The International Monetary Fund, or IMF, has some interesting new estimates out on the size of different countries’ shadow economies.

We in Canada are about to undergo a rare economic experiment in which part of the economy that has been lurking underground—the marijuana production and distribution industry—is about to come above-ground, assuming the Senate cooperates in passing legalization legislation. How and how much that changes GDP will be interesting. It has already increased it a little because Statistics Canada has been paying people to gather data about the industry, including by collecting crowd-sourced data on the current price of weed, which turns out to be substantially less than the $10 per gram most governments are planning on.

The IMF’s definition of the shadow economy doesn’t include illegal activities, however, but rather activities that are in themselves legal but which aren’t being reported, whether for regulatory or tax reasons. In some countries, the shadow economy rivals the reported economy in size. In Bolivia, for instance, according to new estimates (at Tables A.1 and A.2) the shadow economy averaged 62.28 per cent of GDP between 1991 and 2015. In Georgia, 64.87 per cent. In Zimbabwe, 60.64 per cent. (Taking these estimates to the second decimal place is definitely a case of spurious exactness.)

By comparison to these countries, Canada does pretty well, though probably not as well as our self-image as boy and girl scouts might lead you to expect. Our average from 1991 to 2015 was 13.92 per cent of GDP. Switzerland was tops, at just 7.24 per cent. The United States was not far behind, at 8.34 per cent. Austria (8.93 per cent), Japan (10.41 per cent), the Netherlands (10.77 per cent), the United Kingdom (11.08 per cent), New Zealand (11.70 per cent), Singapore (11.90 per cent), Australia (12.06 per cent) and a few other places also did better than we did (assuming having a small shadow economy is best, which you might not think if you’re in it!).

The good news, however, is that our shadow economy has been shrinking. In 1991 it was fully 19.31 per cent of GDP. Since then it has fallen more or less steadily—though with a couple of hiccups—to the point where in 2015 it was 9.42 per cent of GDP, its low for the quarter century.

Any estimate of the shadow economy is going to be iffy, of course. It’s the nature of the beast that people don’t want to reveal their participation in it. So how does the IMF measure it? There are several well-known techniques, among them:

• Surveys of one kind or another that ask people, on the basis of anonymity, to own up themselves or say how much of it they think their friends are involved in.

• Currency demand estimates, which assume most shadow transactions take place in cash and look at whether the amount of cash in circulation is more, less or just about right compared to what’s needed to finance all above-ground activity.

• Comparisons between production data and consumption data, on the assumption shadow-participants care more about, and are better able to hide, their production rather than their consumption.

• Comparing electricity use with official GDP, on the assumption that electricity use grows more or less apace with economic activity. If all of a sudden electricity use spikes compared to GDP, maybe more GDP is being produced than people are letting on. An interesting new variant, which is used to help generate these latest estimates, is to look at nighttime light from satellite sweeps as an indicator of GDP.

• Mixed methods, which relate the shadow economy to economic indicators known to be correlated with it, such as labour force participation rates, marginal tax rates, indices of regulation and so on.

The technique the IMF uses in this study is to combine mixed methods, which can show how different countries’ shadow economies evolve compared to one another, with currency demand estimates that produce an indicator of the absolute size of the shadow economy within each country.

In essence, our shadow economy has been getting smaller because the indicators traditionally thought to be associated with it have been declining. Most importantly, I suspect, though the IMF report doesn’t say so explicitly, marginal tax rates have fallen from their most recent peaks in the late-1980s, when governments were working hard to get control of their deficits.

If that’s the case, it makes you wonder whether the decline in Canada’s shadow economy will continue after 2015, the final year covered in this study. After a steady if modest decline in taxes over the previous 20 years, governments started raising them again, especially at the top end of the income distribution, it won’t be surprising if both the IMF’s model and the real world throw a bigger shadow post-2015. But to know that for sure we’ll have to await the 2020 or 2022 version of this study.