The causes of the deficit problems in Ottawa and at Queen's Park are obvious in retrospect: governments thought that the boom years of the mid-2000's would never end. Ontario government tax revenues grew by an average of 7.3 per cent between 2004-05 and 2007-08 – and so did program expenditures.
In Ottawa, the large and recurring budgetary surpluses had become a source of political embarrassment, and successive governments managed the not-especially-difficult trick of turning a structural surplus into a structural deficit – see here, here and here for my attempts to figure out the mechanics of how it happened.
The same pattern repeated itself in other capitals. Looking back, governments should have realized that the boom years wouldn't last forever, and that running surpluses was the appropriate fiscal strategy.
It's easy to explain the politics behind this mistake: political parties are in the business of getting elected, and they took full advantage of the circumstances that allowed them to pull off the combination of cutting taxes, increasing spending and simultaneously balancing the budget. They may well have thought that this was the payoff from the austerity years of the 1990s.
But fact that this was a mistake is only clear in hindsight. It was by no means obvious at the time that the Canadian economy was enjoying the longest and strongest boom in at least 30 years.
Distinguishing the business cycle from the trend is a messy statistical business at the best of times, and trying to do so in real time with fragmentary data that will be subsequently revised is especially challenging. The Bank of Canada provides what are arguably the best publicly available numbers, in the form of its estimates for the output gap. (See here for the most recent estimates.)
According to the most recent estimates, actual GDP was significantly above potential output during the mid-2000s, sometimes as much as 2 per cent above. (The Parliamentary Budget Office and the IMF have similar estimates.) But if you look through the estimates of the output gap that were being published at the time, a very different story emerges.
The Bank of Canada produces its Monetary Policy Report four times a year, and each MPR publishes preliminary estimates for the output gap in the most recent quarter. If you compare the 'real-time' estimates from the MPRs over the last seven years, you get the blue line in the accompanying graph.
Preliminary estimates for the most recent output gap were positive during the mid-2000s, but only modestly so. Governments looking for an indication of where we were in the business cycle would have reasonably concluded that GDP was at or near potential. Large surpluses are for times when GDP is significantly above potential, and there was little reason to think at the time that this was the case. If output is believed to be at potential, a balanced budget is appropriate.
This shouldn't be interpreted as a failure on the part of the Bank of Canada. They are well aware of the very large error bands associated with their output gap estimates – see, for example, Chart 5 in the April 2006 MPR as well as this study by Athanasios Orphanides and Simon van Norden of HEC-Montréal.
Noisy and incomplete data are an inevitable feature of the environment in which policy makers work. So is making mistakes that made sense at the time.
For more of Stephen Gordon's recent posts, click here.
Follow Economy Lab on twitter