Congratulations to Statsguy

One of my favorite commenters (although we often disagree) is Statsguy.  Tyler Cowen has just called his guest post over at Baseline Scenario a “must-read post.”  Here’s Statsguy:

My primary contention below is that many of these measures used in the composite Heritage Index have nothing to do with less government, and a lot more to do with good government. It is these measures of good government that correlate to economic growth and drive the overall correlation between the “Freedom Index” and positive outcomes. Secondarily, I will argue that many of the other items in the index (like investment freedom) are not causes of growth, but rather outcomes of growth.

I also like the post.  Although I left a comment, I thought it might be useful to repeat the comment here, where more people will see it:

Nice post Statsguy, I mostly agree. It is widely known that richer countries tend to have much bigger governments than poorer countries.  Among really poor countries few people pay income taxes, and it’s hard to collect a lot of revenue. Of course there are exceptions-Brazil has a pretty big government. It’s also true, FWIW, that among rich countries the very richest (Singapore, HK, USA, Australia, Canada, Switzerland, etc) tend to have somewhat smaller governments than the next tier. Norway is an exception.)  My hunch is that the same civic-mindedness that helps countries set up what you call “good government” also tends to lead countries to set up big welfare states (think Denmark.) I believe that in time the Singapore approach to social insurance will be shown to be better.  But despite my right-wing reputation, I regard Denmark and Sweden as highly successful countries.

Tyler Cowen also linked to a very good post at quantoid.

Using the carpentry metaphor, the Heritage Foundation is taking 9 measurements (or 10 depending on the year) to make a single cut. That is, they are using the 10 variables above to produce a single estimate of the economic freedom in every country. A relatively simple investigation shows that both the Fiscal Freedom and Government Spending variables are very poor indicators of economic freedoms and using them will actually produce less accurate estimates of economic freedom. Imagine a very earnest carpenter who takes 10 measurements in order to make precisely the right cut. However, for two of the measurements he was distracted and instead of writing down 12.3 and 12.2 inches, he writes down 21.3 and 21.2 inches. Averaging the 10 measurements will provide a much worse cut than averaging the 9 measurements that were appropriately recorded. I am not arguing that Government Spending and Fiscal Freedom were mis-measured, only that using these will produce a less accurate measure of economic freedom. Technical details of the statistical procedures, theoretical models and investigation that lead to this finding are [here].

I agree that the best way to measure economic freedom is with the 8 non-size of government categories, and when I did this a few years ago I also found Denmark to be the most free market economy.  I prefer the term ‘laissez-faire’ for the small government model measured by Heritage (using all 10 categories) and ‘neoliberal’ for the model of free markets plus social insurance (such as Denmark.)

BTW, In my various recent posts on neoliberalism I took the shortcut of occasionally using the Heritage numbers for all 10 categories, knowing full well that the 8 category subset would have been better.  But I’ve worked with these numbers enough to know that none of my generalizations in those posts would have changed much with the more appropriate figures (which I didn’t have at my fingertips.)



14 Responses to “Congratulations to Statsguy”

  1. Gravatar of Ted Ted
    28. May 2010 at 09:02

    I’ve never much bothered to look into these freedom rankings because they always seem rather arbitrary and pointless, but it was a nice post to read anyway.

    Also, I think the issue with the variables is even greater than statguy seems aware of. Most of the variables could also likely be captured solely with a broader measure of “corruption,” anyway. Protectionism, at least in modern times, is almost always pursued for corrupt means, designed to benefit political elites and insiders (Grossman & Helpman’s famous “Protection for Sale” paper comes to mind). Excessive inflation is usually a hallmark of countries that desire seigniorage revenue, which is, at minimum, a sign of a deeply irresponsible government, if not an outright corrupt one. Excessive regulation of business entry is almost always a result of a governments desire to capture political rents and bribes, particularly at the local level. It seems like a lot of the variables are probably capturing corruption / political rents more than anything.

    What I think I can draw from the Heritage List is that lack of corruption / rent-seeking is probably the most important thing for economic freedom (or, more specifically, political institutions which inhibit the ability to gain rents).

  2. Gravatar of Jon Jon
    28. May 2010 at 10:29

    But his classification of good government versus limit government parameters is non-sense a priori, and is nothing more than a cognitive bias on his part…

    He, himself, deliberately conflates limited government with ‘little spending’ when he classifies property rights as being a question of ‘good’ rather than ‘limited’ government.

    These distinctions are nonsensical. A government that respects property rights is only that does not nationalize capital, one that does not broadly use eminent domain to direct private investment. Both of which are more commonly referred to as limited government ideas.

    To put this in context, the US has a very large military budget, but no one confuses this large outlay with the broader question of neoliberalism because national security is widely regarded and accepted as a public good.

    Statsguy’s post is a fraud. He deliberately conflates two ideas: one whether absolute spending belongs in the index and two whether ‘good’ government or ‘limited’ government is more important. They are independent.

    Another sad moment of herd thinking here in promoting this nonsense. His correlation analysis is interesting but his conclusion and framing a grossly wrong.

    Are you sure you agree with his meta-narrative?

  3. Gravatar of rob rob
    28. May 2010 at 12:11

    Statsguy, great post!

    Scott, I remember you once claiming (and I hope I’m not taking this too far out of context) that most corruption is caused by regulation. I believe the subject at the time was Mexico and I mentioned that, having worked a lot in Mexico, it seemed to me that the incentive for corruption was great precisely because the perception of corruption was so great; nobody trusts anyone and thus the laws are so burdensome that mild offenses (like making a mistake in paperwork) can lead to jail. Since holding any position in the government (or Pemex) puts one at great legal risk, why not engage in graft (everyone assumes you already are) to be rewarded for the risk inherent in your position? (I now see this situation through the lens of Robin Hanson’s claim that we make laws causing nearly everyone to be guilty so that we may punish arbitrarily.)

    I wonder if there are any stats comparing number of laws against corruption vs. level of corruption (which would somehow need to be measured in an absolute sense and not relative to the local laws). I would expect that a fairly corrupt country like Argentina would have many more laws against corruption than its much cleaner neighbor, Chile. If so this would be a case where less government and good government go hand in hand.

  4. Gravatar of rob rob
    28. May 2010 at 12:18

    Of course there are many types of corruption, as Ted above mentions, but I’m talking about something easy to define like bribery.

  5. Gravatar of Mark A. Sadowski Mark A. Sadowski
    28. May 2010 at 13:11

    I was struck by the very high rates of correlation between several of the subindicies (the “nonsize” indicies of course). This suggests that the index perhaps should be revised through grouping some of the various indicies.

    I also was struck by the lower degree of variation in some of the indicies suggesting of course that little of the overall variation is due to those indicies.

    While the Heritage Foundation likes to tout the correlation between the index and GDP per capita it rarely mentions the fact that there is little correlation between the index and GDP per capita growth. One factor may of course be the failure to take into account GDP per capita at the beginning of the period (the phenomenon of convergence).

    However, I am now curious, based on these results, if the inclusion of government size serves to hold back the correlation between the index and GDP per capita growth.

    As an interesting aside there is a statistically significant negative correlation between US GDP growth between 1996 and 2009. The US scored lower in the high growth 1990s than it has in the slow growth 2000s.

  6. Gravatar of ssumner ssumner
    28. May 2010 at 18:18

    Ted, In my post “The Great Danes” I claimed that the non-corrupt countries tend to be much more market-oriented than the corrupt countries. But it also should be said that correlation can go both ways. My hunch is that Singapore and HK are less corrupt than China, despite the fact that all three countries have Chinese cultures. Their systems were imposed exogenously, and seem to have influenced their cultures.

    Jon, I am not sure I agree with his meta-narrative. I am much more anti-big government than statsguy, and I interpret some of the data differently. But I think he raised a number of interesting points that have to be addressed, and can’t just be dismissed. I hope to show later than bigger government (ceteris paribus) actually slows growth, and that the correlation he finds is due to reverse causation and/or third factors.

    rob, I think the causation runs both ways. It is clearly true that more corrupt countries tend to be more statist. The correlation is very strong.

    Mark, See my other answers. I think this data involves lots of cross-currents. As I said, causation can go either way with culture, corruption, etc.

  7. Gravatar of johnleemk johnleemk
    28. May 2010 at 18:29

    The quantoid post is interesting. You don’t ordinarily think of Denmark or New Zealand as being freer than the United States!

  8. Gravatar of Lorenzo from Oz Lorenzo from Oz
    28. May 2010 at 21:38

    If corruption is the market for official discretion one would expect that, other things being equal, more corruption where there is more official discretion (or the discretion has bigger financial rewards).

    In Australia, for example, corruption most often occurs in (1) zoning/land realease matter, where there is a lot of official discretions which are extremely valuable and (2) drug and “vice” control, where the basic official discretions of enforcing the law or not are extremely valuable.

    As you know, I do not much like the term ‘neoliberal’ but I like your distinction with laissez faire a lot, since what is called “neoliberalism” is best understood as economic liberalism in the context of welfare states. Indeed, liberalising reforms have often been undertaken by center-left governments precisely to make the welfare state more sustainable.

  9. Gravatar of scott sumner scott sumner
    29. May 2010 at 05:22

    johnleemk, I also think that New Zealand has fewer petty safety rules that drive me nuts (based on my one visit.)

    Lorenzo, Those are good points. Because they disagree with them, right-wingers often assume liberals are not idealistic. But in fact when liberals got new information around 1980 about the efficacy of markets, they used that new information to better achieve their liberals goals. And that requires resources, hence economic growth.

  10. Gravatar of mbk mbk
    29. May 2010 at 05:26

    Very interesting posts, a good start to taking all these indices with a lot more grains of salt. And finally an attempt to explain why the index numbers often don’t make intuitive sense.

    If you’ve actually lived in a variety of countries and compare your subjective experience to these rankings you get a lot of cognitive disconnects. But in the blogosphere the usual approach is to take the rankings as input data and to generalize from there on about the usefulness of some country’s policies. In reality the situation on the ground is often inextricably complex, between groups of broadly similar countries one advantage is offset by another disadvantage, and the result is simply hard to generalize with a single number.

  11. Gravatar of Scott Sumner Scott Sumner
    30. May 2010 at 07:12

    mbk, I agree it mixes apples and oranges. Instead of being viewed as a small government index, I’d view it as an “Extent to which countries approximate the Sumner ideal” index. So I think it is useful, but not in the way it is advertised.

  12. Gravatar of Tracy W Tracy W
    31. May 2010 at 01:24

    But in fact when liberals got new information around 1980 about the efficacy of markets, they used that new information to better achieve their liberals goals.

    As a Kiwi, I agree with this totally. If you read Roger Douglas (NZ Minister of Finance in the 1980s Labour government most of the time) to Ruth Richardson (Minister of Finance in the first 1990s National government) you can see that they came from very different perspectives. Roger Douglas was all about the social safety net, and using markets to produce resources to raise taxes to provide social services, Ruth Richardson at one point in her autobiography said that if the unemployment benefit was dropped tomorrow everyone on the benefit would be able to find jobs. Roger Douglas was left-wing, Ruth Richardson was right-wing. And they both get called neoliberals.

  13. Gravatar of ssumner ssumner
    31. May 2010 at 06:12

    Tracy, Excellent example.

  14. Gravatar of Taming Leviathan, Part 2 « azmytheconomics Taming Leviathan, Part 2 « azmytheconomics
    27. May 2012 at 07:55

    […] abuses of power. more open countries have bigger governments, but whether a government is good or bad is more important than the size […]

Leave a Reply