In a lengthy article for National Affairs, I wrote the following sentences:
From 1980 through today, America’s share of global output has been constant at about 21%. Europe’s share, meanwhile, has been collapsing in the face of global competition — going from a little less than 40% of global production in the 1970s to about 25% today. Opting for social democracy instead of innovative capitalism, Europe has ceded this share to China (predominantly), India, and the rest of the developing world. The economic rise of the Asian heartland is the central geopolitical fact of our era, and it is safe to assume that economic and strategic competition will only increase further over the next several decades.
These have been the subject of extensive criticism in the left-of-center blogosphere. Let me take the criticisms, as I see them, one at a time.
Criticism 1: I presented incorrect numbers. Accurate data is the foundation of all empirical analysis. Without it, progress is not realistically possible. Paul Krugman wrote a blog post in which he accused me of either gross negligence or dishonesty. The specific charge he made is:
But I went back to Manzi’s source of data, and it turns out that it’s even worse than that. If you use the broad definition of Europe, which includes the USSR, it did indeed have 40 percent of world output in the early 1970s. But that share has not fallen to 25 percent — it’s still above 30 percent.
His assertion is flatly false.
First, Professor Krugman incorrectly identified my source of data.
I have never corresponded with Professor Krugman concerning data sources and analysis for the passage in question, so I can not know on what basis he asserted that the dataset to which he links is my “source of data.” It is not. As per the blog post in which I reviewed multiple data sources for the analysis in question, I averaged multiple sources of data. Professor Krugman has selected only one of these sources, presented it as if it were my sole source, and therefore (obviously) failed to replicate the published result for both the 1970s and the current day. The error is his.
Second, Professor Krugman misread the economic dataset that he did identify.
According to the dataset in question, that part of Europe excluding any part of the USSR had about 40% share of global GDP in 1973 [Cell H59 in the spreadsheet]. According to this dataset, if you add the USSR to this definition [Cell H59 + Cell H102], then such a constructed entity would have had global GDP share of about 44%, not the 40% that Professor Krugman asserted. This error is his as well.
I have invited Professor Krugman to explain how I am wrong, or failing that, to retract his charge.
[UPDATE]: Professor Krugman has responded. I will reproduce his post in its entirety here:
The debate over European economic performance has gone in two directions. One is the usual response of people when they get something very wrong: they start quibbling over minor details (which dataset was Manzi actually using? what about immigration?) to throw up dust clouds. The key thing to remember is that we had a flat assertion that social democracy leads to stagnant economies; that assertion is just wrong.
The other is to point out that even rich European countries have lower GDP per capita than the United States. Indeed. But there’s a story there that is a lot more complicated than a simple table conveys. I wrote about it a few years back.
I’ll leave it to readers to evaluate this response.
Criticism 2: Defining Europe to include Russia and other parts of Eastern Europe is ludicrous when the argument concerns the trade-offs involved in the social welfare state.
Critics argue that since we all know the social welfare state is only present in Western Europe, then it is disingenuous to bundle together Western and Eastern Europe.
In the article, I explicitly defined the ‘European model’ as I used the term by methodically listing out each element of it:
Seen together, these initiatives — shifting government spending away from defense and public safety toward social programs; deeper direct involvement of the government in the operation of large corporations across a substantial portion of the economy; energy rationing in the name of managing climate change; more direct government control of health-care provision; and higher tax rates that probably include a VAT — point in a clear direction. The end result would be an America much closer to the European model of a social-welfare state, which prioritizes cohesion over innovation.
It turns out that Europe as whole is systematically different than the U.S. on each of the listed dimensions. As a further, and more severe statement, it is also true that Eastern Europe as a region (defined as Russia and other west of Urals components of the old Soviet Union and countries of the old Soviet Bloc) and Western Europe as a region (defined as all other countries in Europe from Iceland to Greece) are also each systematically different than the U.S. on each of the listed dimensions.
I’ll take each one at a time
shifting government spending away from defense and public safety toward social programs
The governments of both Eastern and Western Europe have a much higher weighting of social spending than does the government of the U.S. According to the OECD common governmental expenditure classification system, in 2006 the ratio of government spending on the sum of Housing and Community Amenities, Health, Recreation, Culture and Religion, Education, and Social Protection to government spending on the sum of Defense plus Public Order and Safety was at least about twice as high as in the U.S. in every country for which the OECD reports in Eastern Europe (Czech Republic, Estonia, Hungary, Poland, Russia [using transaction code P3CG], Slovak Republic and Slovenia), and every country for which the OECD reports in Western Europe (Austria, Belgium, Denmark, Finland, France, Germany, Greece, Iceland, Ireland, Italy, Luxembourg, Netherlands, Norway, Portugal, Spain, Sweden and the UK).
deeper direct involvement of the government in the operation of large corporations across a substantial portion of the economy
In finance, according to the NBER working paper Government Ownership of Banks (2000), the share of government ownership of the assets of the 10 largest banks was as high or higher in every reported Western European country and in every reported Eastern European country than in the U.S. (where it was 0%).
In products, according to the OECD Indicators of Product Market Regulation, in 2008 product market regulation was heavier than in the U.S. in every reported Western European country (although the UK was almost tied with the U.S.) and in every reported Eastern European county.
Health care is addressed below.
energy rationing in the name of managing climate change
The following Western European countries are ratified Annex I signatories to the Kyoto Protocol that commits nations to defined emissions caps: Austria, Belgium, Denmark, Finland, France, Germany, Greece, Iceland, Ireland, Italy, Liechtenstein, Luxembourg, Monaco, Netherlands, Norway, Portugal, Spain, Sweden, Switzerland and the United Kingdom.
The following Eastern European countries are ratified Annex I signatories: Belarus, Bulgaria, Croatia, Czech Republic, Estonia, Hungary, Latvia, Lithuania, Poland, Romania, Russian Federation, Slovakia, Slovenia, and Ukraine.
The United States has not ratified participation in the Kyoto Protocol.
more direct government control of health-care provision
According to the OECD Health Data 2009, in 2007 the U.S. had lower public expenditure on health as a percent of total expenditure on health than every reported country in Western Europe and every reported country in Eastern Europe.
higher tax rates
According to the OECD Factbook 2009, in 2007 the U.S. reported lower total tax revenue as a percentage of GDP than every reported country in Western Europe and every reported country in Eastern Europe.
that probably include a VAT
The following Western European countries have a VAT: Austria, Belgium, Denmark, Finland, France Germany, Greece, Ireland, Italy, Luxembourg, Netherlands, Portugal, Spain, Sweden, UK, Iceland, Norway, and Switzerland.
The following Eastern European countries have a VAT: Bulgaria, Czech Republic, Estonia, Hungary, Latvia, Lithuania, Poland, Romania, Slovakia, Slovenia, Albania, Belarus, Bosnia, Croatia, Moldova, Montenegro, Russia, Serbia, and Ukraine.
The U.S. does not have a VAT.
There is obviously tremendous variation across the nations of Europe, but at the level of abstraction carefully defined in the article, there are systematic differences between Europe and the U.S. in their respective approaches to the political economy of the social welfare state. In the article, I also carefully reviewed the evidence that actions to date in the late Bush and early Obama administrations, plus the publicly stated intentions of the current administration and congressional leadership, have made and/or would make the U.S. more like a European social welfare state in each of the defined dimensions.
Criticism 3: This comparison does not demonstrate that the social welfare state causes lower economic growth.
As I said almost immediately when this was pointed out, this is exactly correct. No matter what the relative growth rates (however defined) of Europe and the U.S. over any period, such a comparison could never constitute a reliable empirical demonstration of causality, because we could never know the counterfactual of what would have happened had one entity or the other executed some alternative policies. I have argued in many posts over a period of years that even less-naïve econometric methods that attempt to control for other factors are not sufficient for the task of identifying non-obvious causal relationships in human society reliably.
The purpose of the article as a whole, never mind the short passage in question, was not to provide an empirical demonstration that less regulated markets tend to provide faster economic growth under many conditions than more regulated markets. Nor did it ever claim to provide this. The purpose of the article was to describe why even though I (like many, many other people) accept the advantages that less regulated markets can provide, this does not lead to the conclusion that we should advocate a deregulation-oriented path of economic development without considering the balancing consideration of social cohesion, and modifying our proposals accordingly.
As I also said immediately, it is easy for me to see how these sentences could be read naturally to be an asserted demonstration of causality. And as I also said at that time “the responsibility for lack of clarity rests on the shoulders of the author, not the reader.” I acknowledged this, and thanked my interlocutors for improving my expression; I continue to do both.
Criticism 4: The fact that the difference in total GDP performance between Europe and the U.S. is entirely due to different population growth rates proves that it can not have been caused by the social welfare state.
In effect, this argument is that GDP = GDP / Person X Population, and that GDP / Person has grown at about the same rate in the U.S. and Europe over the past several decades, so the only real difference has been population growth. This is a factually (approximately) correct statement. There are two big problems with arguing that this shows that therefore differences in political economy of the welfare state did not matter to this difference in outcome. Both are counterfactual problems. First, just because growth in per capita GDP turned out to be the same in this period, it does not follow that had the U.S. adopted policies more like those in Europe that growth rates would not have have been lower, and per capita GDP converged. It is almost always harder to maintain a lead. Europe and the U.S. had previously been on a generally converging path of GDP / Person, and maintaining this lead in productivity (by this metric) was very far from preordained. Second, it assumes that the political economy of the welfare state can not affect population growth, either through impacting fertility or impacting immigration. There are many plausible models for how this can happen. We can’t know through econometric analysis (whether of the simple “I just know that the welfare state can’t influence fertility or immigration” variety cited in the title of this section, or of the more sophisticated variety that attempt to control for other factors) what U.S. population growth and per capita GDP growth would have been had the U.S. adopted different social welfare policies.
Many variants of these criticisms have been repeated over and again through a certain segment of the blogosphere. I have here answered each of them directly, and I believe comprehensively. In any event, barring new information brought to the debate, this is where I intend to leave it.