We examine the effects of a government electricity conservation campaign on household electricity consumption in a non-experimental setting. Before the Fukushima nuclear accident in 2011, one-third of electricity was generated by nuclear power plants in Japan, but all the plants had shut down by 2013 for mandatory regular checkups. To meet severe supply constraints without increasing the price, the government ran a massive conservation campaign with explicit targets by electric company districts. Exploiting the differential timing of the pre-determined checkup cycle and the heterogeneous dependence on nuclear power generation in the pre-accident period by the electric company districts, we estimate the impact of the target on household electricity demand. We also estimate the price elasticity exploiting the price increase that occurred with a time lag. We show that increasing the conservation target by 1 percentage point reduced the consumption by about 0.4\%, whereas the price elasticity was around 0.3. Households with extra electricity usage complied more strongly with the conservation request while the responses were uniform across income groups.