This paper analyzes how environmental policies that aim to reduce carbon emissions affect asset prices and household consumption. Using novel data, I propose a measure of carbon emissions from a consumer point of view and a carbon consumption growth risk measure. The measures are based on information on aggregate consumption and the carbon footprint for each good and service. To analyze the effects of environmental policies, a long-run risks model is developed where consumption growth is decomposed into two components: the growth rate of carbon consumption and the growth rate of the share of carbon consumption out of total consumption. This paper argues that the long-run risk in consumption growth comes mainly from the carbon consumption growth arising from policies and actions to curb emissions, such as the Paris Agreement and the U.N. Climate Change Conference (COP26). My model helps to detect long-run risk in consumption from climate policies while simultaneously solving the equity premium and volatility puzzles. The decomposition of consumption could lead to identifying the most polluting consumption items and to constructing an investment strategy that minimizes or maximizes a long-term environmental criterion.

This paper studies the role of truly independent nonlinear factors in asset pricing. While the most successful stochastic discount factor (SDF) benchmarks pricing the cross-section of stock returns are obtained from regularized linear principal components of characteristic-based returns we show that allowing for substitution of some linear principal components by independent nonlinear factors consistently improves the ability of the SDF to price this cross-section. We use the Fama-French 25 ME/BM-sorted portfolios, fifty anomaly portfolios, and fifty anomalies plus characteristic-based interaction terms to test the effectiveness of the nonlinear dynamic factors. The SDF estimated using a mixture of nonlinear and linear factors outperforms the ones using solely linear factors or raw characteristic returns in terms of out-of-sample $R^2$ pricing performance. Moreover, the hybrid model -using both nonlinear and linear principal components- requires less risk factors to achieve the highest out-of-sample performance compared to a model using only linear factors.

This paper studies anomaly return predictability across deciles using a set of fifty anomaly variables built using individual stock characteristics. I construct deciles and study their predictability using their own past information, other macroeconomic variables, and limit-to-arbitrage variables. I find that some anomalies are persistent and that there are some predictors which help to forecast the decile portfolio returns. Deciles predictability is not uniform across anomaly variables and predictors. Namely, all deciles are not uniformly predictable but extreme deciles seem to be more often predictable. Stock variance, dividend yield, and dividend price ratio are strong predictors for decile portfolio returns. Most importantly, hedge portfolios are often predictable by the TED spread and Amihud illiquidity measure, which indicate that trading frictions may explain the persistence of these portfolio returns. Furthermore, I use the rich set of five hundred anomaly portfolios to investigate their prediction properties using Deep learning techniques.