The Scatter Plot Against America

Julian Sanchez | March 29, 2005

Lefty economist Max Sawicky links to an interesting Left Business Observer article purporting to show that the Hertiage Foundation's annual Index of Economic Freedom, or at any rate, the claim that economic performance is tightly linked to economic freedom, is bogus. Here's a graph showing economic growth over a seven year period plotted against economic freedom, and the absence of a correlation:

It's actually the almost total absence of a correlation here that makes me suspicious. If they'd shown that the correlation wasn't as strong or tight as Heritage claimed, I'd be open to persuasion. But the idea that there's no link whatever? That there's no statistical link between heavy-handed economic regulation and stunted growth? That one doesn't pass the straight face test. My best guess for what's going on here is that the graph's baseline year is picking up a bunch of late-80s and early-90s liberalization that nevertheless left the liberalizing countries only somewhat more economically unfree. That is: If you're a basketcase economy to start wtih, you can get a pretty phenomenal growth rate (given the low starting point) just by removing some of the most onerous restrictions and unleashing latent entrepreneurial potential. But that's just a guess; I'm curious what the Heritage folks have to say.

Comments

|3.29.05 @ 1:52PM|#

There's also the problem that what constitutes "economic freedom" is a rather mushy concept.

|3.29.05 @ 1:59PM|#

Maybe, just maybe, the reaction to real-world data that conflicts with your political ideology shouldn't be to assume problems with the data.

Monopolies squash growth. Environmental degradation harms growth. Insufficient tax revenue to cover infrastructure, utilities, and security squashes growth. An imbalance in worker/management power squashes growth by restricting the development of a middle class.

Maybe the assumption that regulation per se harms growth, without looking at the specific consequences of the regulation, bears some revisiting.

|3.29.05 @ 2:05PM|#

I'll pick the lowest apple: Deforesting Europe and eastern North America harmed growth, as evidenced the stagnant population in those regions over the last millenia?

Maybe it ain't real-world data...

|3.29.05 @ 2:06PM|#

joe-

Certainly a good point. I think methodological questions are still worth raising, and in the end I suspect that the measures will turn out to be too crude. That's frequently the case with any scientific study: A more detailed look at the data yields more insight than lumping all of the variables into a single parameter. And it's not like this is the only study ever done on the relationship between prosperity and economic freedom.

But, in the spirit "free minds and free markets", my suspicions can only be treated as a hypothesis. Without more information I cannot simply dismiss this study, just as I cannot embrace the findings uncritically.

And now some of the "free minds" on this forum will undoubtedly accuse me of not caring about economic freedom.

|3.29.05 @ 2:07PM|#

"Maybe, just maybe, the reaction to real-world data that conflicts with your political ideology shouldn't be to assume problems with the data."

I'm glad you said that joe. It will come in handy next time you do so. (not counting your following tongue-in-cheek assumptions)

|3.29.05 @ 2:08PM|#

Show me a monopoly not hiding behind some freedom limiting regulation.

|3.29.05 @ 2:11PM|#

"Show me a monopoly not hiding behind some freedom limiting regulation."

A good example of what's wrong with such a measure as this. If government regulation pushed a certain sector towards monolopy rather than market dynamics, and a law was passed to restrict the monopoly, that law would be pro-growth and pro-market. Yet removing that law would be counted here as "pro-economic freedom," despite the fact that it is actually working against market competition, entrepreneurship, and freedom of choice.

|3.29.05 @ 2:12PM|#

To rephrase one of Julian's guesses, it seems to me that the obvious problem here is that they're measuring static economic freedom against change in economic success.

So China, for example, which is degrees of magnitude less economically free and less economically successful than the US, rates higher on the 'growth' category than the US. Is China a counterexample that economic freedom and economic success are connected?

|3.29.05 @ 2:12PM|#

They make two major changes to growth tabulation in going from US Dollars to PPP and in going from GDP to GDP per capita. While each of these is given a reason no consideration is given to what the combined effect, if any, will be.

As I was not familiar with the PPP, I followed their link to read more and found this:

"By using PPPs as conversion factors, the resulting comparisons of GDP volumes enable us to measure the relative social and economic well-being of countries, monitor the incidence of poverty and progress made towards the Millennium Development Goals and target programs effectively, and assist international markets by identifying the relative productivity and investment potential of different countries."

So it appears that PPP is meant to be used with GDP not GDP per capita stats. Would this give the random distribution pattern seen on the charts? I don't know but may be one avenue of explanation.

Also, converting from an absolute value to a rank as they did can lead to a bit of confusion. Rank increase does not necessarily correspond proportionally to value increase. If I'm the 50th highest paid employee at a company of 100 at $50,000/yr and a $10,000/yr raise makes me 25th highest, my rank growth is greater than my value growth proportionally. There would be a notable difference in where I would be plotted depending which variable was used. This is another road of explanation for the differences they site.

It seems to me that this is like some one showing me an orange to indicate that apples don't exist.

|3.29.05 @ 2:14PM|#

Monopolies don't squash growth - excessive government interference protecting entrenched interests squashes growth.

Environmental degredation doesn't affect growth - as evidenced by the amazing increase in wealth in the West in the last century - excessive environmental protectionism does squash growth.

Insufficient taxes do not squash growth - if the infrastucture is needed someone will build it (and usually at a profit), otherwise its wasted spending - that (and high taxes)do squash growth.

Imbalances in worker/management power relations do not squash growth - interference in setting market clearing proces for labor does squash growth.


You're right that the assumption that regulation per se harms growth is bad - worse is the counter-assumption that every problem can be solved by more.

fyodor|3.29.05 @ 2:16PM|#

Human nature: if a study contradicts your beliefs, attack it as a fraud. If a study backs your beliefs, unquestionably embrace it.

joe, at least give Julian props for even blogging this study in the first place and for merely taking a measured tone of skepticism against it versus some more hysterical reaction. If someone told you they had proof that up was down you might be hesitant to believe it immediately. No one study should be taken as verbatim truth until its methodology is thoroughly reviewed and its conclusions considered along with other studies along the same lines. At the same time, anyone with a serious and open mind will consider all the evidence and not just that which backs their own beliefs. (Duh...)

|3.29.05 @ 2:17PM|#

"Show me a monopoly not hiding behind some freedom limiting regulation."

"A good example of what's wrong with such a measure as this. If government regulation pushed a certain sector towards monolopy rather than market dynamics . . ."


But again the root cause here is the government regulation pushing towards monopoly - instead of adding another layer to counter that why don't we instead see what happens when the regulation is cut back in this example?

|3.29.05 @ 2:19PM|#

Joe:
If the range of freedom in the sample ran from, say, the U.S. or Hong Kong to something along the lines of European welfare states, I'd be a little less reflexively dismissive and a little more open to the prospect of efficient regulation. But that's not the range: We're talking about plenty of countries where the regulation in question is of the sort no sane economist of any political disposition would defend. So I'll stick by my "doesn't pass the straight face test"

Jim Henley|3.29.05 @ 2:22PM|#

I had a CFO who was fond of saying, "An analysis done once is no analysis." At a minimum we need to do the same analysis by the scores from other years. One always wonders about study setup in these cases too. Why 1996? (First year of index?) Why 2003? (Last year of data?) Why base the comparison on the score in 1996 and not some average (however weighted) of the score in subsequent years? Suggested next step: take the slices for EF20, 40, 60, 80, 100, 120 and within each cohort plot how their EF score changes in subsequent years.

Gimme Back My Dog|3.29.05 @ 2:26PM|#

chthus hit it on the head. They use rank instead of absolute value because it makes their case look a lot better. Like most distributions, there is very little difference between elements in the middle of the distribution. So the rankings are dominated by samples which are effectively placed at random. It is no surprise that the results look random in this case.

If you follow the link to the change in freedom vs. economic growth chart, you can see that there is a strong correlation between the top twenty in both groups (nine samples vs. three if they were randomly distributed)

|3.29.05 @ 2:29PM|#

Computer: $880
Internet Access: $39.95 per month
Reason Subscription: $19.97 per year
Being lectured to by joe on open-mindedness: priceless

|3.29.05 @ 2:32PM|#

There are no EF scores 20,40,60, etc. These are the relative ranks. This was one major change they employed, got a different result, and then try to say it reflects back on things before they made changes. As an illustrative example: in 2005 Hong Kong ranks #1 with a EF score of 1.35, Belarus ranks #143 with an EF score of 3.99. A score less than three times higher yet a rank more than 100 times higher. Graphing out rank v. graphing out score are entirely different animals. It's just a card trick they are doing here.

Link for the info I used above.
http://www.heritage.org/research/features/index/

|3.29.05 @ 2:32PM|#

It's basically laughable to claim that the correlation is essentially random, as their plot would seem to indicate. That the most repressive regimes behind the old iron curtain were economically decrepit while the greatest increases of freedom in human history has just happened to coincide with the most astounding level of wealth, historically speaking, for even the poor in this country, could be simply be a mater of chance is an argument which I feel no need to take seriously absent a mountain of compelling evidence.

Having said that, the argument is irrelevant, even if we grant its extraordinarily dubious conclusion. In other words, even if it's true: So what? The argument for economic growth rests on its correlation with freedom, not the other way around. Freedom is the inherently good, while economic growth is the rather pleasant side-effect. So if economic growth comes with freedom, great, all the better. But even if you were to prove it doesn't, that hardly means we need less freedom, it simply means we would have to accept less growth.

|3.29.05 @ 2:36PM|#

The axes in the image don't have units attached, so it's hard to say what is really going on here. chthus's post makes it sound like these are rank statistics rather than absolute numbers, in which case the usual correlation coefficient would be completely meaningless (I have not RTFA so I don't know if this is the case here or not). Even so, for this sample size a scatter plot should visually indicate *something* if there is a significant rank correlation. It may be that the connection is less than we want to believe; it may also be that the parties here are using different definitions of "growth." There may be time lag factors involved, as Julian said. At any rate, it is on the surface a counterintuitive bit of data, so it needs to be examined more closely, and with an open mind.

|3.29.05 @ 2:39PM|#

The charts tell me that no country's GDP declined aboslutely from 1996 to 2003 which is nonsense (no negative numbers) and no country's IEF score declined from 1997 to 2003 (also nonsense). The IEF is not on a scale of 0 to 140 so I'd like to know what they did to the IEF numbers.

For a course, I once ran GDP per capita (in constant dollars at PPP) against Heritage's IEF, Transparency International's Corruption Perception Index, Freedom House's survey of politcal rights and civil liberties and the UN's Human Development Index. Because the surveys don't go very far back I did not use change because long recessions (e.g. Japan) or bubbles (e.g. the US) skew the results.

The best correlation was between low corruption and high per capita GDP (R2 of about .8), then the UN HDI (not surprising as GDP per capita GDP is a component of the HDI), then the EFI (R2 about .5 as I remember) and last was the Freedom House data (R2 of about .4, as I recall). What it comes down to is the rule of law is arguably more important than anything else if you want to become rich, which conclusion makes perfect sense if you are an liberal economist (it's the uncertainty, stupid).

If anyone's interested, I can post the actual R2 numbers later today.

SP

|3.29.05 @ 2:47PM|#

Chuck,

They don't use units because a rank is simply a unitless ordering. They mention this as the case when they say:

"In the exercises that follow, the 125 countries with scores for both 1996 and 2003 were ranked by their freedom scores and their growth rates, and their positions in the two sets of rankings compared."

|3.29.05 @ 3:10PM|#

I would not mind living in a nation that ranks dead last in growth so long as median personal income is very high and everyone is free.

|3.29.05 @ 3:18PM|#

After further consideration, I've got to call bullshit on this whole article.

1) The first graph is all but meaningless. They even acknowledge that the first graph doesn't compare the same measure when they state, "Running the numbers on the report's measure, the improvement in the index vs. GDP expressed in 1995 U.S. dollars, produces somewhat better correlations (.33, to be precise)" in introducing the second graph.

2) They are trying to compare a different application of the same measure by trying to compare change in value with change in rank. This fails to recognize several things including that states with high freedoms to begin with will continue to have growth despite relatively poor showing in freedom increases. The US ranks quite low in advances in economic freedom from '96-'03, barely above NoKo, and yet it has better growth. Surprise surprise.

3) Even with the change to overall rank, some effect is still noticeable. They attempt to explain this away and go so far as to state, "An analysis of the second chart by the naked eye would conclude the relationship is as good as random."

This is not true. The naked eye shows that 9 of the top 20 freedom increasers by rank show up in the top 20 by growth, whereas one shows up in the bottom 20.

This is typical stat manipulation. change a few things and set up a false comparison in the first graph, show it to be entirely different. Then make an actual comparison with a few things still alterred, belittle the results and make a false plea to "the naked eye."

|3.29.05 @ 3:42PM|#

The argument for economic growth rests on its correlation with freedom, not the other way around. Freedom is the inherently good, while economic growth is the rather pleasant side-effect. So if economic growth comes with freedom, great, all the better. But even if you were to prove it doesn't, that hardly means we need less freedom, it simply means we would have to accept less growth.

Yes, to a libertarian. Same as with gun rights--it's the freedom that's important, improved crime stats are really beside the point.

Yet, to actually win the argument (in the real world), you have to convince people who are more interested in crime stats or economic growth than in freedom.

|3.29.05 @ 3:55PM|#

Here comes Don to tell us libertarians we live in a dream world. Christ, we know already, Don. If people would just drink our kool-aid and take a nap with us, all the world's problems would be fixed! :)

tomWright|3.29.05 @ 4:06PM|#

Adding a single drop of water to a perfectly dry desert increases it hydro-freedom factor by 100%.

But that dosn't make it a rain forest.

Tom

|3.29.05 @ 4:06PM|#

um, the financial times said pretty much the same thing here* and here** :D they're both under password if you don't believe me!

---
*Martin Wolf: Big spending doesn't mean less growth - "What we must abandon is a debate that takes the form of 'public sector bad, private sector good', or the other way round."

**Martin Wolf: The issue is better, not higher spending - "Some argue that higher taxes automatically make a country uncompetitive. But countries do not compete like companies."

|3.29.05 @ 4:35PM|#

chthus--

As I said, I did not read the article; for the time being, I am just going by Julian's post and the comments. It wouldn't have hurt to label the axes "growth rank" and "freedom rank", which would have worked just fine to convey the system of measurement being used.
It's just kind of sloppy, that's all.

|3.29.05 @ 4:39PM|#

This sort of reminds me of the lists that I see that rank the best cities to live.

When you look at the details you see that cities get high marks for having many libraries and parks. Whoo-hoo! Outdated books and dog poop -- that's what I call living!

Having plenty of Wal-Mart and Barnes & Noble stores rank high on my list.

|3.29.05 @ 4:41PM|#

If the chart can be believed, then it's both a case against freedom AND a case against regulation. Economically speaking.

|3.29.05 @ 5:03PM|#

jc has a good point. What would it mean if the neither the libertarians or the regulators could affect the ecconomy?

|3.29.05 @ 5:08PM|#

jc-

Absolutely, and I would argue that makes it a de facto case for freedom. Like the "case" for gun control, the removal of freedom (in this case through taxation and regulation) can only be sold on the basis that it provides some benefit. If no real benefit can be had, then why take away the freedom?

Unfortunately, the world we live in often gets it backwards. Without a real benefit, then no freedom can be granted by the gov't. I mean, why should we let people run around with guns starting unregulated businesses if there's no real benefit to be had?

Mark Bahner|3.29.05 @ 5:38PM|#

I'd say the old saying is right: "Figures don't lie, but li...er, leftists can figure."

Let's look at some of the lowest ranking countries in economic freedom:

North Korea: North Korea had an estimated GDP per capita (PPP) of $920 in 1995. By 2003, it had increased to an estimated whopping $1300.

That's an increase of 41%! Woohoo! One of the world's fast-growing economies! That proves that economic freedom is meaningless!

Riiiigght.

I notice that the Leftist Business Observer article is anonymous. That's too bad. I was going to offer the author a free (one-way) airline ticket to North Korea. Provided the author signed a contract agreeing never to come back.

|3.29.05 @ 6:54PM|#

Mark, actually the author is Doug Henwood. His e-mail is [e-mail]

|3.29.05 @ 7:39PM|#

Absolutely, and I would argue that makes it a de facto case for freedom. Like the "case" for gun control, the removal of freedom (in this case through taxation and regulation) can only be sold on the basis that it provides some benefit. If no real benefit can be had, then why take away the freedom?

Like how no one has shown liberalized concealed carry laws to increase crime (and they may even reduce crime); states (like California)continue with strict carry laws.

|3.30.05 @ 12:33AM|#

That Lefty dude's stats are, (here is good stat for ya) 100% meaningless. First off, we don't know the regressions R-squared, f-stat and whether the errors are normally/randomly distributed.

If you don't understand what I just said, don't comment on this story.

|3.30.05 @ 10:30AM|#

Stretch,

The de facto case for regulation is always based on moral grounds. I can't wait for the plot diagram showing a correlation or lack of correlation between morality and freedom.

|3.30.05 @ 12:35PM|#

Low hanging fruit, indeed, comparing America to subsaharan Africa and assuming the amount of vegetation destroyed is the only relevant factor in economic growth. No other variables to eliminate there!

Let's try something closer to apples-to-apples: Haiti vs. the Dominican Republic. Haiti was extensively clearcut. If you fly over Hispanola, the border is quite obvious - Haiti is brown and barren, whereas the Dominican is lush and green.

Mark Bahner|3.30.05 @ 12:46PM|#

"Mark, actually the author is Doug Henwood. His e-mail is [e-mail]"

Hi, Frederick. Actually, I'm not sure I would trust the signature of a man who is sooooo dishonest and/or divorced from reality that he would claim that the Index of Economic Freedom is "meaningless"...and completely un-related to a country's economic situation. As Julian Sanchez commented, it doesn't even pass the straight face test.

I've pulled GDP per capita data from this website:

http://www.photius.com/rankings/

...and Index of Economic Freedom data from the Heritage website. I plan to do my own analyses. I'll withhold my nomination for a Nobel Prize in Economics for Doug Henwood until I'm finished.

;-)

|3.30.05 @ 3:17PM|#

Let's try something closer to apples-to-apples: Haiti vs. the Dominican Republic. Haiti was extensively clearcut. If you fly over Hispanola, the border is quite obvious - Haiti is brown and barren, whereas the Dominican is lush and green.


From the 2005 Index of Economic Freedom:

Haiti: Rank: 145, Score 4.04 (1 is best; 5 worst), Catagory: repressed.

http://www.heritage.org/research/features/index/country.cfm?id=Haiti

Dominican Republic: Rank: 121, Score 3.54 (1 is best; 5 worst), Catagory: mostly unfree.

http://www.heritage.org/research/features/index/country.cfm?id=DominicanRepublic

The Dominican Republic is both freer and more prosperious. And, if joe is correct, has a better maintained environment.

|3.30.05 @ 3:23PM|#

Let's see if these links work:

Haiti

DominicanRepublic