Previous research has found that there is a link between local yearly weather deviations and climate change perceptions of the public. This study explores whether policy stances of MP candidates are affected by random variation in local yearly weather. We find negative relationship between how important matter climate change is for Finnish MP candidates and yearly deviation in local snow fall. Snow fall is a salient weather measure and local yearly deviation are random thus establishing a causal link. We use election advice application data, where candidates answer set of policy questions on a scale before the election, of Finnish MP candidates in order to get a measure of climate change (policy) attitudes. In order to explore the mechanism behind, we exploit the fact that in Finland election constituencies are geographically large, meaning voters and candidates do not necessarily share the same local climate. Using mean snow fall deviation of candidate’s voters, weighted by the share of candidates votes she got from the municipality, we find that the coefficient for candidate’s home municipality weather becomes insignificant whereas the coefficient for mean weighted weather of voters stays negative and significant. This suggests that candidate react to the voters, not the weather they experience themselves, when forming their climate change policy stances.