We introduce investors with preferences for green assets to a general equilibrium setting in which they also prefer consuming green goods. Their preference for green goods induces consumption premia on expected returns, which counterbalance the green premium stemming from their preferences for green assets. Because they provide a hedge when green goods become expensive, brown assets command lower consumption premia, while green investors allocate a larger share of their portfolios towards them. Empirically, the green-minus-brown consumption premia differential reached 30-40 basis points annually, and con- tributes to explaining the limited impact of green investing on the cost of capital of polluting firms.

We study ESG and non-ESG mutual funds managed by overlapping teams. We find that non-ESG mutual funds include more high ESG stocks after the creation of an ESG sibling, and the high ESG stocks they select exhibit superior performance. The low ESG stocks selected by ESG siblings also exhibit superior performance and despite being more constrained, the ESG funds outperform their non-ESG siblings. The latter result is consistent with fund families making choices that favor ESG funds. Specifically, ESG funds tend to trade illiquid stocks prior to their non-ESG siblings and get preferential IPO allocations.

Flooding is the most costly natural disaster in the US. To protect collateral value against flood risk, many mortgage borrowers are thus required by law to maintain flood insurance. However, compliance is loosely enforced and lapsed policies are common. This paper hypothesizes that with a high monitoring cost borne by lenders, credit supply will depend on borrowers' insurance incentives. Exploiting an exogenous premium rise ($266 per year) which disincentivizes some borrowers to buy flood insurance, I show that lenders increase the corresponding mortgage denial rates by 0.8 percentage points (3.54% of the mean). This effect is gigantic, 80 times larger than that of lowering a borrower's annual income by $266. I rule out alternative demand-side explanations and provide evidence to support the mechanism that lenders internalize ex-post monitoring costs into ex-ante restrictions on credit supply.