Evaluation of similarity models for expected utility violations

A body of work proposes a decision cost argument to explain expected utility (EU) violations based on pair similarity. These similarity models suggest various measures over the risky pairs that define decision costs and benefits. This paper assesses the empirical modeling success of these similarity measures in explaining risky choice patterns showing EU independence violations. We also compare model fit for these similarity models relative to EU and to a selected generalized EU model. Although the candidate models exhibit some degree of substitutability, our results indicate support for models that use relatively simple measures as instruments for similarity.     

The Complementarity between Tax Avoidance and Manager Diversion: Evidence from Tax Haven Firms

We investigate whether tax avoidance and manager diversion are complementary when the costs of diversion are low by comparing dividend payouts, performance, and overinvestments of tax haven firms versus other multinational firms based in countries with weak and strong investor protections. Desai and Dharmapala (2006, 2009a, b) and Desai et al. (2007) set forth a theory of tax avoidance within an agency framework (the D&D theory) based on the assumption that tax avoidance and manager diversion are complementary when the corporate governance system is “ineffective” (i.e., the manager's expected costs of diversion are low). Tax haven firms are corporate groups whose parent firms are incorporated in tax haven countries that are not the countries where the groups’ headquarters or primary operations are located (i.e., their “base” countries). We argue that tax haven incorporation potentially lowers the costs of diversion for managers of firms based in countries with weak investor protections. Using a sample from 28 base countries, we provide evidence that manager diversion and tax avoidance are complementary for tax haven firms based in countries with weak investor protections but not for tax haven firms based in countries with strong investor protections. Our results are consistent with the complementarity assumption underlying the D&D model and provide additional insights into the potential impact of the decentralization of the global firm.     

Estimating the Prevalence and Cost of Yield-Switching Fraud in the Federal Crop Insurance Program

Producers who manipulate and switch their reported crop-yields between separately insured units can increase their insurance indemnities substantially. A statistical model that identifies potential yield switching is developed. The unrestricted statistical model is singular and is identified by imposing a mixture of system-estimable and system-nonestimable restrictions. Lower bound estimates of yield-switching fraud incidence and costs are obtained by applying the model to 207,067 multiple unit producers who purchased crop insurance in 1998.     

Can Normality of Yields Be Assumed for Crop Insurance?

Normality of crop yield residuals pooled to the state and regional level is examined using a procedure that accounts for trend and heteroscedasticity. Normality tests are conducted using farm-level yield data from over 200,000 producers of six crops in seven U.S. states. The results indicate consistent non-normality of crop yields. The effects of assuming normality on insurance premium rates are examined. Assuming normality is found to generate premiums that can differ substantially from premiums derived using data-based empirical distributions.
Montana Agricultural Experiment Station Journal Series No. 2001-7. The Risk Management Agency, USDA, also provided support for this research. The views expressed herein are the authors' and do not necessarily represent those of Montana State University or the Risk Management Agency.
Les auteurs ont analysé la normalité du rendement résiduel des cultures à l'échelon de l'État et à l'échelon régional au moyen d'une méthode tenant compte des conjectures et de l'hétéroscédasticité. Ils ont utilisé les données sur le rendement fournies par plus de 200 000 cultivateurs produisant six cultures dans sept États américains pour effectuer des tests de normalité. Les résultats révèlent l'anormalité constante des rendements. Les auteurs s'interrogent sur les conséquences d'une normalité hypothétique sur les primes d'assurance. En supposant des rendements normaux, on obtient des primes sensiblement différentes de cedes calculées avec une distribution empirique des données.     

A Comparative Study of Alternative Approaches to Estimate Productivity

Theoretically, for single output-single input, annual productivity are expected to be identical across index, non-parametric programming and parametric statistical approaches. The following models within each approach is considered—index (Tornqvist-Theil and Ideal Fisher), the non-parametric programming (Malmquist input, output and graph; Malmquist total factor productivity) and parametric (Input and Output; total factor productivity) regression. Empirically, for single output-single input, this research show differences in annual productivity and productivity growth rate between and within each of the three approaches using Nebraska agriculture data from 1936 to 2004. The annual productivity growth rate from 1936 to 2004 was identical across non-parametric Malmquist output, input, graph and Malmquist total factor productivity, and parametric Malmquist total factor productivity. However annual productivity estimated by parametric Malmquist total factor productivity is identical to Ideal Fisher productivity.

     

Are Crop Yields Normally Distributed? A Reexamination

This article demonstrates that normality test procedures that include individual detrending of short-term panel data can severely reduce the power of normality tests and strongly bias normality tests in a Type II direction. An alternative error component implicit detrending procedure is suggested that demonstrates higher power for the distributions examined. Both procedures are applied to a large data set with normality of yield residuals being rejected. Assuming normality is shown to reduce potential premium rates for a large number of producers in an existing crop insurance product.