The perils of artificial precision

One thing you have to admire about the Conference Board of Canada is their chutzpah. Not only is their "Centre for the North" actually located in the south, but their economic forecasts are remarkably ballsy.

One thing you have to admire about the Conference Board of Canada is their chutzpah.

Not only is their “Centre for the North” actually located in the south, but their economic forecasts are remarkably ballsy.

Their most recent press release said “Yukon’s real GDP will rise by 3.7 per cent in 2014.”

If I recall from English 12 at F.H. Collins, “will rise” is the “future simple” tense and is used when one has certainty about the future.

Less self-confident forecasters like the International Monetary Fund prefer the subjunctive and lard their press releases with weasel words like “forecast,” “projected” and “expected.” Their calls are followed by paragraphs hedging their bets with lots of talk of “overshooting” and “downside risks.”

The conference board people probably don’t really mean “will rise.” They know they’ve been criticized for their epic miss in 2013. Their Yukon forecast said “real GDP will increase strongly by 6.3 per cent in 2013.”

Note again that embarrassing future simple. A few months later they had revised 2013 down a whopping 5.7 points to 0.6 per cent. Economic modelers live in fear of incidents like this, where not just the numbers change but also the “key message.” Instead of the robust growth forecast by the conference board, the Yukon barely avoided a recession.

As fun as it is, we shouldn’t beat up on the conference board too much. Nothing is harder to predict than the future, to paraphrase Yogi Berra.

However, the conference board forecasts do highlight the perils of forecasting in general and of artificial precision in particular.

The conference board’s new annual territorial forecasting model has over 300 equations, including lots of behavioural formulae. It breaks our economy up into 25 sectors and uses input-output techniques to estimate the future. They worked around data restrictions in the North using “calibration techniques to estimate key relationships among economic variables in the model.”

This all sounds very sophisticated. It gets them to quite precise estimates. Remember that the missed 2013 estimate was “6.3 per cent” not something like “around 6 per cent.”

However, let’s remember that we live in a very small economy where individual decisions like whether to delay the F.H. Collins rebuild or cut a shift at the Minto mine can significantly impact the economy.

So how should we think about the latest conference board forecast of 3.7 per cent growth in 2014 followed by 3.7 per cent in 2015? And what about their call that there will be four operating mines in 2017, including Victoria Gold’s Eagle mine?

The first thing is that the conference board economists have done some useful work. The figure of 3.7 per cent probably is somewhere in the middle of the range of probabilities.

But we should also ask them what their range of likelihood is, like political pollsters do when they say things like “plus or minus 3 per cent, 19 times out of 20.” The 2013 forecast of “6.3 per cent” would have been much more useful to people if it had been put in context with a range of likely outcomes.

Consider the difference between these two forecasts: “6.3 per cent and we’re 90 per cent sure it will be between 5-8 per cent” versus “6.3 per cent but could range from a recession to 10 per cent growth.”

Unless you have a feel for the range of possibilities, a mid-point estimate is a very dangerous thing to make decisions with.

As for the forecasts that four specific mines will be in operation, that is very gutsy. One could joke that if the conference board economists are really that confident they should quit their jobs and plough their life savings into Victoria Gold shares.

The Conference Board report is also a bit strange in how it has only three paragraphs on government spending tucked away at the end, despite the fact that our billion-dollar transfer payment is the single biggest driver of economic activity in the Yukon. One significant point that media reports missed was the conference board’s prediction that the Yukon government’s “own-source” revenues will grow more slowly than transfer payments between now and 2026.

If true, this means the Yukon government will become even more dependent on Ottawa over the next decade.

In the end, economic models and their forecasts are useful but one has to know their limits. If you’re making big decisions, you need to supplement them with other intelligence. If you ran the social media chatter from Yukon miners or realtors through one of the fancy, new text analysis systems that some hedge funds use, you might get quite a different forecast from the conference board’s model.

Keith Halliday is a Yukon economist and author of the MacBride Museum’s Aurore of the Yukon series of historical children’s adventure novels. You can follow him on Channel 9’s “Yukonomist” show or Twitter @hallidaykeith