Watson … Maybe we’re a nation of Mike Duffys.
Just in time for Canada Day, the Bank of Canada research department has brought out a study that concludes the country’s underground economy is a lot bigger than previous estimates suggest, and that almost a majority of us engage in the morally dubious practice of under-reporting our incomes. The under-reporting may not go so far as funnelling expenses through speech writer-friends, as it is alleged Senator Duffy did. But it seems the idea that personal gain should be a priority over fastidious accounting is not one you only find in the Red Chamber. Canadians at large are also familiar with the notion.
The Bank’s researchers, Gregory Dunbar, who works both in the Bank’s Currency Department and at the University of Ottawa, and Chunling Fu, who is at the Vancouver School of Economics of the University of British Columbia, estimate that between 35 and 50 per cent of households under-reported their incomes in the years 1998 and 2004 and that the amount of under-reporting amounts to between 14 and 19 per cent of GDP. Today, 19 per cent of GDP represents $376 billion, roughly the size of Quebec’s economy (which is not to suggest Quebec is solely responsible!) and about $90 billion more than the federal government spends in a year. So, not exactly chump change.
Past studies of “sheltered income,” the genteel phrase Dunbar and Fu use to describe the unreported income, have operated by assuming it’s really only possible to hide income in certain kinds of businesses: restauranting, renovation, construction, self-employment and illegal activities like drugs and prostitution (excuse me: “sex work”). Researchers then estimate the scale of these activities — one study suggested that in 2004, there were 17,500 marijuana grow ops in British Columbia alone — make rough guesstimates about how much of each industry’s activity is capable of being hidden, and then tote up to get estimates of the underground economy that usually range from three to five per cent of GDP
The Bank’s study is different. It’s a detailed statistical exercise that compares people’s consumption to their income. Consumption data are taken from two Statistics Canada Surveys, the Survey of Financial Security, which focuses on households’ income, debts, savings, use of pawnbrokers, and so on, and the Survey of Household Spending, which focuses on exactly that, how much households spend on different purposes in a year.
When you weave these data sources together — and the research paper spends about 25 pages describing in very careful detail how the weaving was done, what the potential empirical and logical pitfalls were and how each was addressed — you can compare people’s stated consumption with their stated income. And since StatsCan gives survey respondents the option of letting it link directly to their CRA files, for many people, their stated income is as they stated it legally for income tax purposes.
It turns out in a surprisingly large number of cases, especially among people who don’t have very high incomes, that their consumption exceeds their income, even when a margin for error is built in. “So what?” you say, from time to time, people go through an income dry spell, but they keep on consuming, a deficiency they finance by borrowing. Except that the surveys include data on how much they’re borrowing and their consumption exceeds even that.
Economic theory suggests people generally proportion their consumption to what Milton Friedman christened their “permanent income.” If you see them repeatedly consume more than their current income, you can conclude either that they’re not rational or that their income is actually higher than they have been reporting. The Bank’s researchers aren’t willing to give up on the rationality postulate, so they conclude there’s under-reporting going on.
It’s a purely statistical exercise, so there’s no smoking gun. The StatsCan files are linked demographically, by personal characteristics, not by Social Insurance Numbers, so there’s room for inexactness there. And Dunbar and Fu don’t provide juicy stories or helpful hints about precisely how people go about underreporting their income. It probably wouldn’t do for the Bank of Canada to be publishing a handy Internet guide to defrauding the taxman. Our monetary and fiscal authorities are supposed to work together. (The researchers do provide a tool the tax collector can use to judge the likelihood a taxpayer is under-reporting: If mortgage payments get above 36 per cent of declared income or rental payments above 60 per cent, that’s a telltale signal.)
But, as statistical exercises go, it’s pretty comprehensive. And you can imagine, given the explosive nature of its conclusions, that it has been vetted with even more than the Bank’s usual vigilance.
What are those conclusions? First, don’t just look at the selfemployed for under-reporting of income. They don’t especially stand out in this exercise.
Beyond that, and more radioactively, “reported income as a poverty measure is misleading… (A)bout 65 per cent of households in the low-income range (less than $20K) under-report their income. Thus, tax and transfer policies based on reported income may have unwelcome, and regressive, social-efficiency costs.”
What does that last sentence mean? If you give people income transfers because their stated incomes are low but in fact their actual incomes are higher, that’s socially inefficient and regressive because money is going to people who shouldn’t really be getting it. They’re higher up the income scale than you intend your anti-poverty money to go.
A third possible conclusion, which in fact the researchers don’t make, is that if you raise income tax rates, you may get even more under-reporting.
In an election year, maybe in any year, conclusions like Dunbar and Fu’s are bound to come under political attack. I hope the Bank stands by its researchers and that debate about their findings revolves around their methods and data, not their or the Bank’s motivations, politics or whatever.
New study says under-reported income of Canadians could be as high as 19 per cent of GDP