There were a couple of puzzling things in my recent post on the sources of the federal defict, and I think I've solved the puzzle. I made a mistake.

Or rather, Chris Matier of the PBO solved the puzzle for me when I asked them if they knew what was going on. What I had missed was a change in the way the numbers were calculated for personal income tax revenues and for transfers to persons. If I had read the March 2006 Fiscal Monitor more closely, I would have noticed that PIT revenues would henceforth be reported gross of payments for the Child Tax Benefit. Previously, PIT revenues had been reported net of the CTB transfers. So starting in April 2006, the numbers for PIT revenues and for transfers to persons would increase by the same amount. The effect on the balance was a wash. Since I used the April 2006 FM as the source of my April 2005 numbers – that is, after revisions – my series started their rise in 2005.

This explains both the rise in PIT revenues and in transfers to persons I found in that previous post. When you use the numbers in Cansim's Table 380-0007, the surge in PIT revenues in 2005-06 disappears:

As does the increase in transfers to persons:

(That spike in 2005 was the one-time payment I removed from the graphs in the earlier post.)

So the conclusion to that last post should have pointed to

• Transfer payments to provinces that increased by about 0.5% of GDP
• Lost GST revenues of about 0.75% of GDP

I think I have it right this time. The sum of these two effects is approximately the size of the swing in the government balance between 2004 and 2008.