Hardly a week goes by when the pages of our nation’s newspapers aren’t graced with someone worried about our productivity gap. Econowatch has been a bit of an exception, noting that there are some issues with the ways in which we measure productivity which imply we probably shouldn’t worry about it as much as we do.
I believe there’s a factor in the productivity story to which we don’t pay anywhere nearly enough attention: the misuse of Statistics Canada’s Input/Output model to keep telling us that costs are benefits and that productivity doesn’t really matter. A wonderful example of this was provided last week by the Canadian Energy Pipelines Association (CEPA). Lest you think I am beating up on the fossil fuel industry, let me assure you that the obsession with low productivity does not stop at fossil fuels; the green jobs crew more than pull their weight.
The CEPA study is the perfect example of what’s wrong with using I/O models to measure the benefits of infrastructure, either existing or proposed. Ideally, pipelines would be cheap to build, would use little energy, and thus would allow resources to reach their markets at minimal cost, maximizing the value realized from extraction of our scarce resources. By enhancing the value of resources, efficient transportation would enhance the implied productivity of labour in the resource sector. The method used in the CEPA study would value a pipeline system which was costless to operate at something approaching zero, and leave us all wishing for more expensive pipelines so that they’d have more economic impact. If you can’t see how that’s backwards, you’re just not giving it enough thought.
The CEPA study looks at the annual operating revenues of pipelines, and treats those as a shock to the economy – money which would not otherwise be spent. Of course, nothing could be further from the truth. Operating revenue of pipelines is revenue not earned by resource production or refining, depending on the dynamics of the particular energy market, or money not saved by consumers at the pump. This is not money which is magically added to the economy. Rather, it’s part of the value of resources lost to the costs of transporting them to market. The higher the transportation costs, the lower are taxes, royalties, net revenues to producers, margins of refiners or the higher are consumer costs. These lower margins would either mean less employment or lower wages or both in the upstream sector, while higher costs would lower welfare for consumers, all as a result of the lower productivity that we’re told is a source of great benefit to the economy. It’s as close to zero sum as you can get.
I wrote a few years ago about the economic case for an oil bucket brigade, but I never thought anyone would take it seriously. Of course, it’s a ridiculous option because it would have absurdly high operating costs and would negate any value in the resources being transported via inflated transportation costs. However, if you used the method CEPA uses, a shift from pipelines to bucket brigades would produce a boom of epic proportions. The logical conclusion of their study is that, if the current transportation network is a large benefit to the economy, then a more expensive transportation network would be an even larger benefit.
Perhaps, with this in mind, we should consider shutting down the existing pipeline network and replacing that low-cost, high-efficiency transportation with oil transport by rail or truck? At the very least, we should clearly abandon new pipelines in favour of transportation options which will generate higher total revenues. Somehow I doubt that’s the point that CEPA was trying to make.
The more disturbing upshot of the use of I/O models in this way is what these methods would tell you about a true shock to the economy delivered from the oil transportation system. An oil spill, in the spirit of broken windows, would show up in an I/O model as a veritable economic boom. A billion dollar clean-up bill would ripple through the economy and employ people in the same way as any other expenditure, but in no way would it be a net benefit to society. Sure, people would be employed in the clean-up, and businesses would see demand for their services increased, but these costs would be paid for by insurance companies, pipeline company shareholders, and other members of our society. This is money that would otherwise be spent on other things, which would now have to be deployed restoring a landscape to something resembling its previous state, and not adding any net value. Not to mention, of course, that the environmental costs of the spill would not be tabulated in any way by the I/O model. I am sure that CEPA would not endorse the idea that oil spills are good for their economy, but that’s what their method would tell you – higher costs mean higher benefits.
Input/output models lull us into a false sense that costs are actually benefits, so we’re fine no matter what. They allow business and government leaders to believe that a project executed inefficiently is better for the economy than one executed efficiently. The next time you’re faced with these numbers, take them to their logical conclusion. If they tell you that a pipeline would be worth more to our economy if it cost more to operate or if you built it twice rather than once, ignore the numbers. If the method would count an oil spill as a benefit, it’s not worth the paper it’s printed on.
These models have their uses (see Stephen Tapp‘s comments here and here), but with every study I see using the I/O model to validate lower productivity, I become more and more convinced that setting all macro multipliers to zero by fiat would be a welfare-improving policy. That way, people would have to justify projects on their merits with respect to the welfare of Canadians, and that would be a refreshing change.