Carmen & Vincent Reinhardt have a paper (PDF) about an empirical survey of what happens in the wake of big financial crises. It’s all bad news—you’re looking at about a ten-year span of deleveraging and reduced growth on average, and it’s not rare for per capita income to remain below pre-crisis levels at the end of that period. Near the end they express some agnosticism as to causation:
What we observe, of course, is an association. Growth falls and the
unemployment rate remains high after a severe economic dislocation. That observation, itself, is not informative as to the balance between changes in aggregate demand and
The outcome could materialize as a consequence of the failure of policy makers to
provide sufficient stimulus after a wrenching event in an economy where rigidities give
ample scope to demand management. An important role for credit in supporting
spending might imply that an associated collapse in financial intermediation lengthens
and deepens the downturn (with the unavailability of credit serving as the propagating
mechanism discussed in Bernanke, 1983). In such circumstances, slow growth might be a self-fulfilling prophecy produced by timid authorities who neither supported spending nor dealt with the capital-adequacy problems of key financial institutions.
I think their data actually sheds more light on this question than their analysis suggests. Negative shocks have both real and nominal elements. If governments tend to respond to big shocks with too much stimulation of demand, we should see higher-than-usual inflation as a result. Conversely, if governments tend to respond to big shocks with too little stimulation of demand, we should see lower-than-usual inflation as a result. And Reinhardt & Reinhardt show that unusually little inflation is the typical result:
So that seems pretty clear-cut to me. Like the United States today, most governments faced with big financial crises in the postwar era engage in a quantity of demand-stimulation that’s insufficient to offset the purely nominal impact of the negative shocks. That’s not to deny that real shocks are also present or that limits exist to what demand-stimulation have accomplished, but simply to point out that government usually don’t test those limits and thus end up inflicting avoidable suffering.