Prices and Portfolio Choices in Financial Markets: Theory and Experiment∗

Most tests of asset pricing models address only the pricing predictions – perhaps because the portfolio choice predictions are obviously wrong. But how can asset-pricing theory be right if the portfolio choice theory on which it rests is wrong? This work builds and observes experimental markets in which risky and riskless assets are traded. Risk aversion is a robust phenomenon in experimental settings, thus offering an opportunity to study asset pricing in an environment where the crucial variables can be observed or controlled, unlike in the field. The experimental data are consistent with historical field data not only in generating sizeable risk premia, but also because portfolio choices are wildly divergent from theoretical predictions, e.g., portfolio separation of CAPM. Yet there is clear support for the associated pricing predictions: the market portfolio becomes mean-variance efficient, for instance (the central pricing prediction of CAPM). Apparently, the experimental markets exhibit the price-allocation puzzle implicitly assumed in most tests of asset pricing models. We build a structural model that can explain the price-allocation puzzle. Following a route familiar from much applied work, it is based on perturbations of individual demand functions. We develop an econometric test of the structural model. Unlike field data tests, ours verifies the consistency between prices and portfolio choices. In addition, it relies on large cross-sections (of agents) and not on long time series. The empirical distribution of our test statistic across 83 samples is consistent with the theory, except for a tiny bias in pricing evidently induced by shortsale constraints.