Stick to facts, not factoids

Let’s stop citing OECD growth forecast

Afactoid you hear a lot these days from conservatives and Conservatives alike is the OECD’S forecast that among all its members Canada will have the slowest economic growth to 2060. That the OECD does forecast this is fact. But conservatives, of all people, are usually skeptical of long-term forecasts. They’re certainly skeptical of long-term climate forecasts. They’re also typically skeptical of big international organizations. So the forecast itself should be regarded as factoid. It’s therefore interesting to see conservatives embrace it as if it were gospel according to Adam Smith, Milton Friedman or Ronald Reagan.

The OECD probably is too big, activist and rich, as many conservatives believe. But it employs many good economists, gives them great support and has been working on long-term forecasting for several years now — partly in response (ironically) to complaints from climate economist William Nordhaus that there are no good long-range economic models to go along with long-range climate models. This is probably because “few scholars or institutions have the temerity to attach numbers to a distant future,” as OECD Policy Paper No 22 (July 2018) puts it before then going ahead and taking a crack at such attachment.

When you hear the OECD has forecast Canada’s economic growth for the next 35 years or so, you think they probably sent a large team here, visited our factories, labs, distribution centres, government operations centres and so on and did a critical analysis of what seemed promising, how well we were introducing best practices, and so on. But, no, economists do this sort of thing sitting at their computer screens, parsing cross-country data and, what economists do best, making assumptions.

In general, three things produce output: labour, capital and the know-how with which they’re combined. You get more output over time — your economy grows — if your labour force grows, your capital stock grows or your knowhow grows. Labour force growth is at least somewhat predictable: you can make informed estimates about births and deaths, labour force participation, education rates and so on. You can also make assumptions about investment rates, and therefore how much your capital/labour ratio grows over time.

Alas, most studies of economic growth find what accounts for it is in the black box: know-how, or, as the OECD terms it, labour efficiency. To forecast labour efficiency the OECD uses a “convergence model.” It assumes countries that are behind in productivity gradually catch up. The farther behind you are, the greater your potential to catch up. Unfortunately (in this regard only!) Canada has high productivity already so is doomed to slower productivity growth than many other countries.

If you look at “trend labour efficiency” in Table 1 of OECD Policy Paper 29 (October 2021), ours is only 0.7 per cent per year for 2030-60 — the same rate as one other country: the U.S., our most common comparator. That our productivity growth has been lagging theirs is one reason the Bank of Canada wants us to break the glass on our productivity emergency. Our having the same forecast productivity growth as they do therefore can’t be terrible news.

True, Europe and the OECD on average are at 0.9 per cent. One reason North America lags is that productivity growth is correlated with trade and the centre of gravity of world trade is shifting to Asia. The OECD estimates the average distance of our trade in 2060 will be 1,750 kilometres longer than now.

Where does labour efficiency grow faster to 2060? Slovakia and Portugal at 1.2 per cent per year. Czechia and Turkey (1.1); Costa Rica, Ireland, Israel, Italy, Korea, Luxembourg, Poland, Slovenia, Spain (1.0), and so on. Small differences do compound over time, but so does the influence of small errors in forecasting.

We trail the U.S. (and others) in growth of real GDP per capita. Our forecast is 0.8 per cent from 2030-60. The Americans’ is 1.0 per cent. Where does the difference come from? From growth of capital/worker (0.2 per cent for us, 0.3 per cent for them), the share of the population that’s active (-0.2 for us, -0.1 for them); and the potential employment rate (no change for us, growth of 0.1 per cent for them). These are not big differences — 0.1 per cent in each case — and some are “natural.” (How much people work is largely a question of age.)

How could we improve? The OECD favours labour market reforms that would get more women into the workforce and persuade more older folk to stay there. But working more hours or years, though it does raise GDP per person, doesn’t necessarily raise well-being.

What’s key to that is higher productivity, which OECD working paper 1385 (from 2017, now a little old) looks at. Canada already has the fourth highest levels of education. We’re best for low and stable inflation (or were from 2006-10, the years the study uses). We regulate our product markets, a productivity drag, more than the U.S., U.K. and the Netherlands (the leaders in this area) but less than the OECD median. Our R&D and patenting numbers aren’t actually bad: 15th and sixth, respectively. In terms of income inequality, whose relation to productivity is tenuous, we’re right in the middle of the OECD pack.

Conclusion? It’s all very interesting work, very professionally done. But policy debates in this country should stick to the facts, not the factoids.