John Howard’s biographer, David Barnett, has a piece in the Drum today arguing against the use of fiscal stimulus in recessions and against Keynesianism in general. Setting aside the philosophical and theoretical arguments he makes, I’d like to examine his empirical claim, namely that fiscal stimulus has not worked, and the somewhat peculiar methodology by which he comes to this conclusion.
“It hasn’t worked…. In Australia, unemployment has just gone up, instead of down. At 5.4 per cent it is more than a percentage point higher than it was when the Howard-Costello government went out of office. That is around another 130,000 people without jobs.”
He knows, or should know, that the appropriate comparison isn’t between what was and what is, but rather between what is and what otherwise would have been. To assess the efficacy of a particular course of action we need to know the counterfactual: what would unemployment be today if the government had not implemented its fiscal stimulus? Economists can and do differ about the answer to that question, but to evade it entirely by making a comparison between 2007 and 2010 and thereby imply that all other things remain equal is disingenuous at best.
Interestingly, though, Barnett seeks to establish a counterfactual of sorts by drawing a comparison with Canada. He praises the Canadian Government, saying “Canada did not stimulate. The Canadian government responded to the GFC by cutting back on its expenditure. Canadian exports rose”. Notice how he shifts the goal posts, using exports rather than unemployment as the metric to evaluate the efficacy of macroeconomic policy.
Nevertheless, his argument about Canada provides us with the opportunity to follow Barnett’s own chosen methodology by comparing 2007 (pre-crisis) unemployment with present unemployment and imputing the difference to a failing of public policy. To be clear, I think this approach is not particularly useful, but Barnett seems to be of the view that it is appropriate, so we will follow it.
If we follow this approach, we see that unemployment increased by a greater amount in Canada than in Australia, off a higher base, both in absolute terms and as a proportion of the labour force.
|Unemployment rate – November 2007||Unemployment rate – October 2010||Number of additional unemployed people|
So, on that measure, it certainly seems odd to suggest that a simple comparison of pre-crisis and post-crisis macroeconomic aggregates in Canada and Australia demonstrates the folly of fiscal stimulus.
But what if we accept Barnett’s suggestion that it is international trade that should be the barometer of policy success or failure? Well, our balance on goods and services (our exports less imports) looks pretty healthy to me: