Wednesday, March 10, 2010

The Super-Economy in one picture

We noticed earlier that when Sweden is compared to the U.S, Sweden does better for some part of the income distribution. However, if we want to compare systems, we should look at comparable population groups. For this reason I have graphed Swedish income compared with income of Americans with Swedish ancestry.

Read my previous post on my method, which combines GDP per capita data with income distribution data. Read an earlier post on the rationale of using Swedish-Americans rather than all Americans as the comparison group to Sweden.

The graph is income per capita in Sweden and the U.S (for Americans with Swedish ancestry) for 10 income groups, based on official Swedish statistics and census data. I define you as American with Swedish ancestry if the main ancestry group is Swedish. Americans with Swedish ancestry have a 55.8% advantage in income compared to people in Sweden; very close to the figures I estimated using similar underlying numbers and a somewhat different methodology (is a good sign).

The results of the comparison is striking.

Swedes under the American small-government system beat Swedes in the Swedish welfare system for almost 90% of the income distribution. Among the first 10th percentile the Swedes in Sweden do better. By the 15th percentile or so the Swedes in the U.S have caught up, and vastly outperform Swedes in Sweden for the rest of the income distribution.

Here is the same exact picture with income differences filled in.

The median is 42% higher for the Americans compared to the Swedes.
Overall, the middle 60% of the population earn 46% more in the American Super-Economy versus welfare state Sweden.

Now, these figures are slightly unfair to Sweden, because we are comparing ethnic Swedes in the U.S with total Sweden, which includes immigrants (many of whom have not done well in the job-market). For the aggregate figures this is not important, since immigrants are few. Non-immigrants earn 3% more than average of Sweden, simply since most of the average is made up of non-immigrants. The poor in Sweden are disproportionally immigrant however and I would guess Sweden would do even better in the first and second decile if we excluded immigrants. Nevertheless, the advantage in earnings for Americans is massive, indicating that there are large losses in income associated with having a welfare state.

The claims of the left, such as the claim that welfare state is better for ordinary people, are the results of incorrect comparisons. Sweden is homogeneous, whereas the U.S is not. Second, Swedes are a uniquely productive people (as indicated by their performance under the U.S system, and Sweden’s performance before the expansion of the welfare state), which should be taken into account.

The fact that the aggregate results are not even better is due to homogeneous Sweden having innate advantageous in income distribution. Most of the problems of the U.S are concentrated a few distinct populations and geographic areas (such as Appalachia). When we make apples to apples comparisons, it is clear that the American system is better for almost everyone, and almost as good even for the very poor.

The Swedish system is a losing proposal for almost all the population. Unless someone spends their entire life in the bottom 10% of the population, they are better off under the American system.

Here are the lessons from this comparison:

* The U.S would be even richer and have much less poverty if it was made up entirely of Swedes.

* Sweden would be richer than the U.S if Sweden adopted American institutions.

* A large (Swedish-size) welfare state with an American population would do much worse than Sweden.

If you want one picture to demonstrate that free markets and limited government are better for almost everyone in society, this is it.


  1. Interesting post.

    I have two questions that might already been answered in the post, but is not obvious to me.
    1) Does the income take into account all transfers and government services? For example a US family might spend USD15000 on healthcare while a Swedish family might spend UDS150. (I see you have discussed this in an earlier post, so I guess you have tried to adjust for this).
    2) If we assume all Swedes are very productive in the labour market. So much so that most are in, say top 20% of the US income distribution. And assume the US system is such that it rewards the top 20% much more than egalitarian Sweden. How might this affect your conclusion? The way I see it is that if this is true, the US Swedes need the bottom 80% to achieve their incomes, while this 'resource' is not available to Swedish Swedes (not sure I am making sense here).

    Thanks for some great posts.

  2. But are the Swedes who moved to the US likely to be more productive individuals (and families) than the Swedes that stayed home?

  3. Outstanding. I've been following your blog since you began and have really enjoyed your insights.

    Is the data "disposable" income or gross income? Marginal tax rate differences between the countries would make this even greater if the graphs are based on gross incomes. As we finance guys always say: "AFTER tax cash flow".

  4. 1. I have adjusted the data using GDP per capita so that ON AVERAGE absolutely everything is taken into account.

    However for constructing the distribution data some forms of income are missing.
    For Sweden all government transfers are taken into account.

    For the U.S a few are not taken into account, but on the other hand some taxes are not taken into account either. Food stamps, and the EITC are famously not taken into account in the U.S and thus depress the income of the very poor.
    Since the poorest 10% in the U.S get a lot of government aid, like the poorest 10% in Sweden, my judgment is the data is a reasonable reflection of their relative living standards.

    I would say though that percentile 2-4 in the U.S may be relatively worse off, since they earn much that they don't get government transfers, especially health care as you mentioned.

    2. I do understand you second point. Maybe it is the size of the market, and they end up in the top, whereas in a small economy they would not have this advantage.

    My first answer is that Swedes in America are not spectacular in terms of having very high average income (like, say, Jewish Americans). The more important point is that there are few of them if poverty. My second answer is that the European economy is so integrated now that Swedes in Europe have the same chance of being "at the top" if they had the correct institutions.

    3. Nick: that is a very important point. However as I explained in length in my previous post there is absolutely no evidence for this:

    * Something like 1/3 of the entire population moved to America, so it was not a tiny elite.

    * The people who moved were the poorer members of society, not the richest (although they may have been more motivated)

    * One generation of selection is not enough to create long lasting genetic effects (even if say those who moved were smarter, the effect would diminish for their children and grandchildren).

    What I worry most about is that the people who remember they are Swedish are different, and better off, than people who just say they are "American".

  5. PS.

    Nick: I just calculated the I.Q of Americans with Swedish ancestry, using the relible NLS97.

    Americans with Swedish ancestry have an I.Q 1.5 points higher than the Swedish average (which is identical to the average for non-hispanic white Americans).

    The difference between Americans with Swedish ancestry and non-hispanic white Americans is too small to be statistically significant.

    I would guess there was almost no selection in moving to America, and a small selection in terms of recalling that you are Swedish (my theory being that dumb people don't know their ancestry).

    At any case the effect is tiny.

  6. Seven ominous plots and then one plot to rule them all.....

    So "matched" Americans have higher incomes than those that were left behind. Anyway if we correct for differences in work hours (wikipedia report OECD numbers from 2004 that hints at 35% difference, essentially equalizing the medians above; 1777h vs 1316h) where do we end up ?

    How much will current exchange rates influence the results, naively a dollar worth 10 or 6 SEK would have some impact ? I guess, applying purchasing parity would favour US numbers enormously.

    What would it look like if Sweden was replaced by "the better half" Norway ?

    Now my American colleagues say they envy my vacations and "health care", "free higher education" and "safe living with less crime". Being a materialist, I envy their cheap electronic gadgets.....


  7. Yes, because we all know how crime ridden the scandinavian-american areas of the U.S are...

  8. There are a few obvious problems with this thought experiment - the US is a society where income is divided pretty sharply between ethnic groups, causing the selection of a certain ethnic population to remove a large part of the less fortunate ones, whereas the same does not hold true for your selection of swedish nationals. To a reasonably large degree, swedish-americans are also concentrated in less poverty-ridden states in the US, which I did not see you mention correcting for. Thirdly there is a very significant historic effect that I know well enough from even my own family as I am of norwegian descent - since the emigrants from the scandinavian countries had families back home, many of the ones who did not prosper in America upon immigrating there would simply return to their native land as their hopes for a better life were not fulfilled. This also creates a selection (although I certainly admit that the effects of this is hard to estimate several generations later) towards a successful part of the population that is not seen in those that do not have a better alternative to return to if they do not manage to realize the American Dream.


Google Analytics Alternative