than other regime types, including dictatorships, and not only during periods when the political regime is changing. In particular, a dyad of limited democracies is more likely to be involved in a militarized dispute than any other dyad (including “mixed” dyads, where the two countries have different regime types). Thus, while full democratization might advance the cause of peace, limited democratization might advance the cause of war. We also find that as the environment becomes more hostile, fully democratic countries become more aggressive faster than other regime types.

This paper provides evidence that firms that have consistently met or beaten analysts’ earnings expectations (MBE) provide more frequent “bad news” management forecasts than firms with no established string of MBE, particularly when existing analyst forecasts are optimistic. This suggests that firms with a consistent MBE record are more likely to guide analysts’ expectations downward to avoid breaking the consistency. Subsequent analyst forecast revisions following bad news management forecasts issued by these firms are dampened, implying that analysts suspect that these forecasts may be opportunistic. The relation between management forecasts and MBE consistency is stronger after Regulation FD.