I formally relate the consequences of climate change to the time series variation in weather extensively explored by recent empirical literature.  I show that whether conventional fixed effects regressions underestimate or overestimate the effect of climate on actions (such as adaptation investments) depends primarily on whether actions are intertemporal substitutes (owing to resource constraints) or intertemporal complements (owing to adjustment costs).  I also derive the conditions under which fixed effects regressions can recover the marginal effect of climate change on payoffs.  I show that these conditions become less restrictive when regressions control for lags of weather and for forecasts of future weather.