More nonconcavities in information processing functions

The productivity of (human) information processing as an economic activity is a question that is raising some interest. Using Marschak's evaluation framework, Radner and Stiglitz have shown that, under certain conditions, the production function of this activity has increasing marginal returns in its initial stage. This paper shows that, under slightly different conditions, this information processing function has repeated convexities with ongoing processing activity. Even for smooth changes in the signals' likelihoods, the function is only piecewise smooth with non-differentiable convexities at points of conditional changes of action. For linear likelihood functions the processing value proves to be piecewise linear with convexities at these levels. This revised version was published online in June 2006 with corrections to the Cover Date.     

A common-features analysis of Amsterdam and London financial markets during the eighteenth century

We examine the financial linkage between the London and Amsterdam financial markets using stock prices recorded in each market over the period 1723-94 in conjunction with tests for common trends, cycles, and regime shifts. These tests reveal a surprising degree of integration between the markets as their prices move together in both the short and long run. Moreover, shocks to the assets translate quickly and accurately between markets. It also appears that Dutch investment did not destabilize London markets and stock prices in London were the primary determinant of prices in Amsterdam.     

Prevalence and Correlates of Gambling Problems Among a Nationally Representative Sample of Brazilian Adolescents

The objectives of the study are: (a) to provide the first prevalence estimates of pathological gambling among Brazilian adolescents using an age-specific instrument in a nationally representative sample; (b) to investigate the extent to which adolescents participate in gambling activities in a developing country; and (c) to correlate different levels of gambling behavior with demographic variables. Multistage cluster sampling selected 3,007 individuals over 14 years of age from Brazilian household population. A total of 661 participants were between 14 and 17 years old. The Lie/Bet Questionnaire and the DSM-IV-MR-J were used for assessing problem and pathological gambling. 2.8% scored positive on the screening questionnaire, while 1.6% were classified as problem and pathological gamblers. Factors associated with problem and pathological gambling were male sex, not currently studying and considering religion as not important. Less than 4 months elapsed between the age of regular gambling involvement and the first gambling problem. Prevalence rates were quite similar from recent studies which used nationally representative samples. The association of problem and pathological gambling with male sex, school drop-out and low religiosity supports the Problem Behavior Theory. The fast progression to problem gambling adds evidence that adolescents may be more vulnerable to the effects of gambling.