Are forecasters too bullish?

Here is the latest from the Conference Board:

Its outlook projects Canada’s economy to grow 1.7 per cent this year – a far more bullish prediction than the Bank of Canada, which on Tuesday revised downward its growth forecast to one per cent this year.

What is interesting is how the CP report calls them “bullish”. Back in February, I testified before the House  Finance Committee on the economic outlook and what it means for the federal budget. Glenn Hodgson of the Conference Board, with whom I testified, thought I was too pessimistic and had this to say:

We were actually the most optimistic of all the private sector forecasters right now on our Canadian outlook. We think the Canadian economy can grow at 2.5% this year, but it will be very uneven sectorally and geographically. So the west is best–we think 3% growth or beyond is quite attainable for all four western provinces. Central Canada is very challenged because of the heavy reliance upon sales to U.S. consumers. The industrial heartland is very much challenged. I’m sure you’ve heard from many manufacturers that have talked about the challenges they’re facing.

Not that the Bank of Canada is much better. Its revised forecast of 1% is well below its own estimates earlier this year, which were in the mid-2% range.

I’m pretty skeptical of economic forecasters – the same ones who repeatedly get quoted in stories about the economy without any reference to whether their track records merit such media play – though I recognize that their job is pretty difficult. But it seems to me that when slowdowns or recessions come on they are almost uniformly missed by forecasters. If it is sunny today the best forecast for tomorrow is more sun. I’d like to test this out statistically, so if anyone out there wants to do some pro bono data gathering, let me know.


  • Hi Marc,

    I believe you’re on to something. I think analysts are generally trend followers. I’ve done some data crunching on the trader positioning reported in the Commitments of Traders data from the U.S. Commodity Futures Trading Commission. (See my blog.) It generally shows the “large speculators” (investment firms and hedge funds) are positioned the wrong way at key market junctures. Specifically on your question, I know I’ve read some studies on the phenomenon of analyst excessive bullishness, but a quick check of my own files didn’t turn up those reports. I’m sure you could find them easily enough by Googling around a little.


  • Hi Marc,
    I think labour market data, combined with “real” investment outlays gives a worthy trend line to use when you are asked to comment on the future performance of the economy. If full-time job creation turns down, as Erin showed in a post here it has in the last three months, that is an indication of a slowdown. Business’s invest when they expect a profit. When they are not investing … and the labour market is turning down, you can safely predict a slowdown.
    John K. Galbraith, who else, said it best about forecasting all the time for a living. There are two types of economic forecasters, those who know they are wrong, and the ones who don’t.

  • In chapter 5 of “The Death of Economics”, Paul Ormerod refers to a survey published by the OECD in June 1993 (as part of their regular ‘Economic Outlook’ if I’m reading his limited bibliography correctly) where various IMF/OECD/National Government forecasts were compared to a naive forecast that next year’s growth & inflation would be the same as the last over the period 1987-1992 and it was found that the naive forecast was just as good or better than the models.

    Don’t know if they’ve tried again since then or not.

  • There is one file that I would keep an eye on, and that is the payroll file. It feeds into the SEPH and is an administrative data file. That file to me is the one that carries a lot of weight, but is typically under-represented in the media as a relevant indicator. Like any administrative data source it has bumps and pot holes, but to me is far superior than any survey for forecasting. It is fairly timely and does not suffer from many of the issues that other indicators do.

    I do also agree with Cameron D. that the LFS measure of full time jobs is a good indicator as well.

    GDP estimates are about as useful as flaked metal in grease. The data sources in the near term GDP estimates are about as accurate as an elephant gun.

    By the way did you happen to notice the elephant sitting silently by the side of the Bank of Canada’s inflation report. I think his name was full time employment. I wonder what they feed him to keep him so silent?


Leave a Reply

Your email address will not be published. Required fields are marked *