Trevor Tombe: How to create two jobs for every Canadian worker

With each new subsidy, tax incentive, or project, we’re told some quantity of jobs follows. No doubt you’ve seen the headlines. The Trans Mountain Pipeline can create 3,300 permanent jobs and East can create over 4,250. Bombardier, which is looking for government financial support, claims responsibility for 64,800.

The numbers aren’t pulled from nothing, they are available from what exactly are called economic impact studies. They are utilised in marketing campaigns and often are even mandated by governments. The application process for Alberta’s new subsidy for petrochemical plants, for example, requires them.

Unfortunately, such research is easy to misinterpret, rarely explained, and may make bad public policy look great. Let’s explore their limitations.

What’s really the net effect of the new petrochemical plant on overall employment? Roughly zero.

Suppose you spent $1 million on beer. This spending would employ a brewer or two, however it doesn’t stop there. Breweries must also buy inputs – grains, fuel, and so on. Suppliers then employ more workers, and themselves buy more inputs. The cycle continues on as well as on.

Statistics Canada does the math in what is called an “Input-Output Model.” For breweries, $1 million in sales will “create” over $1.8 million as a whole sales and 4.6 jobs throughout the economy. These are called the direct plus indirect effects.

Related

Trevor Tombe: Alberta gets a ‘fair’ share of royalties, after allJoe Oliver: Canadians should beware the effects of global warming – policies

Next, workers are paid income, which they’ll spend, spurring yet another round of sales, hiring, and input purchases. These are called the induced effects. For breweries, it adds another 1.4 jobs. Overall, that’s just over six jobs per $1 million in new brewery sales.

Unfortunately, these studies have serious problems.

First, why consider only worker incomes when calculating induced effects? How about business owners? Won’t additionally they spend additional incomes? When we account for this, something interesting happens. The more business income that will get recycled as spending, the larger the induced effects.



And there’s no limit. Under some tweaks towards the model, an additional $1 in spending can imply millions C heck, billions C of recent jobs! If both workers and firms spend all of any additional income, the “induced” effects become infinite!

Obviously, this really is silly. While students learn this startling result, the substitute nature of induced effects is not widely appreciated.

This raises the next problem: The models assume personnel are in limitless supply. Accumulate all of the direct, indirect and induced jobs in the Canadian economy using Statistics Canada’s standard numbers and also you get about 36 million jobs. No matter that there are only 18 million workers for everyone, the model doesn’t care.

In reality, total employment is largely pinned down by how many workers you will find. More jobs in a single part of the economy mean fewer jobs in another. Interest in labour is important, obviously, but same with supply.

So what’s really the net effect of a new petrochemical plant on overall employment? Roughly zero. Workers, with few exceptions, would otherwise have been employed elsewhere. The plant doesn’t increase employment, it shifts employment.

This leads us to a final point: The models take into account benefits, although not costs. Want a subsidy? Point to the concentrated and visual gains. Desire a larger subsidy? Point to large spillover gains from indirect and induced effects.

The costs, on the other hand, are less visible C though they’re no less real. Once we saw, workers that shift into one activity are no longer available to operate in another. Money to supply subsidies should also come from somewhere. Typically, the money is raised through distortionary taxes that lower business activities. Income taxes lower employment, and corporate taxes lower investment. These costs are absent from typical economic impact studies.

So how can we improve? Better models are available. Instead of simple “input-output” models, we are able to use what are known as “general equilibrium” ones. They’ve benefits, including recognizing that personnel are in limited supply.

We can also put excess fat on income and productivity, and fewer on job creation claims. Producing electricity having a fleet of stationary bicycles, after all, would employ a military of workers but is hardly sensible policy.

Finally, we ought to push back against those making dodgy claims. Results from economic impact studies, particularly those regarding induced effects, aren’t what they are cracked as much as be.

Trevor Tombe is definitely an Assistant Professor of Economics in the University of Calgary.