This week’s edition of Embassy newspaper contained a very interesting briefing insert on the Canada-EU CETA talks.Â Below is a commentary from me critiquing the ubiquitous but unbelievable claim that free-trade with Europe would boost Canada’s GDP by $12 billion, create 80,000 jobs, and life incomes by $1000 per family.
$12 billion. That number keeps popping up in debates about the CETA. Itâ€™s the increase in Canadian GDP the federal government claims will result from the deal.
The number has been repeated so often, it now shows up unsourced in routine reporting on the negotiations. CETA â€œis a massive trade deal expected to boost the Canadian economy by $12 billion,â€ opened a recent, typical article.
But where did this factoid come from? Very few reporters or policy-makers know (or bother citing) the original source. It was generated by a computer model built by three European economists, retained by the Canadian and EU governments to work on the 2008 joint economic study of the CETA.
There are now two subsidiary factoids, both derived from the $12 billion claim, that also appear in virtually every DFAIT news release â€“ and trickle down into more stories. Ottawa now claims the $12 billion gain in turn will generate 80,000 new jobs, and boost incomes by $1000 per family.
Letâ€™s learn more about these claims. The European economists used a tool called a â€œcomputable general equilibriumâ€ (CGE) model. It consists of hundreds of mathematical equations assumed to describe various relationships in the economy. By fitting particular numbers to those equations, you can simulate the working of an economy (like an enormous game of SimCity). But remember: a CGE model is not empirically grounded. Any model can be tailored to look like any economy, but the modelâ€™s predictive power depends entirely on the validity (or lack thereof) of its assumed equations.
The European model makes assumptions that are typical of its genre, but not remotely realistic. These include: constant full employment (so no-one can be unemployed due to imports), balanced trade (so a countryâ€™s total output cannot be undermined by a trade deficit), no international capital flows (so companies cannot shift investment abroad), and no impact from fluctuating exchange rates. Quaintly, the model assumes Canada consists solely of one â€œrepresentativeâ€ household, fully sharing income from all sources.Â (I have catalogued and critiqued these assumptions in detail in other work, including Out of Equilibrium, published by the Canadian Centre for Policy Alternatives, and in scholarly form in â€œEconomic Models and Economic Reality: North American Free Trade and the Predictions of Economists,â€ International Journal of Political Economy 33(3), pp. 28-49.)
Eliminating tariffs produces trade gains when each country specializes in what it does best. But this traditional mechanism only explains $2 billion of CETA benefits. The modelers had to go further, with more farfetched assumptions, to boost their prediction. They assume that invisible, unspecified non-tariff barriers will be fully eliminated by the CETA. They assume Canadian service providers will do as much business in Europe as European firms currently do. Finally, they assume Canadians will save a strong share of new income, all of which is invested in new capital here (thus spurring even more growth). This latter effect alone accounts for over half the predicted $12 billion. Given record consumer debt and growing hoards of corporate â€œdead money,â€ this saving-and-investing assumption is downright bizarre.
The subsidiary claim that CETA will produce 80,000 new jobs is more than unrealistic. It is intellectually dishonest. Remember, the CGE model assumes constant full employment. Thatâ€™s essential, because it prevents any loss in total output from a lack of competitiveness. The predicted GDP gains do not come from more employment, they come from higher productivity. By propagating the findings of the model in this manner, DFAITâ€™s spin-doctors are showing they do not understand the methodology of the model they commissioned.
Iâ€™ve discussed this with Richard Harris at Simon Fraser University, who co-wrote the first Canadian CGE trade models. He confirmed that models of this type predict productivity effects, not job-creation. â€œThe argument about gains from trade is about welfare and productivity gains, not employment gains,â€ he wrote. â€œHowever that argument is difficult to use politically to sell trade liberalization, except under normal employment growth.â€
The $1000-per-family factoid is the most outrageous of all. DFAIT simply took the $12 billion gain in GDP and divided it by the number of families in Canada. That assumes that every additional dollar of GDP translates directly into family income. In fact, higher GDP never fully trickles down into income (due to depreciation, retained earnings, indirect taxes, and other leakages). And these days, whatever does land in individual pockets is increasingly captured at the top of the income ladder â€“ never reaching the â€œaverage family.â€
From 1980 to 2010, Canadaâ€™s real GDP more than doubled (dwarfing the tiny impact predicted by the CGE model). Yet real median family income in 2010 was exactly the same as thirty years earlier. How will we get $1000 per family from CETA, when the median family received nothing from 30 years of growth?
The claim that CETA will boost Canadaâ€™s GDP by $12 billion is highly questionable. If the underlying CGE assumptions are relaxed, then the impact of CETA could well be negative (as Iâ€™ve found in my own research). The 80,000 jobs and $1000-per-family claims are worse than dubious. They are dishonest and manipulative, and should not be repeated by serious commentators.
- A Response to the 2017 Saskatchewan Budget (April 23rd, 2017)
- The Alternative Federal Budget 2017 (March 20th, 2017)
- Louis-Philippe Rochon’s Top 10 Economic Predictions for 2015 (January 11th, 2015)
- (Macro) Econ 101 (December 16th, 2014)
- CGE models and carbon tax incidence (November 24th, 2014)