I've just come across this report from Goldman Sachs examining 44 major fiscal corrections in the OECD since 1975. I'm not sure what to make of it, except that it seems to be getting conclusions from the data that satisfy the authors' prior beliefs.

In a review of every major fiscal correction in the OECD since 1975, we find
that decisive budgetary adjustments that have focused on reducing
government expenditure have (i) been successful in correcting fiscal
imbalances; (ii) typically boosted growth; and (iii) resulted in significant bond
and equity market outperformance. Tax-driven fiscal adjustments, by contrast,
typically fail to correct fiscal imbalances and are damaging for growth.

Their analysis includes case studies of Ireland, Sweden and Canada that show there is another possible explanatory variable - change of government. These three countries, which had large turn arounds in growth, all had change of government before the fiscal contraction was put in place. It is quite possible that the animal spirits that provoked the change in government, themselves led to renewed activity in the economy.

There's another implication for government spending that is strong in their analysis but doesn't even make it to a single phrase in their conclusion, which is what makes me suspect that they have a conclusion and interpret the data to fit that conclusion. In their regressions of average growth relative to OECD (Table 1) the strongest coefficient is for government investment i.e. government spending on capital works has a strong effect on subsequent growth than either increases in tax or reduction in current government spending.

So how to prime growth? Shift government expenditure from current spending to investment and cut taxes. But my caveat still holds - none of their regressions are believeable while we have a major regression variable not taken into account - change of government.