Special Issue 01
Words by

Anti-growth safetyism

12th October 2022
22 Mins

Is a build up of generic regulations together causing us to be three times poorer than we need to be? Probably not. But the insidious rise of risk aversion is still a big drag on economic growth.

Where is my flying car?’ asks J. Storrs Hall, with justified indignation, in his 2021 book of the same name. If it weren’t for meddling, shortsighted regulators preventing innovation in fields like aviation, energy, and nanotechnology, we already would have flying cars, according to Storrs Hall. Without those regulations that were gradually layered upon the American public over the course of the twentieth century, we would have also grown far more rapidly. Citing a paper published in the prestigious Journal of Economic Growth, Storrs Hall concludes that the median household income in the United States would have been $185,000 in 2011  , three times higher than it actually was. Such a difference in economic performance is hard to even contemplate, although much of Where is My Flying Car does exactly that.

We don’t think regulation, as most economists think of it, is exactly the right framing. What Storrs Hall really describes as preventing innovation isn’t simply the red tape of too many rules; it is what we will call Anti-Growth Safetyism. While Anti-Growth Safetyism sometimes leads to or consists of regulations, it’s also a broader cultural force. It prevents any innovation from occurring in areas like aviation or energy because the public is misinformed about risks or because politicians and bureaucrats prioritize CYA (Cover Your Ass) over anything else they could be doing with their time.

Anti-Growth Safetyism differs from, say, the minimum wage or a rule that all coal power plants must install specific pollution-curtailing devices. Such regulations add incremental, annoying costs to doing business without fundamentally preventing innovation from happening. Anti-Growth Safetyism simply stops innovation in its place by erecting barriers in the name of safety (for people or the environment) that are basically impossible to overcome, such that many of our technologies are stuck in the 1970s. For example, nuclear power is required to keep radiation ‘As Low as Reasonably Achievable’ – an infinite standard that eats up the returns of every new technological gain. The pernicious attitudes toward innovation underlying Anti-Growth Safetyism also underlie phenomena like the bizarre behavior on the part of regulators in the United States with respect to the Covid-19 vaccine, where it wasn’t just the regulations at issue but the operation of the bureaucracy itself, with slow approvals likely killing off a few hundred thousand Americans. Anti-Growth Safetyism is about covering your ass, not safety.

What we want to figure out is what we know about Anti-Growth Safetyism. Is economic research, including the work Storrs Hall cites, able to measure it? Is it impoverishing us to the degree Storrs Hall suggests?

Although regulations often seem irrational, self-defeating, or even pathetic, it was not all that long ago that commentators generally sympathetic to market-driven progress expressed qualms over calls to further deregulate the economy, as if those calls were all a plot by ideologically blind conservative absolutists. As Jonathan Haidt writes in The Righteous Mind:

“I think liberals are right that a major function of government is to stand up for the public interest against corporations and their tendency to distort markets and impose externalities on others, particularly on those least able to stand up for themselves in court (such as the poor, or immigrants, or farm animals). Efficient markets require government regulations . . . [It] is healthy for a nation to have a constant tug-of-war, a constant debate between yin and yang over how and when to limit and regulate corporate behavior.”

And Steven Pinker in Enlightenment Now:

“Right-wing libertarians (in their 21st century Republican Party version) have converted the observation that too much regulation can be harmful (by over-empowering bureaucrats, costing more to society than it delivers in benefits, or protecting incumbents against competition rather than consumers against harm) into the dogma that less regulation is always better than more regulation.”

That is to say, regulations are not simply needless red tape – they often serve legitimate and important social purposes.

We do not know if Haidt or Pinker would have the same view about regulation today. Certainly, however, some of the problems of the regulatory state that draw the ire of Storrs Hall have become more publicly evident in recent years, as the leadership of organizations like the FDA or CDC on issues from sunscreen to vaccines to imported baby formula has been, in our view, indistinguishable from that of a death cult.

Perhaps reasonable people disagree. However, Storrs Hall’s claim, if true, would blow most practical concerns out of the water. (‘An externality here, an externality there, I’m sorry, I can’t hear you over the noise of $185,000 for the median household.’) Storrs Hall correctly points out that the major environmental (and externality) issue of the day, global warming, would simply not exist were we to have the means of generating power (i.e., nuclear) he envisions in the absence of regulation. But even if other issues had arisen in the absence of the modern regulatory apparatus, $185,000 for the median household can paper over an awful lot.

This figure originates with the work of economists John Dawson and John Seater, in a paper titled ‘Federal Regulation and Aggregate Economic Growth’. What they did was to take the theoretical work of another economist, Pietro Peretto, who created a formal model of how taxes affect economic growth, and conceptualize regulations as acting similarly to taxation within his formal model. Dawson and Seater count the number of pages in the Code of Federal Regulation to measure the level of regulation in the United States over 57 years and apply Peretto’s model. By bringing data to this model they show that regulation may be a more important determinant of economic growth than taxation. More speculatively, they use their estimates of how regulation affects the economy to imagine the size of the US economy in the absence of new regulation. But in no way should the exercise be interpreted as akin to a controlled experiment.

Dawson and Seater find that if you imagine that tax rates and the numbers of pages of regulations are the only drivers of changes in the trends of economic growth rates, you can, without accounting for any other factors, predict the trends pretty well. If you take this seriously, it implies that two percentage points have been knocked off growth every year by the sheer amount of regulation. Due to the nature of compounded growth, the median household would have $185,000 if we did have an extra two percentage points for just over half a century.

There is little to object to when this is an exercise carried out in a scholarly journal. We want to know what the effects of regulation might be under different modeling assumptions. But for several reasons, it does not seem fair to assess that as a baseline estimate for what we really could have had were it not for all that regulation. For reasons we will explain, it is very difficult to translate modern empirical work on regulation into a reasonable guess as to what would happen in a counterfactual world where we never made it impossible to innovate in areas like aviation and energy via regulation.

Why? Let us start with what we view as the origins of the modern scholarly literature on regulation, ‘The Regulation of Entry’, a paper published in 2002 in The Quarterly Journal of Economics by Simeon Djankov, Rafael La Porta, Florencio Lopez-de-Silanes, and Andrei Shleifer. They create a new data set for a large number of countries that measures the cost and time involved in setting up a new business. They find that countries with lower levels of regulation tend to also have more democracy and less corruption. Surprisingly, there doesn’t seem to be any relationship between regulation and things like product quality, which runs counter to the story that regulations (generally) fix market imperfections.

This was followed up in 2006 by a short paper in Economics Letters by Simeon Djankov, Caralee McLiesh, and Rita Maria Ramalho, which finds that going from the worst quartile to the best in business regulation corresponds to about two percentage points of growth. (Neither of these papers is anything like a controlled experiment, but we will set aside that concern for now.) This finding ostensibly supports the findings of Dawson and Seater, and Dawson and Seater say as much in their own paper.

But let’s take a step back. First of all, what were Djankov, McLiesh, and Ramalho comparing? The variation of business regulations across the world. If you are comparing the countries with the very worst business climates, where it takes literally years of your life and multiples of the country’s average income in fees and such to open a business legally, to a country like Hong Kong, yes, it is somewhat plausible that you could bump up the rate of growth by two percentage points in the former by simply deregulating.

The first problem with extrapolating from this number to the United States is that the United States is already in the first quartile in the measures Djankov and his colleagues are using.

There’s something of a contradiction here – how can the United States have both the best regulatory climate in the world while also having such a heavy-handed regulatory environment that our median household income is a third of what it could be?

It’s possible that the US is that regulated and other countries are simply worse. The better reading of the facts in our view is, however, that the number of pages in the Code of Federal Regulation and the business regulations measured by Djankov and his colleagues are measuring very different things. And neither of these measures has very much to do with why you don’t have a flying car.

Rather, we prefer to demarcate regulations into one of the three following categories:

  1. Djankov and his colleagues were measuring what we will call ‘regulations against capitalism’. A country has these regulations when it basically does not want markets to function and, instead, for everything to go through the government. These regulations are almost always bad and we would hope that those like Haidt and Pinker would agree. The United States does very well on these, but we suspect that there would be agreement that more deregulation in this category would largely be seen as desirable. According to the Doing Business report, until reforms took place after 2013, it took 690 days to legally set up a business in the country of Suriname and it would cost more than a year of average income in fees. At some point this century, it would cost literally ten times or more than the average yearly income to legally set up a business in Sierra Leone, the Democratic Republic of the Congo, and Angola. In the United States, in the report’s most recent year (2020), it would take four days and one percent of yearly income. There’s room for improvement – in New Zealand, it takes a half of a day and 0.2 percent of yearly average income – but the United States is rather reasonable, and often quite good, in terms of ‘regulations against capitalism’.
  2. Regulations akin to the data from the Code of Federal Regulation built by Dawson and Seater we will call ‘compliance regulations’. Think of endless paperwork, Occupational Safety and Health Administration (OSHA) regulations, financial rules, or rules concerning environmental protection. Maybe many of these are good, and they are probably what scholars like Haidt and Pinker have in mind.
  3. And the last category is Anti-Growth Safetyism. This encompases both a cover-your-ass culture and the manifestation of this culture throughout the state and private firms alike. The culture of Anti-Growth Safetyism means that rather than having to explain why a regulation is worth the cost of less economic dynamism, political entrepreneurs can score points by promising to protect the public. If we could have had a flying car and do not because of regulations, it’s because of Anti-Growth Safetyism, not because the United States has too many regulations against capitalism or ‘compliance regulations’. Evidence about the effects of regulations against capitalism or compliance regulations are uninformative about Storrs Hall’s hypothesis.

We do not think there is much evidence at all for the claim that, in the United States, incomes would triple were it not for regulations against capitalism (the United States is not that bad) or compliance regulations (they are not a big enough deal). But it may be possible that Storrs Hall is correct and US median household income would be $185,000 in the absence of Anti-Growth Safetyism – we just cannot say because there aren’t data sets on Anti-Growth Safetyism.

It is true that France and Japan have (or had) somewhat less onerous of a regulatory structure concerning the construction of nuclear power plants, but not in such a way that would support new world-changing innovations. And French nuclear costs continued to rise after the 1970s, like those in the USA – the French state was simply more willing to absorb those costs. There just isn’t meaningful variation in the data that would allow us to convincingly approach the question.

The Dawson and Seater paper (using compliance regulations) neatly circumvents the lack of variation across countries by looking at regulation within the United States, but, as we stated, we do not think that the greater amount of workplace signage the American health and safety bureaucracy requires firms to hang on their walls, or yet another disclosure form a firm on Wall Street must fill out, is that reasonable of a proxy for how a safety-obsessed society makes it impossible to innovate in areas like aviation.

If Italy cut its thicket of regulations against capitalism to the more modest levels of Sweden, we can be somewhat confident about the positive effects because we know what Sweden looks like. However, a prediction about what the US would have looked like if we had not engaged in Anti-Growth Safetyism is far more speculative, because no relevant comparison exists. Indeed, as Storrs Hall argues, a focus on safety seems to have arisen in all developed countries.

So while Storrs Hall could be right – that eschewing Anti-Growth Safetyism could lead to massive gains in median income – he can’t prove it from the studies he cites, nor any study. But we do have one suggestion for establishing a better-grounded counterfactual for what the United States would look like in the absence of Anti-Growth Satefyism. How would the world look differently if policy differed, and how would it concretely change economic performance? Robert Fogel explored just this when he won the Nobel Prize in Economics in 1993, specifically when he challenged the widespread view that the development of railroads was essential for economic development in the United States. Fogel explored which technologies would have been used in the absence of railroads (probably canals), and how much more it would have cost if the earlier technologies were to have been used instead. You trace out the cost of delivering goods to market using railroads versus canals, and then you can ballpark how important the railroads were. He found that railroads only explain a few percentage points of GDP (not GDP growth). The historians who preceded him were wrong.

But in a more recent work, a 2016 paper in The Quarterly Journal of Economics titled ‘Railroads and American Economic Growth: A “Market Access” Approach’, by Dave Donaldson and Richard Hornbeck, the authors take a similar approach to Fogel, but get a different result. Instead of just looking at the next-best technologies, which assumes all the benefits of railways are internalized into the transport-spending decision, they look at the land market. Railroads lowered the cost of bringing agricultural products to market so much that it became economical to use vastly larger plots of land, substantially increasing productivity and therefore agricultural land values across huge swathes of America. This more comprehensive estimate comes out much larger than the impact as judged by Fogel, so maybe the historians who preceded him were not totally wrong (though they lacked the knockdown evidence for their case).

What we suggest is that to estimate counterfactual US median household income, grounding the analysis in a method like Fogel or Donaldson and Hornbeck, future inquisitors of the lack of flying cars can get something more tangible. Storrs Hall speculates that energy would have become too cheap to even meter – we probably should not start our counterfactual with that assumption. We can instead speculate that in the absence of Anti-Growth Satefyism, nuclear power may have simply improved in its technical efficiency at the same rate as other heavy industries over the same period of time, and in turn muscled out expensive forms of power generation across the country, plus most fossil fuels not used in combustion engines (and presumably all coal, with measurable health and environmental benefits associated with its elimination). In the spirit of Donaldson and Hornbeck, we could consider where else the effects of cheaper, clean power may arise, perhaps by considering industries where innovation appears especially responsive to cheap energy.

This description is consciously less optimistic of what the counterfactual would look like than Storrs Hall, but we are stating the potential analysis in terms of something that could be done, and would be believable, whatever you think of other claims made in Where Is My Flying Car? We suspect that nanotechnology, functioning as Storrs Hall describes, would begin putting us in the realm of ‘$185,000 median household income’ in a similar counterfactual, if it were taken on face value. But that is an exercise waiting to be performed.

Another way of trying to get an answer on the cost of Anti-Growth Safetyism from the academic literature that currently exists is to take all the studies on the effects of regulation, and violently shake them until something that vaguely resembles the research question we want to ask pops out. Maybe here we can find some positive support for Storrs Hall’s hypothesis. Fortunately, just this year, the academic journal Regulation & Governance published a review article by James Broughel and Robert W. Hahn, titled ‘The Impact of Economic Regulation on Growth: Surveys and Synthesis’. We can use that article for our purposes, and a working version of the paper is free online.

For studies to begin corroborating the finding of $185,000 average household income in the United States, they would need to check several boxes. First, just looking at GDP growth is not really sufficient. GDP growth occurs for essentially three reasons: More labor is employed, more investment in capital takes place, and we make our existing labor and capital more productive. The last of these is called total factor productivity (TFP), and it’s what we really care about (to their credit, Dawson and Seater do focus on TFP). Some policies may allow us to use more labor (think of a policy lowering the unemployment rate), but that can’t have a permanent effect on the growth rate of GDP – there’s only so many people you can employ. But something that was not obvious to economists until the work of Robert Solow (another Nobel laureate) is that the same kind of thing happens with investment, and you cannot look to investment to permanently improve your growth rate, either. It’s just TFP that can have permanent effects on growth rate.

Another subtle point: We must have at least some reason to believe that the improvement in growth caused by deregulation through TFP is itself something permanent, not transitory. Let’s say a country banned pressure cookers, but lifts that regulation. We would expect a boom in the pressure cooker industry, but it would have a minimal effect on the permanent rate of growth, because it doesn’t impact how well we are able to continue making improvements in how we apply our capital and labor. This distinction is made in the figure below. Line A represents a permanent increase in the rate of growth. Line B does not. Dawson and Seater’s estimates suggest that the effects of deregulation look like A, not B (though we are actually unconvinced that this is fully addressed in their paper).

A model of deregulation’s effect on the growth rate.
Joe Ronan

Another related issue is that the study that could answer our question must somehow zero in on the effects of regulation for the United States and those countries similar to the United States. The paper by Djankov and his colleagues, for example, appears to be measuring what happens when an overregulated country becomes like the United States, not the effects of regulation in the United States. Even if you are looking at TFP, it must not be gains to TFP that the United States has already achieved. It needs to be something like the production of a new technology.

Okay, so given all those conditions, what shakes free from the literature? A paper published in the Review of Economics and Statistics in 2013 by Renard Bourlès, Gilbert Cette, Jimmy Lopez, Jacques Mairesse, and Giuseppe Nicoletti, ‘Do Product Market Regulations in Upstream Sectors Curb Productivity Growth? Panel Data Evidence for OECD Countries’, is by far, the most on point. The paper focuses on industries in 15 OECD countries, meaning countries that are similar to the United States or close to the United States. The measure used is from a product market regulation database available for a select number of countries. The data is a bit all over the place, and it can probably be thought of as a mix of all three kinds of regulations discussed above. But it does consider the effect of regulation on total factor productivity, the type of technological progress that can give a lasting boost to economic growth.

What makes the paper most intriguing is that it shows that not only is the effect of product market regulation on total factor productivity negative among just OECD countries, but the effect is larger the closer a country is to the productivity frontier. This suggests product market regulation is preventing true innovation on the frontiers of economic growth, not just causing temporary jolts to growth rates among the OECD countries with weaker economies. In other words, these regulations seem to have a greater effect on the speed at which we discover things like cutting-edge battery technology, rather than simply how quickly managers in Greece can adopt the competent management practices of German firms. Finally, while it still is not a controlled experiment, it is considerably closer to one than the estimation by Dawson and Seater or by Djankov and his colleagues because it looks at natural experiments: instances of deregulation and the situations before and after them.

Based on the regulations they do track, they estimate regulation as a whole is cutting US GDP growth one percentage point per year. Although the model seems to indicate that removing regulations would improve innovation, the paper does not indicate (and really cannot indicate) if the effect on growth would be permanent. But even if we pessimistically assume that the effect on growth rates ‘only’ lasts a few decades, not more than a half century, the cumulative effect is large. Since this measure still isn’t really data on Anti-Growth Safetyism, it cannot disprove the $185,000 claim. It is always possible that the extra untracked issues mean we would indeed have flying cars and $185,000 median household income in the presence of a regulatory environment unafflicted by Anti-Growth Safetyism.

If we did take this one-percentage-point figure, though, and naively apply it as if it perfectly fit the original Dawson and Seater methodology to try to work out what incomes today would be in the absence of the increases in postwar regulation (and Anti-Growth Safetyism), then we still get a pretty big number. Median US household income is about $70,000 today. If growth had been one percentage point higher since the late 1940s, it would be $117,000 today. That’s still a remarkable figure.

What would we have to do in order to properly measure Anti-Growth Safetyism? It might be a question of creating new data sets. In recent years, economists and other social scientists have tackled a range of questions by mining newspaper articles, social media posts, books, and journals for various keywords. Perhaps words with safety valence could be identified and their frequency measured, using variation over time, and, say between states or countries to identify variation. Perhaps there are big cases that had trend-breaking impacts on safety consciousness, and we could look at growth before and after with event studies. Perhaps we could, with theory, get a more granular understanding of how safety concerns generally have their impacts. In all of these areas, the main contributions are still to be made. Yet without good measures of Anti-Growth Safetyism it is hard to be sure, with Storrs Hall, that it is the reason we don’t have flying cars.

So we might have to settle for merely nearly doubling our incomes, with Bourlès et al. Such an increase in income would represent an incredible result for a single set of policy reforms. It is usually not possible for reforms to produce results this large – with the exceptions of migration policy and housing policy, none this large comes to mind. Big, bold policy reforms in regulatory policy, especially with respect to strangulating the misanthropic ideology of Anti-Growth Safetyism, have the potential to put an end to the economic malaise and stagnation faced by societies in recent decades – a conclusion we can reach even after subjecting the hypothesis of Storrs Hall to rather close scrutiny.

More articles from this issue

Nanotechnology’s spring

Words by Eli Dourado

Nanotechnology sometimes sounds as much like science fiction as artificial intelligence once did. But the problems holding it back seem solvable, and some of the answers may lie inside our own bodies.

Read more