Guest Post: Economic Freedom Indexes Do Seem to Measure How Capitalist a Country Is: A Response to Ben Burgis
This is a guest post from Tibor Rutar, an Assistant Professor of Sociology at University of Maribor, Slovenia.
In my book Capitalism for Realists, I argued that economic freedom indexes, such as the third-party sourced index produced by the Fraser Institute, are a good – if imperfect – measure of how capitalist (or perhaps even “neoliberal”) a country is. This is an important matter, because if true, then we would be able to measure whether the “more capitalist” countries score better or worse on various standard measures of social wellbeing, such as the prevalence of poverty and inequality, the extent and quality of democracy, and so on. This would help us settle certain aspects of the great debate around the correlates, effects, and overall desirability of capitalist institutions.
Of course, there are those who disagree. One critical thesis I aimed to examine and respond to in the book was the one produced by Ben Burgis in 2021 for the popular Marxist magazine Jacobin. His main argument, which he has now restated in a YouTube video, was that the many positive correlations that exist between higher rates of economic freedom and “good” social outcomes (such as low poverty rates, the presence of democracy, higher GDP per capita; see the figure below for an example) to which libertarians such as Jason Brennan and Peter Leeson refer to when defending capitalism, are in fact uninformative or spurious. Burgis does not challenge the existence of the correlations themselves. But he does deny that they tell us anything about how good or bad capitalism is. His thought is that economic freedom indexes are not really measuring capitalism, so if the indexes show positive correlations with good social outcomes, one cannot say that capitalism correlates positively with good social outcomes.
Why not? Burgis argues that of the five categories composing the Fraser Institute’s index of economic freedom, only three are truly relevant to measuring capitalism. The three categories are (1) size of government, (2) freedom of international trade, and (3) regulation. The two irrelevant categories, according to Burgis, are (4) sound money and (5) legal system and property rights. He adds that one subcategory composing the legal system and property rights dimension might also be relevant to capitalism, namely security of property rights, although he also has some issues with that.
In my book, I wanted to recalculate each country’s index of economic freedom in accordance with Burgis’ critique. My idea was that if the recalculated index (one eschewing the ostensibly non-capitalist measures of the index) still turned out to correlate just as strongly with good social outcomes as the original index does, then Burgis’ critique fails. It fails because it would be demonstrated that the original index of economic freedom is in fact quite a robust measure of capitalism, meaning that a country’s score does not change much if we take out some components of the index (say, the level of corruption or the volatility of inflation). I showed in the book that this is, indeed, the case. It does not matter whether we measure capitalism with the original or the recalculated index, higher levels of economic freedom (measured in whichever of the two ways) still mostly correlate positively with good social outcomes.
Note, of course, that this does not mean the correlation is causal. If anything, I think this would potentially be the greatest challenge to anyone who would mindlessly tout economic freedom indexes as knock-down evidence that capitalism causes nice things. I go to great length in the book to point this out. We cannot say on the basis of simple correlations that capitalism (using the recalculated index) causes low levels of poverty, high levels of democracy, and so on. One has to control for confounding factors and the potential issue of reverse causality to establish this stronger claim, and many studies do. But what we can say even without those studies (which, by the way, come to quite mixed findings on issues such as inequality, corruption, and the environment!) and causal designs is that Burgis’ particular critique of the index as a measure of capitalism does not stand. When excluding what Burgis says are non-capitalist categories from the index, we still arrive at roughly the same simple empirical finding pointed to by Brennan and Leeson. In other words, we can say that capitalism (however measured) is correlated with, or consistent with, reductions in poverty, increase in democracy, and so on. This is important, especially because some anti-capitalists think that bad social outcomes are just rising and multiplying under capitalism. But given the mentioned correlations, this is simply not the case (whatever the exact causality at play).
In a recent YouTube video, Burgis aims to respond to my critique. He says that the demonstration in my book fails because I did not recalculate the index on the basis of his actual suggestion, which had ostensibly been that one should eschew both the sound money and the legal system and property rights categories completely. Instead, I included in my recalculation one of the subcategories of the latter (namely security of property rights), which Burgis says is problematic because we do not know which kinds of rights are being protected. If the subcategory really measures against risk of nationalization, then that is fair game, but if it does not, then we ostensibly cannot use it.
Perhaps I really was not careful enough in the book to home in on that point, instead just taking Burgis at his word when he wrote that the protection of property rights subcategory “is the only one that sounds remotely relevant to the capitalism-versus-socialism debate”. But, fair enough, to really test his thesis, we should be stringent and remove the subcategory that is only “remotely relevant”, and instead look at the others with which Burgis did not have explicit qualms. So, let me first describe the remaining three categories eligible for inclusion, and then we can look at whether, as Burgis intimates, things fall apart with the properly recalculated, “capitalism-only” index of economic freedom.
The first category eligible for inclusion on Burgis’ account is the size of government. This category measures how large government consumption of a given country is, how generous its transfers and subsidies are, how much government investment there is, how high top marginal tax rates are, and what the extent of state-owned assets is. It awards individual scores on all of these dimensions (on a scale from 0 to 10, 10 being most economically free) and then produces an aggregate score to represent all of the individual components of the size of government. The more capitalist a country is in this category (namely, the smaller its government consumption is, the lower the taxes are, and so on), the higher its score of economic freedom. To give a couple of examples, France is scored quite low on this category (5.08) in 2019, the year I am looking at, so one could say that on this category, France is not really all that capitalist. The same goes for Finland or Norway (5.13 and 5.06, respectively), which are near the bottom of the world in these terms (close to Venezuela, which scores 4.34 on the size of government dimension). In contrast, as one would expect, the United States score significantly better (7.32), which means they are much more capitalist in this aspect. So, too, does Yemen with a score of 8.27 due to its lack of government consumption and its extremely low top marginal taxes.
The second category is freedom of international trade. This category measures revenue from trade taxes, the tariff rate, non-tariff trade barriers, compliance costs of importing and exporting, regulatory trade barriers, black market exchange rates, financial openness, capital controls, and freedom of foreigners to visit. Venezuela again scores poorly (4.09), and so cannot be said to be very capitalist on this dimension. In fact, it is almost in last place globally. United States is awarded a much higher score (7.83), although Sweden and Norway are awarded an even higher score (8.49 and 7.91), and so can be said to be much more capitalist when it comes to international trade.
The last category is regulation. This category measures bank ownership, private sector credit, interest controls, hiring regulations and minimum wage, firing regulations, centralized collective bargaining, hours regulations, mandated costs of worker dismissal, conscription, administrative requirements, regulatory burden, starting a business, impartial public administration, licensing restrictions, and tax compliance. Not surprisingly, United States score very high (8.68). Somewhat surprisingly, Sweden also gets a high score (7.81), as does Norway (7.93). Looking into the data for the individual subcategories, we can see that Sweden gets low scores for hiring and firing regulations (4.31) and centralized collective bargaining (4.30), which one would expect. These two aspects of Sweden are not at all capitalistic. However, bank ownership, private sector credit, interest rate controls, starting a business, and licensing restrictions garner it very high scores, (8, 9, or higher). This is even more so the case with Norway, which also gets a high score for regulatory burden (7.78, somewhat better than the US).
What happens when one recalculates the index so that it reflects only the aggregate scores for the size of government, freedom of international trade, and regulation categories of economic freedom (for the year 2019)? Is this new, ostensibly more appropriate measure of capitalism inconsistent with the original, ostensibly inappropriate measure of capitalism? I do not think it is. In fact, the correlation between the two indexes amounts to a very high r = 0.91 (or an R2 = 0.83). In other words, the countries that are ranked to be the most capitalist according to the original index of economic freedom are mostly the countries that still turn out to be ranked as most capitalist according to the recalculated index of economic freedom (namely, one totally eschewing the sound money and legal system and property rights category). In yet other words, are countries that are placed high on the original list of economic freedom solely (or primarily) there due to an accounting trick? Do their high scores merely (or primarily) reflect the non-capitalist fact that they have a robust legal system, low corruption, and low inflation? No, they are not and they do not.
To an extent, this extremely high correlation is obviously the product of autocorrelation, namely the fact that both indexes have, by definition, at least 3 components that are completely identical between them. So, what would be a more demanding test? First, we can check if the recalculated index (based on the three categories) correlates positively or negatively with the two excluded categories. It turns out that the connection is positive, with the R2 amounting to 0.5, which is quite a strong correlation. In other words, it is not the case that, say, a country’s legal system score pulls in the opposite direction of the same country’s size of government score, as Burgis’ charge of spuriousness would imply.
Second and more definitively, we can simply look at whether the recalculated index correlates with various “good” social outcomes in a similar manner to that of the original index. For demonstrative purposes, I picked the share of population in absolute poverty (the $3.65 threshold). Relating the original index and (decreasing population share in) absolute poverty results in a R2 = 0.39, while replacing the original index with the recalculated one gives a somewhat lower R2 = 0.28.[1] The connection is not as strong, so Burgis is not completely wrong in his skepticism, but his point that the correlation between a more appropriate measure of capitalism and “good” social outcomes would disappear definitely does not hold.
To end, let me make one additional note which is actually very important to me. I am not a right-winger, I am not a libertarian, and I am not a neoliberal. I am social democrat that would want the world to look like Sweden or Norway. One of the best books I read in a while is sociologist Lane Kenworthy’s Social Democratic Capitalism, which by the way also utilizes economic freedom indexes (particularly their regulation aspect). I would want the world to score low on certain dimensions of economic freedom/capitalism, such as firing and hiring regulations and centralized collective bargaining. I am happy that on this aspect, Sweden and Norway are not more capitalist. Likewise, I would want the world to score relatively low on the size of government (although not too low!). In other words, I would like to see quite high marginal tax rates and more transfers. Sweden gets a lowly 1 (on the 0–10 economic freedom scale) for its top marginal taxation and government consumption, because both are so high. And it gets a middling 4.95 for its transfers and subsidies, because Sweden’s welfare state is quite generous. I like that! To my mind, being non-capitalist (or only semi-capitalist) on certain dimensions is great. This is something about which I was quite clear in the book.
I am not the neoliberal shill Burgis probably thinks I am. I hope this helps him see that my response to his critique of economic freedom indexes does not come from a place of tribal ideological dismissal. Moreover, I am saying all of this more generally because I have no particular dog in this fight. If Burgis continues engaging with me and shows me a more definitive argument for the spuriousness of economic freedom indexes as they relate to capitalism, I will be happy to concede that we should stop relying on them.
You can listen to podcasts featuring Tibor here and here. Tibor is the author of two books:
Rational Choice and Democratic Government: A Sociological Approach
Capitalism for Realists: Virtues and Vices of the Modern Economy
[1] Almost the same happens when looking at extreme poverty ($1.9 threshold), although the R2 values are lower (0.19 and 0.14, respectively).
I'd love to see why the studies that control for more factors are mixed.