The market's beliefs about its own future evolution are an essential determinant of asset prices. Yet, these beliefs cannot easily be identified from prices as they are altered by stochastic discounting. Under conditions described by Ross (2015), beliefs and the discount factor can be jointly and uniquely disentangled from market prices of financial derivatives. The approach has been widely criticized for its poor performance on empirical prediction tasks. In this paper, I show how to impose ordinal constraints on the stochastic discount factor to mitigate the method's fragility. To that end, I introduce a new class of matrices whose Perron vector meets an ordinal condition on the sequence of its entries. Using simulations, I show that the original recovery method fails at approximately identifying true probability distributions from prices, while the constrained version produces reliable predictions. Finally, using data on Futures and its European options prices on the S&P 500 between 2007-2021, I compare different methods to predict the distribution of future returns, including known variants of Ross' recovery. The results suggest that ordinal constraints on the implied stochastic discount factor make the method a competitive forecasting instrument.

This paper investigates how consistently investors use their beliefs in making trading decisions. To this end, we design a customized online experiment that records contemporaneous forecasts and investment decisions. Our findings reveal that decisions to sell are substantially less beliefs-driven than decisions to buy. This discrepancy can be largely attributed to selling decisions made in the face of paper losses. The larger inconsistency in belief utilization complements potential shifts in investor risk preferences, corroborating the existing literature. We further find that in the presence of some return predictability, lower belief-use consistency adversely affects investment performance.

We demonstrate that investor satisfaction and investment behavior are influenced substantially by the price path by which the final investor return is achieved. In a series of experiments, we analyze various different price paths. Investors are most satisfied if their assets first fall in value and then recover, and they are least satisfied with the opposite pattern, independent of whether the final return is positive or negative. Price paths systematically influence risk preferences, return beliefs, and ultimately trading decisions. Our results enable a much more holistic perspective on a wide range of topics in finance, such as the disposition effect, risk-taking behavior after previous gains and losses, and behavioral asset pricing.