It is well-established that governing parties, on average, lose votes. Yet what explains this cost of governing remains largely unknown. We propose better measurement, as the commonly used dichotomous measures do not properly reflect the true range of government participation across political contexts. This research note brings in portfolios held by each party. By differentiating in this way between parties, we obtain novel insights into the cost of governing. Furthermore, we estimate how the effects of well-known factors on the cost of governing depend on both the number and type of portfolios a party holds. To do so, it leverages an original party-election year and post-election government formation data set in 40 democracies from 1966 to 2021 which calculates new continuous and categorical measures of government participation. We demonstrate that our measures add predictive power across a variety of contexts and show how the economy affects the cost of governing.