Dynamic America, Poor Europe

Paul Krugman recently wrote in a column for the New York times that Europe has an economy as dynamic as the US, and that the US should be "Learning From Europe". His main argument for this was that from 1980 per capita income only increased slightly faster in America than EU.15 (in fact the US grew 72% and EU.15 grew 66%).

Other economist, such as Tyler Cowen and Greg Mankiw pointed out that Krugman ignores the much higher levels of per capita income in the US. This is important for two reasons. First, it is known in growth theory that in well functioning economies poorer regions/countries grow faster than richer ones, since they can adopt "low hanging fruits". It is remarkable that the US grew faster than Western Europe despite starting off at a higher level. Second, and most importantly, when measuring the productive powers of an economy it is the level that matters, not growth. Albania may have grown faster than Singapore one year, but no one would argue that Singapore is less productive than Albania.

From a policy perspective, if America follows Paul Krugman and Barack Obama in "Europeanizing" the American economy, what is most likely to happen is that levels of income will drop to Europeans ones (or lower, taking into account demographic differences). Once this happens, it is hardly a consolation that America will grow at the same pace as Europe.

During the current recession, the US GDP fell by a 3.3%. Theoretically if the US adopts European policies and immediately decreases to the levels of EU15, its per capita GDP would fall by 26.5%, 8 times worse than The Great Recession! (in practice the convergence would probably happen through years of reduced growth).

Here I will make three comparisons, two that have been done before (but that Krugman and his readers needs to be reminded of), and one that I believe is original.

1. First, let us compare the latest publicly available per capita GDP of 18 western Europeans countries and the US. We see that the US per capita GDP is $45.500, compared to $33.500 for EU15. Each American produces 36% more than each member of the EU15.

2. Second, let us compare the production of American States with European countries.



If France were to became an American state, it would be the 50th poorest, below Arkansas.

The EU.15 as a whole, which Krugman presents to his readers an economy as dynamic as the US, would be the 49th poorest state, below Alabama, a State that Paul Krugman ridiculed in 2005. Few Americans consider Alabama a dynamic state, because of the low average income (even though, hardly surprisingly, the poorer southern states have much faster per capita growth rates than the rich states such as New York). Why than should we consider Europe a dynamic region?

Even the richer European countries do not fare well against American states (the exceptions being oil rich Norway, financial city state of Luxembourg, free market Ireland and capitalist utopia Switzerland). Denmark and Sweden barely inch ahead of Kentucky, below Louisiana, New Mexico and Missouri. Minnesota is 34.4% richer than Sweden.


3. These comparisons are still to generous to Europe. The United States is demographically more diverse than Europe, and for example includes a larger proportion of low skilled immigrants. We also know that some groups have cultural traits that make them more productive than others, such as trustworthiness and work ethics among Scandinavians. I believe that the most fair comparison of American policy environment and formal institutions and European ones would compare apples to apples, and take cultural differences into account.

Of course America did not did not drop out of the sky, it is largely populated by Europeans. All western European nations sent a considerable share of their population to America. Therefore I will compare each European group in America with each host country (some host countries have in turn taken in large groups of non-European immigrants. I will deal with this issue later, although it does increase the advantage of the US by a few percent). The implicit theory behind this comparison is that formal institutions (such as tax rates) and informal institutions (such as work ethic) determine income, that immigrants bring some of their informal institutions with them, and that we therefore should control for what economists refer to as "fixed effects" for each national group.

I discuss method and some problems below. All countries are included, except Spain, which Census does not include (due to the problem with separating European Spanish with people from Latin America who speak Spanish). I believe generally the figures to be most reliable for groups with large immigrant shares, but less reliable for the British.

The GDP per capita for Americans from EU.15 is $53,000, compared to $33,500 for E.U15 itself. Those of European descent in America on average produce 58.6% more than they do in Europe.




In absolute terms, the $19.600 per capita wealth gap between Americans of European descent and Europe is as large as the gap between the Europe (the EU.15) and Turkey. In percentage terms the gap is almost large as the one between western Europe and Hungary.

Being the habitual cheater he is, Krugman asks his readers to compare America to Paris, London and Frankfurt.

"For those Americans who have visited Paris: did it look poor and backward? What about Frankfurt or London? You should always bear in mind that when the question is which to believe — official economic statistics or your own lying eyes — the eyes have it."

If you stop and think about this for a second the problem becomes apparent. These are not representative cities, they are three of the absolutely richest areas of Europe!

According to Eurostat, which contains GDP per capita figures for European regions, each inhabitant of London produces 65.3% more than the UK average. The figure for inner city London, the area most American visiting would see, is an 279%. That is not a typo. Inner city London is the richest region in Europe. Paris has a per capita income 272% higher than the French average. Lastly Frankfurt, the financial hearth of continental Europe, has a per capita GDP 278% higher than Germany as a whole. (Stockholm earns 37% above the Swedish average, for those curious).

These figures overestimate the wealth disparity somewhat, since cities contain more jobs than people, especially important for financial centers (one reason why Luxembourg is so rich). But statistics are not fooling us, central Paris is rich, which our eyes would confirm. The suburbs and most of the rest of France is not. The correct comparison for Frankfurt would be Manhattan, not the US average.

A meaningful question is not if 3 of the richest cities of Europe (randomly chosen by the dear professor for the benefit of his readers, no doubt) are backward, it is if typical Europeans earn less than typical Americans. Which they do, which the data confirms, and which an in-depth journey in Europe or a comparison of starting wages of various professions would verify. The lesson? Trust your eyes, just never trust Paul Krugman.

Suggested by Tyler Cowen, I also constructed a "virtual Europe". This is what would happen if each European country had a per capita GDP level of their descendants in America (so instead of multiplying the Italian-American income by 17.8 million Italian-American we multiply it by 58.9 million Italians). The per capita income of this "virtual Europe" would be $53.600. Soon Europe will attempt a new version of the (largely failed) Lisbon Strategy. They should keep this figure in mind in setting their goals.

The total GDP of EU.15 is 13 trillion dollars. Assuming Spain performed at an average level, an EU.15 working at American wealth producing figures would have a total GDP of 21 trillion dollars! If we believe all the differences in outcome between Americans and Europeans is due to culture and policy (Krugman has no problem assuming all differences in health outcomes is policy alone), the loss in income for Europe from following Krugman's advice instead of current American policy is 8 trillion dollars, every year.

Milton Friedman famously said that there is no poverty among Swedish-Americans. Indeed their poverty rate is only 6.7%, half the national average. It would be interesting to compare absolute poverty between Swedes in Sweden and their American cousins.

For one country, Sweden, I have calculated the figures when excluding immigrants to Sweden. This reduced the American advantage from 55.4% to 50.5%. If we believe that this is representative of Europe (Sweden has a higher share of foreign born than most other European nations) the American advantage should be around 53%.

Krugman's dogmatic ideology aside, there is no doubt which economy that is most "dynamic".


/Tino Sanandaji, PhD student Harris School of Public Policy



Method

I have calculated per capita GDP for the US and 18 western European countries. I have also calculated Gross State product for 50 U.S states and Washington D.C. In addition, as far as I know for the first time I have calculated imputed per capita GDP of Americans by European ancestry, and compare them with those in the home country. My sources for population and Per capita (purchasing power adjusted) GDP are OECD Factbook 2009 (latest available GDP is for 2007, latest available population 2008).

My source for relative income for Americans of various ancestry is Census 2006-2008 American Community Survey 3-Year Estimates. I use per capita income of each group relative to per capita income of the US (so if a group is 10% above average it will have a per capita GDP = 1.1*45489). (click on Selected Population Profiles).

My sources for Gross State Products is Bureau of Economic Analysis, Per capita real GDP by state (chained 2000 dollars) in 2007. I show the results adjusted to the OECD GDP per capita figure (so again if a state is 10% above average it will have a per capita GDP = 1.1*45489).

UK includes British, Welsh, Scottish, Scoth-Irish and those that simply identify as "American".

British excluding "American" has population 41,181,917, per capita income $59775.5 and poverty rate 6.8%.

French includes French-Canadian and Basque. German includes Pennsylvania German.

A small group identifies as "Scandinavian", and a the group that call themselves "European" are included in EU.15 (of course technically some of the Scandinavians are Norwegian and Icelandic).


Some issue that should be taken into account:


1. The people who left may have been the genetic elite of the country, making the comparison unfair.

While this may be true for some countries, the opposite is true for most, it was the poor who left (although they may have been more motivated). For Sweden for example the number of emigrants was equal to about 1/3 of the entire population in 1850. One third of entire population group is quite unlikely to have been be very selective.

Most importantly, since the emigration took place more than one hundred years ago this effect would have disappeared through reversion to the mean. Unless we believe in extreme genetic determinism one generation of selection (on say motivation or IQ) will not have large impacts, especially since we know that the selection itself was not very strong. Indeed there is no difference between the IQ of Europeans in America and in Europe.

A slightly different argument would be that the cultural elite emigrated and have maintained their norms. While this is simply not true for most countries (consider south Italians and the Scoth-Irish) I will note that English Episcopalians are an elite in the US, and that one of the most important regions of emigration in Swedish was entrepreneurial "Småland".

2. The ancestry groups include recent immigrants (who are more likely to be elite).

This is true, but the effect is small. For example there are 4.3 million Americans with Swedish ancestry in the US, and only 55.000 people born in Sweden.

3. We are comparing Europeans in the US with a mix of Europeans and immigrants in the European countries.

Again this is true, but the effect is modest, since all the European countries so far have too few immigrants to seriously effect the numbers (and some immigrants are high-skill). Again using Sweden as an example, those born in Sweden earn 3.2% more than the Swedish average. This reduces the advantage of the US from 55% to 50.5%. The same adjustment should be made in other European countries with immigrants groups that have low income.

4. People who identify their ancestry in the US are more skilled than average.

I believe this is the most important problem. 6% of the population simply identity themselves as "American". This groups earns 12% less than the national average. This groups is particularly large in the Appalachian region, home of the white underclass, settled by the Scotch-Irish. Especially the result for British and Irish immigrants will be upward biased by this group (while for example Swedes are not likely to be affected as much).

In order to err on the side of caution I have included ALL those who identify as "American" as British, even though some are not even white.

5. The share with British ancestry are underestimated.

This is my guess, since groups have mixed a lot, people identify their recent ancestors more than the older, where the British were dominant. In 1790 63% of Americans were British. Campbell Gibson has estimated that 49% of the 1990 population was attributable to the 1790 population. In the 2000 census the sum of British and "American" (mostly British) was 20%, less than what we would predict from Gibsons estimate.