Why so many forecasters get it wrong

By Lethbridge Herald on December 11, 2025.

Roslyn Kunin
Troy Media

Most forecasts fail for one simple reason: they assume people won’t change their behaviour when circumstances change. That mistake keeps producing dramatic but misleading predictions.

One example appeared recently in a news story warning that food costs for the average family will rise by $1,000 in the coming year. Prices for many items are expected to climb, especially meat and notably beef. 

Estimates like this often come from Canada’s Food Price Report, which looks at how much the average household’s grocery bill may change in the year ahead.

How the $1,000 figure was reached is straightforward. Statisticians priced a typical family’s grocery cart and added the expected increases. Buying the same mix of groceries next year would cost $1,000 more. The problem isn’t the math. It’s the assumption that families won’t change what they buy. In reality, shoppers adjust, and the real increase is almost always lower.

Statistics Canada data backs this up. When beef prices rose sharply in 2022, many households cut back or shifted to cheaper meats. Households don’t simply absorb higher costs; they rebalance their budgets in real time. They buy less meat, choose lower-cost options or keep the pricier items and make up the difference by eating out less often. 

Forecasts built on a fixed “shopping basket” overshoot because they assume people stand still even as prices move.

And if grocery prices rise as forecast, we will likely spend more overall, but nowhere near as much as the headline suggests. That’s why dramatic grocery warnings mislead: they ignore how shoppers actually respond to higher prices.

That same faulty assumption shows up well beyond grocery prices.

For example, U.S. President Donald Trump may have forgotten that tariffs are a tax. He treated them as easy revenue, likely by multiplying the value of imports into the United States by the tariff rate.

But when the price of anything goes up, people buy less of it. The same is true for businesses. The expected tariff windfall was never realistic because it assumed buying habits wouldn’t change.

We see the same pattern with workplace benefits. Paid sick leave is a good example. Employers likely estimated the cost by looking at how many sick days people took in the past. 

In Canada, only federally regulated workers are guaranteed paid sick days. Elsewhere, it depends on provincial rules or employer policies.

Once sick days were paid, usage rose. Part of that is positive: workers who couldn’t afford to lose a day’s pay no longer came to work sick. 

But part of it is simply people taking a day for a mild illness because they can. When a benefit is subsidized, use goes up and early cost estimates rarely hold.

Changes to policies or prices rarely stop where planners expect. They ripple through people’s behaviour in ways that often surprise. Governments should consider how people might react before launching new rules or programs, but too often they don’t. 

That matters because budgets and policy plans are regularly built on forecasts that assume nothing will shift, even though real-world choices almost always do.

And that is the core problem: too many forecasts ignore the simplest reality: people adapt. Any model that overlooks that will continue to mislead Canadians.

Dr. Roslyn Kunin is a respected Canadian economist known for her extensive work in economic forecasting, public policy, and labour market analysis. She has held various prominent roles, including serving as the regional director for the federal government’s Department of Employment and Immigration in British Columbia and Yukon and as an adjunct professor at the University of British Columbia. Dr. Kunin is also recognized for her contributions to economic development, particularly in Western Canada.

The views, opinions, and positions expressed by our columnists and contributors are solely their own and do not necessarily reflect those of our publication. 

© Troy Media