Nick and I argued that Barro's 1979 tax smoothing result is due to his quadratic loss function. We identified what we dubbed a "precautionary taxation" motive when there's a Laffer curve in taxation.