From whence innovation?
Posted by Daniel Hall on January 14, 2008
I have been sitting on this post for almost two weeks now, trying to work out a satisfactory (and clever) way to do a back-of-the-envelope calculation that would help me get a rough handle on the degree to which innovation gets spurred by regulation versus emerging from the rigors of the competitive market. Alas, I haven’t come up with anything yet, so I’m choosing to post a more qualitative discussion for now.
This finding is key for three reasons. First, it demonstrates that regulations are an effective way to generate innovation. Second, it shows that tighter regulations can be consistent with continued output growth. And finally, it reveals that tighter regulations in developed countries need not lead to massive leakage of production and pollution to places with loose pollution rules.
Well, that sure sounds nice — regulations are good for you! — but it’s not necessarily what this research implies. The problem with claiming that “regulations are an effective way to generate innovation” is that Levinson isn’t trying to disentangle cause and effect; indeed, he notes early in his post the inherent challenges facing those who set out to conduct causal analyses. Levinson’s research shows that technological improvement (innovation) has accounted for 60% of the decline in U.S. manufacturing emissions from 1972 to 2001. (The rest is due to changes in what we consume and changes in where things are made.) But how or why this technological innovation has occurred is a much different question than assessing its relative contribution to total reductions.
Environmental regulations may be driving technology innovation, but Levinson’s analysis doesn’t shed light on whether or to what extent this is the case. Companies have private incentives to reduce emissions to a degree as well — for example, efficiency improvements that save companies money are a big part of the reason why there is a secular trend towards a more energy-efficient economy. This question — does regulation induce innovation? — is an important and interesting one, but it needs a more direct treatment before we jump to an answer.
Enter reader Tidal, who comments that:
I think it rather important to note that he does not consider CO2 as a manufacturing emission. US CO2 emissions have risen quite dramatically over that time frame, albeit with falling “intensity” vs. GDP. Granted, there have been fewer incentives for US manufacturing firms to reduce CO2 over the same period.
This is where I hoped to develop some type of numerical back-of-the-envelope estimate of the reduction in CO2 emissions and compare it to the pollutants that Levinson is looking at. What is going on with CO2 emissions over the same time period, and can it tell us something about how much unregulated emissions may have improved (or worsened) on their own?
First, a point of clarification: Tidal is correct that U.S. CO2 emissions have risen rapidly, but he is wrong to pin it on industry. Increases have been driven primarily by transportation and the electric power sector. Industrial CO2 emissions, meanwhile, have been nearly flat over the last 15 years. (See Figure 3 in this summary paper, or for further info see the EPA or EIA emissions reports.)
The EIA collects detailed data on CO2 emissions from U.S. manufacturing through the Manufacturing Energy Consumption Survey. Their most recent report examines emissions from 1991-2002. Note Table 7 near the end, which shows that the emissions intensity of manufacturing (CO2 emissions per dollar value of output) fell at an average annual rate of 1.4 percent over that 11-year period — a period in which, quite clearly, CO2 emissions were unregulated.
The problem with moving from this fact to even a rough idea of what produced this change is at least two-fold. The first is practical. Like Levinson if I wanted to understand the impact of technological change alone then I would need to isolate it from compositional effects — the changes in consumption and trade that led to a different mix of products being manufactured in the U.S. (And indeed these could be quite important — although Levinson estimates that trade can explain at most 28% of the total change in manufacturing emissions from 1972 to 2001, his analysis suggests its relative importance increased in the latter part of that period.) This analysis could be done for CO2 in the same way Levinson does it for the regulated pollutants, if, that is, I had access to Levinson’s detailed breakdown of the value of production and shipments in each of the 470 industries in the 6-digit North American Industry Classification System (NAICS) codes, and I also had a breakdown of CO2 emissions for each of these industries. I don’t.
Even if I did, however, the second — and frankly much bigger — problem is that I don’t have a way to control for the likely collinearity between these regulated pollutants and CO2. It is possible that a desire to cut costs and compete more effectively led firms to innovate new ways to reduce overall costs through increasing efficiency, thus leading to reductions in energy-related CO2 emissions. Such actions would also have likely led to some reductions in these other air pollutants (which are often associated with energy use).* Or, looking at it from the other side, it could well be that some firms faced with a regulatory limit found that the most effective way to reduce pollutants such as NOx or SO2 was to improve the energy-efficiency of a manufacturing process, which would then simultaneously reduce CO2 emissions. I don’t have a way to separate out which reductions were prompted by regulation, and which by a market-imposed desire to cut costs.
There’s also a second thought experiment that I’ve been pondering that concerns the relationship between regulation and innovation. Imagine that the U.S. and Europe both enact similar pollution-reduction laws — similar levels of stringency, similar compliance mechanisms and penalties — for the same sets of industries. In theory the effective “spur” to increase innovation (and reduce pollution) would be the same in each region. But I suspect the outcomes would not be the same: we would observe different levels of effective innovation between regions. Is it possible to empirically examine this question through the histories of regulation and innovation in the U.S. and Europe? I don’t know, but I’d be interested if anyone has tried. (This seems the second potential research program I’ve suggested for some whip-smart and energetic economist to take up.)
The final point I’ll make about this innovation-regulation relationship comes from Levinson’s original post. He states that recent empirical work “does seem to find that those industries that have seen the largest increase in pollution abatement costs in the US have also seen the largest increase in net imports” (citing a forthcoming paper). The implication is that there may be limits to how burdensome regulation can become before it starts to result in leakage to regions with less stringent rules.
My objective with this post isn’t to deny that there is a relationship between regulation and innovation, but to highlight some of the complexities involved. Claims that “regulations are an effective way to generate innovation” sound enticing, but if unsubstantiated they frequently serve only to perpetuate myths about the ready availability of free lunches. Cost-benefit analyses may well show that regulations produce net benefits, but we shouldn’t be glossing over costs to arrive at that conclusion.
*Such a scenario could be part of why the actual costs of complying with environmental regulations are frequently lower than ex ante projections suggest.