A preliminary estimate for Canadian 2010Q1 GDP growth

It's time to update my series of posts (2009Q1, 2009Q2, 2009Q3, 2009Q4) in which I try to take the GDP numbers from the first two months of a given quarter, mix them with the LFS numbers from the third month, and boldly produce what Statistics Canada dares not attempt: a preliminary estimate for quarterly GDP growth. The BEA released theirs yesterday (3.2%); here's mine: 5.5%.

This is the fifth time I've done this, so let's all take it as read that that the result of this exercise is a probability distribution, not a single number. Its standard deviation is 0.5%, and the interquartile range is [5.1% , 5.8%].

And since I've been doing this for awhile, it's time to see how these preliminary estimates have held up. Here is a table comparing the 'backcasts' with the numbers produced by Statistics Canada:

Quarter	WCI estimate	First StatsCan release	Latest data
2009Q1	-6.9%	-5.4%	-7.0%
2009Q2	-3.4%	-3.4%	-3.5%
2009Q3	-0.4%	0.4%	0.9%
2009Q4	4.0%	5.0%	5.0%

For comparison purposes, here is a table comparing the preliminary BEA releases with most recent data (that is, after revisions that have been made so far):

Quarter	Preliminary BEA estimate	Latest data
2009Q1	-6.1%	-6.4%
2009Q2	-1.0%	-0.7%
2009Q3	3.5%	2.2%
2009Q4	5.7%	5.6%

So far, my worst number was for 09Q3, where I underestimated growth by 1.3 ppts. Then again, the BEA's preliminary number for that quarter was off by 1.3 ppts. If the BEA can live with errors of that size, then so can I.

Comments

[brendon]

My model (a quarterly projection model with an dynamic IS-LM structure) says 4.9 - 6 with a mean of 5.4

[Stephen Gordon]

Dang. Always makes me nervous when other models give similar results...

[Just visiting from Macleans]

Ummmm. Stupid question. Who cares about the accuracy of predicting history?

In other words, why should non economists care?

[brendon]

well, my model is often right, if that helps.

[Stephen Gordon]

Ummmm. Stupid question. Who cares about the accuracy of predicting history?

In other words, why should non economists care?

Policy decisions are contingent on the current state of the economy. So it's of interest to know the current state of the economy.

[Nick Rowe]

Yep. Controlling monetary policy is like driving a car, trying to keep the speed constant, when you can only look in the rearview mirror, and the speedomoeter only tells you want speed you were doing last week. So anything that brings the speedo a little bit more up to date can help.