Dept of Being Savaged by a Dead Sheep

by Kieran Healy on July 15, 2007

Someone I believe to be Megan McArdle weighs in at the Economist blog on the laughable graphic run by the WSJ the other day. Brad DeLong is not impressed, nor is Mark Thoma (in part because comments are misattributed to him in the post), and nor am I. She singles me out for membership in “a special category of wrong,” I think mostly because my Ph.D is in sociology and not economics.

In his own special category of wrong is Kieran Healy, who vindicates an old professor’s precept that “sociologists rush in where angels fear to tread.” In this case, he lectures a bunch of economists (and economics journalists) on their temerity for throwing Norway out of the set: … I do not deny that policy wonks on all sorts of issues are far too fond of deleting “outliers” that defeat their neat, ideological story of cause-and-effect. …

There is a very good reason that economists, on left and right, are often willing to agree to “Omit Norway” as a first order of business when discussing developed economy GDP. For they are indeed “doing something up there in the fjords to push the observed value up to the top of the graph” and contrary to Mr Healy’s querulous assumption, we know exactly what it is: they are pumping hydrocarbons out of the North Sea as fast as their white-skinned little hands can run the pumps, and selling those hydrocarbons abroad at what are currently very nice prices. Since “find oil” and “experience a rapid appreciation in fossil fuel prices” are not replicable policy recommendations, it is often useful to remove those clever Norwegians from the data set. (Particularly in cases like this, because so much of Norway’s government revenue comes from corporate taxes on oil and gas companies, which are basically lease fees by another name.) Similarly, economists on the left are usually willing to throw Luxembourg (and Monaco, etc) out of set, because “Become an international banking and tax haven” is not generally a policy recommendation they want to see proven out.

Megan doesn’t like me much in part because she has a long history of making an ass of herself on the topic of economics and its status as a social science. Again, just bear in mind that my original post on this was prompted by people saying “I want to see this graph/regression estimate but with Norway (or some other country) removed as an outlier.” After the inevitable snipe at sociology, she concedes most of the point I was making, which was just that when working with small cross-national datasets—e.g., of broadly comparable rich industrial democracies—deleting some countries as outliers in order to make the quantitative analysis go more smoothly is generally a bad idea. It typically doesn’t make much theoretical sense and it also usually isn’t technically necessary because you can use more robust estimators than OLS if you need to. This is just a corollary of a more general principle that one shouldn’t throw away observations to make your life easier. The post was intended as a corrective to the several commenters on various blogs who seemed to think that picking which cases to drop was the starting point of a quantitative analysis.

This simple point applies to the debate about the WSJ graph, but the fact that the figure was basically a dishonest piece of rubbish sort of drowns it out. The original graph included Norway, but probably for the dishonest reason that you could draw the ridiculous curve around it as a result. When discussion started about what the “right” line should be, some people began with the assumption that Norway should be left out just because it was well off the main cluster of points. My post gave some reasons why that wasn’t really necessary, theoretically or methodologically, although—as I said—the basic stupidity of the graph meant that this was something of a subsidiary point.

Megan asserts that Norway’s status as an oil producer means that “economists, on left and right, are often willing to agree to ‘Omit Norway’ as a first order of business when discussing developed economy GDP” because “’find oil’ and ‘experience a rapid appreciation in fossil fuel prices’” are not replicable policy recommendations.” But what exactly is the claim here? If we’re talking about discussing the determinants of GDP in general in developed economies, then every rich industrial democracy has a raft of important, specific, non-replicable features of this sort. A policy of discounting them would quickly mean your “first order of business” would leave you with no cases to discuss. Such features might include, “Have No Feudal Past,” “Have a very large proportion of your population under 25 at the right time,” “Be an island with big coal deposits and start growing rapidly in 1780,” “Have a lot of nice ports and alluvial plains,” “Have a lot of Protestants,” “Be in the middle of Europe,” “Have three to twenty times as many people as the other guys,” or what have you.

Norway certainly isn’t ruled out from consideration for its oil any more than the particular features of other countries rule them out of play. (Besides, Norway is not Nigeria—clearly its government is managing its oil endowment in ways that are within the domain of policy, as opposed to gifts of nature.) It’s facts like this that give rise to what Stephen Gordon, in a related post calls the “vagaries of cross-country GDP data.” My original post, which Megan quoted at length but seems not to have understood, made that point. When generalizing across countries you will want to be clear about your scope conditions, which can include stuff like Norway’s oil industry if it’s relevant. This is different from dispensing with the case altogether. Moreover, demanding a restriction of discussion on GDP growth to only those variables that are strictly within the reach of policy might make it easier to believe that policy-makers can control the world, but you will certainly be ignoring a lot of information about the much more important question of why certain countries are rich—especially those that stubbornly do not fall into line with your favored policies. Of course, such considerations are far from the minds of believers in the Laffer Curve As Economic Law. Nor are they much to the fore in Megan’s mind, her oft-stated belief in the scientific content of economics notwithstanding.

If we’re just talking about the Laffer Curve as policy in particular, rather than GDP growth in general, I presume Norway was included in the original figure precisely because of its apparently high corporate tax rates, which allowed a curve to be dishonestly drawn around it. But as subsequent discussion has shown, in the end there’s really no reason to “throw out” or “drop” Norway in order to say why the curve is stupid, or to calculate a sensible and robust regression estimate for the relationship at hand, in contrast to what some people were originally saying when they saw this ridiculous figure.

{ 2 trackbacks }

Peter Levin’s Rethinking Markets » Blog Archive » Trumping knowledge
07.17.07 at 12:44 pm
Crooked Timber » » Poking the Sloth
07.17.07 at 7:57 pm

{ 89 comments }

1

Kilroy Was Here 07.15.07 at 9:26 pm

It’s interesting to note that Norway and Canada are very similar. Ask the folks in Calgary if oil is helping out their economy? As the folks in the NW Territories if the commodity boom is helping at their mining operations. If we are throwing out Norway for their oil, then we have to throw out Canada as well.

Kilroy Was Here

2

David 07.15.07 at 9:33 pm

Brad Delong makes the point that the reason Norway shows as an outlier on this graph is because oil severance taxes are counted in proportion of revenue [the y-axis] but not in the corporate tax rate [the x-axis]. Corrected, Norway’s effective corporate tax rate zooms from 28 percent to 52 percent, and Norway’s data point moves well to the right, becoming consistent with a linear relationship between rates and proportionate revenue.

3

abb1 07.15.07 at 9:44 pm

Dammit. Thank God we have finally developed good unmanned bombers. Now we can bomb ’em Norwegians (without putting the lives of our brave men and women in uniform at risk) until they pass a sensible oil law, like everybody else.

4

notsneaky 07.15.07 at 9:44 pm

For the record, I’d just like to say that when *I* crack jokes about sociologists, I only do it in the same spirit as, say, the punks hate the hippies (even though they share a lot of common ground), or the people from the west side of town rag on people from the east side of town (even though they really live in the same place) – i.e. because you’re suppose to.

Most real academic work in sociology that I’ve read from the past 10yrs or so (I hesitate to comment on earlier stuff) has been really good. Some of it in terms of statistical sophistication is quite impressive. I do think that a lot of “pop-sociology” that gets occasionally bandied around is annoying but then of course people complain about Econ 101 all the time too, so it’s just par for the course.

And the two disciplines – econ and soc – complement each other much more than many folks realize/admit.

5

P O'Neill 07.15.07 at 10:18 pm

In the text of the WSJ editorial (which no one seems to have discussed, the chart being so laffable), the main country example used was not Norway but Ireland. But in the spirit of —

Similarly, economists on the left are usually willing to throw Luxembourg (and Monaco, etc) out of set, because “Become an international banking and tax haven” is not generally a policy recommendation they want to see proven out.

shouldn’t “become a haven for tax-shifting multinationals” be another of those not widely-replicable recommendations that therefore earns the toss?

6

John Emerson 07.15.07 at 11:05 pm

It’s interesting that even at the level of “The Economist”, glib, lightweight winger ideologues are favored over liberals and even centrists.

As I’ve been saying for a couple of years, when capital gained control of the media we should have expected the terrible results we’re seeing.

(Chomsky, Nader, et al have been saying this for decades, of course, but the Democrats now whining about the Times and the Post, etc., ignored them with extreme prejudice.)

7

martin g l 07.15.07 at 11:31 pm

Yes, the oil is just peachy (thanks, dinosaurs. You guys rock!). And we also invest the oil money in the Petroleum fund. It contributes a sizeable chunk to the state budgets (though not too sizeable, to combat inflation). Last I checked, it was the biggest fund of its kind in the world. NOK 1.980 trillion, says Wikipedia.

8

Elijah 07.16.07 at 12:07 am

“they are pumping hydrocarbons out of the North Sea as fast as their white-skinned little hands can run the pumps.”

Winger dogwhistle. Translation: “And Norway doesn’t have Negroes so they don’t count.”

9

Eric Nilsson 07.16.07 at 12:11 am

The writers of the blog at The Economist site are generally poor economists.
For instance: here .

10

engels 07.16.07 at 12:41 am

They’re not economists, though, are they? What training does eg. McArdle have in economics?

11

engels 07.16.07 at 12:46 am

(From an admittedly shallow reading of her oeuvre I had come to the conclusion that she was not really a “numbers person”, full stop.)

12

josh 07.16.07 at 2:39 am

“As I’ve been saying for a couple of years, when capital gained control of the media we should have expected the terrible results we’re seeing.”
When did not capital not have control of the media?
This does seem a sad decline from the days when E.H. Carr could explain that he got all his news from the Financial Times, since the capitalists needed to have accurate information, and thus their publications tended to be more truthful. (Still true of the FT to a great extent, from what I can tell; and indeed of the non-loony portions of the WSJ — though how long that’ll remain, god knows).

13

walt 07.16.07 at 3:13 am

That comment is by E. H. Carr? I always heard it attributed to Chomsky.

14

Megan McArdle 07.16.07 at 3:20 am

I obviously can’t confirm or deny the authorship of that post, but . . . no. No, I’m not the least bit embarassed, and will continue not to be so, despite your repeated assertions to the contrary. No, I am not going to start being embarassed because you have a PhD and you tell me I ought to be; that got old around the age of twelve for me. No, economists should probably not make throwaway jokes about sociologists, and sociologists should definitely stop making it so much fun by rising to the bait every time. No, I can’t say I’m not flattered by the quasi-obsession that Crooked Timber displays with me, and my putative authorship of posts on Free Exchange.

As for the meat of the post–if you’re this sensitive about a four-year-old crack about sociology, surely you can understand why an economics blogger might find it slightly risible to have a sociologist lecturing economists on throwing out outliers, without the slightest apparent suspicion that there might be a reason for omitting that particular outlier? I mean, yes, all economies are special, like butterflies and children. But natural-resource based economies are special in a fairly regular way, and a way which most economists agree makes them difficult-ish to compare to economies that are not dominated by a single globally traded commodity. Norway is not, of course, Saudi Arabia–but oil and natural gas are a big enough share of GDP that a) the Norwegian government has a source of revenue that is not subject to deadweight loss (but is *incredibly* vulnerable to price change in the underlying) and b) swings in the price of fossil fuels will tend to pretty much completely swamp the effect of any non-lunatic public policy, such as tax rate changes, or education improvements, or whatever.

The problem with Norway is that one way or another, it’s almost always sitting out there well off the curve–only where it sits depends on what a barrel of Brent Crude closed at last quarter. If it was up, Norway’s public policies seem to be pure genius, and if oops, it fell, the Norwegians need to change everything! Now! Brad’s reclassification of the oil excise taxes removes that problem in this particular graph . . . but only by opening up all sorts of other problems, like how to classify spectrum auctions, and what to do about America’s weird international tax policies and . . . that’s when you begin to realise that there’s a reason so many people rely on the OECD harmonised measures. They are imperfect, but they are at least standard.

15

Charles Kuffner 07.16.07 at 3:31 am

What training does eg. McArdle have in economics?

I know she has an MBA from the U of Chicago – as it happens, she was in the same class there as my sister-in-law. Doesn’t quite answer your question, but there you have it.

16

John Emerson 07.16.07 at 4:22 am

What I think has changed is not that people with money control the media (always true), but that all big media groups are publicly traded corporations which are either expected to return maximal return per unit of capital (like any other corporation), or else which are run at a loss as propaganda organs.

Perhaps I should have said, “When finance gained control of the media”

17

josh 07.16.07 at 4:49 am

Walt — I think I read it in Jonathan Haslam’s bio of Carr; but I read that book several years ago, and my copy’s in some other state or country at the moment, so I can’t check. (It could well be true of both Carr and Chomsky, of course)

18

josh 07.16.07 at 4:49 am

John Emersron — fair enough.

19

seth edenbaum 07.16.07 at 4:52 am

a small interruption since the thread I was closed prematurely (it was winding down).
engels,
we agree, all in all. I meant no offense.
groucho.

on with the abuse of “jane”
I got no complaints.

20

Ginger Yellow 07.16.07 at 10:27 am

“It’s interesting that even at the level of “The Economist”, glib, lightweight winger ideologues are favored over liberals and even centrists.”

Why is this surprising? The Economist is a hardline libertarian/free-market magazine. It makes no attempt to hide its ideology.

21

Chris Bertram 07.16.07 at 11:43 am

And there I was contemplating whether to ask Megan to be my “friend” on Facebook ….

22

Chris Bertram 07.16.07 at 11:47 am

BTW, here’s McCardle’s advice to me in an earlier post

http://www.janegalt.net/archives/009555.html

“There’s a valuable lesson for bloggers in there, one that I have only learned in painful embarassment: never make personal attacks unless you are very, very sure that you understand the subject so well that there is no possibility of error.”

23

Matt 07.16.07 at 12:18 pm

Wouldn’t that rule out _any_ personal attack by ‘ol Jane?

24

Barry 07.16.07 at 12:36 pm

Matt, that assumes that Megan was capable of feeling personal embarassment over being wrong. She’s got that glib, knowitall, never admit error attitude that’ll take her far as a propagandist. Considering that she has an MBA from one of the top 10 B-schools in the world, the fact that she’s sunk to propagandist says seomthing about her,

25

Kieran Healy 07.16.07 at 12:54 pm

if you’re this sensitive about a four-year-old crack about sociology … No, economists should probably not make throwaway jokes about sociologists, and sociologists should definitely stop making it so much fun by rising to the bait every time. No, I can’t say I’m not flattered by the quasi-obsession that Crooked Timber displays with me, and my putative authorship of posts on Free Exchange.

Don’t flatter yourself, Megan. As far as I can see, it was you who singled me out to begin with for being in a “special category of wrong.” And if someone does that without knowing what they are talking about, then I will “rise to the bait.” At least you’re happy to admit it’s just bait, as opposed to anything with any intellectual content.

As for the sociology sniping, you did that in the post on Norway, not the one from four years ago. I brought the old one not because I’m sensitive to “a four-year-old crack about sociology,” but to remind people that you’re as shallow, prone to error and resistant to correction on this stuff now as you were back then.

And accusing people of being “obsessed” with you or whatever is just the typical reaction of someone who doesn’t like it when the targets of their uninformed hackery don’t just sit there stunned but actually respond — as Thoma did on his site, and as I do above.

26

john m. 07.16.07 at 1:31 pm

…that’s when you begin to realise that there’s a reason so many people rely on the OECD harmonised measures. They are imperfect, but they are at least standard.

Or, put another way, they are in fact both incorrect and incomplete but at least they are incorrect and incomplete in a standard fashion. Interesting unreasoning.

27

John Emerson 07.16.07 at 3:03 pm

Brd DLng clms tht Th cnmst ws wrth rdng vn fv yrs g. Glt’s hrng ws nt th cs f th dcln, bt prhps mrks th ndr.

Nt t ply th Kt Crc / Wnktt crd r nythng, bt Glt s rprtdly ct s bg. Bt ‘m cnfdnt tht shnngns wr nt nvlvd, nd t wld b hdsly sxst vn t sggst sch thng.

28

wood turtle 07.16.07 at 3:53 pm

There must be a reason why they call it the “Laffer Curve.”

29

wood turtle 07.16.07 at 4:06 pm

Also, does the Laffer curve assume normality? If so, then the data definitely do not appear normal, and should be normalized before graphing.

30

Glorious Godfrey 07.16.07 at 4:30 pm

<> Glt s rprtdly ct s bg.

< hrf="http://www.rsn.cm/nws/shw/119057.html" rl="nfllw"> Jdg fr yrslf.

<> Bt ’m cnfdnt tht shnngns wr nt nvlvd, nd t wld b hdsly sxst vn t sggst sch thng.

Hds nd nclld fr, ndd.

31

notsneaky 07.16.07 at 6:12 pm

Also, does the Laffer curve assume normality

No. But that’s the thing about the “theoretical” Laffer curve, outside the end points (tax rates of 0 and 100%) it can pretty much look like anything (well, to be honest, some shapes are more plausible than others, maybe).

32

Steve LaBonne 07.16.07 at 6:20 pm

But it is so based on theory! For imaginary values of the variable “theory”.

33

Megan 07.16.07 at 6:27 pm

on with the abuse of “jane”
I got no complaints.
and the rest of the ad hominem attacks.

I have complaints. There’s plenty of content there if you need to critique something. It demeans all of us to have the threads here devolve into nasty personal attacks. Provocation is a reason to answer the argument itself. It is not a reason to slam the person herself.

34

mq 07.16.07 at 6:39 pm

That last was Megan non-Mcardle of “From the Archives”, not the subject of the post.

And while I’m not a big fan of Megan McArdle-the-writer, the nasty sexual dig in comment #27 was truly mean and uncalled for. Emerson should apologize, and/or the post should be edited.

35

Kieran Healy 07.16.07 at 6:53 pm

Enough with the sexist insinuations.

36

LizardBreath 07.16.07 at 7:05 pm

Good for you on the editing, and I say this as a friend of one of the people making the unpleasant comments.

37

Megan 07.16.07 at 7:54 pm

Thanks for re-directing the thread and editing that.

Not Megan McArdle

38

Troll 07.16.07 at 8:16 pm

I regret my error. The Economist gave a platform to a worthless, dishonest, incompetent writer for perfectly fine reasons which no one should snigger about.

39

Henry 07.16.07 at 8:47 pm

In re: “the quasi-obsession that Crooked Timber displays with me,” a quick count suggests that we’ve mentioned Megan McArdle in precisely four posts this year, one of which was a link to a bloggingheads that I did together with her, another of which was Kieran sympathizing over the loss of her dog. In contrast, we’re clearly completely obsessed with poor Matthew Yglesias (21 posts and counting – he should be getting worried).

40

Sebastian Holsclaw 07.16.07 at 9:27 pm

“But what exactly is the claim here? If we’re talking about discussing the determinants of GDP in general in developed economies, then every rich industrial democracy has a raft of important, specific, non-replicable features of this sort. A policy of discounting them would quickly mean your “first order of business” would leave you with no cases to discuss.”

Here you kinda go out of your way to compound your initial error. You are making a magnitude error. Yes, we shouldn’t excise 5% here and 2% there until it all goes to zero espeically when 0.5% in wine gets balanced out by 0.8% in beer. But we should note when 45% of it seems to be one non-replicable thing–especially when it makes Norway’s apparent brilliance or stupidity highly dependent on whether or not oil prices (something not under Norway’s control) are up or down any given year.

41

elliott oti 07.16.07 at 9:42 pm

especially when it makes Norway’s apparent brilliance or stupidity highly dependent on whether or not oil prices (something not under Norway’s control) are up or down any given year.

Really? Why not try reading up on Norwegian macro-economic policy first, e.g. at Wikipedia, before making statements like this?

42

elliott oti 07.16.07 at 9:45 pm

But we should note when 45% of it seems to be one non-replicable thing

That’s another thing. The US has a population of 300 million. The Netherlands 16 million. Is this a replicable thing? Will we see these two countries swapping internal market sizes any time soon?

43

nick s 07.16.07 at 10:27 pm

I believe that Megan’s best friend’s cousin had a tax consultant who told her that Norway was, like, totally an outlier.

44

Sebastian Holsclaw 07.16.07 at 10:40 pm

“That’s another thing. The US has a population of 300 million. The Netherlands 16 million. Is this a replicable thing?”

Now you’re just being silly. That is what both economists and sociologists would recognize as ‘controlled’ for when you use things like GDP per capita.

45

LizardBreath 07.16.07 at 11:09 pm

Seriously? You think there are no economic effects due to the scale of the economy? Can I quote you on that the next time ‘reasons why universal health care can’t be provided in the US’ comes up?

46

Kieran Healy 07.17.07 at 12:01 am

You think there are no economic effects due to the scale of the economy?

E.g., to quote an authoritative source, “the division of labor is limited by the extent of the market.”

47

Arr-squared 07.17.07 at 12:21 am

“surely you can understand why an economics blogger might find it slightly risible to have a sociologist lecturing economists on throwing out outliers”

This is a punchline, right? An economics _blogger_ resents being called out on an issue of quantitative methods by a PhD in a reasonably quantitative discipline?

So please spell it out for me, Ms. McArdle, because I don’t understand. Why would an economics blogger find it slightly risible to be criticized by a sociologist?

48

Sebastian Holsclaw 07.17.07 at 12:53 am

“You think there are no economic effects due to the scale of the economy?”

To the extent that oil is the dominant factor of Norway’s economic health? Is that what you mean by economic effects due to scale? Becaue I think it is pretty safe to say that the economic effects of scale don’t equate to around 40ish percent of the economic success of the United States. We’re talking about why you might not want to curve Norway in to certain types of economic comparisons right?

49

Barry 07.17.07 at 1:43 am

(re: using standardized OECD cross-national measures)
“Or, put another way, they are in fact both incorrect and incomplete but at least they are incorrect and incomplete in a standard fashion. Interesting unreasoning.”
Posted by john m.

A shorter term might be ‘least bad’.

Once one starts doing things like comparing features of countries, one is using somebody else’s data (or rather, “somebody elses'”), which was collected using their processes for their purposes.

50

MQ 07.17.07 at 2:13 am

the economic effects of scale don’t equate to around 40ish percent of the economic success of the United States.

I wonder about that. There’s a reason we dominated mass production and mass marketing so early — the first truly united continental-scale mass consumer market. Plus, you know, our size and distance from other countries allowed us to win a number of wars without getting damaged ourselves.

Anyway, this whole Norway thing is pretty trivial compared to the ridiculous Laffer curve graph in the first place. Megan M. has given plenty of proofs of hackishness in the past, but this post of hers is pretty minor. I suspect it mostly set Kieran off because of the typically snotty dismissal of everyone who isn’t a right-wing economist.

51

LizardBreath 07.17.07 at 2:22 am

Actually, what I was explictly responding to was this:

That is what both economists and sociologists would recognize as ‘controlled’ for when you use things like GDP per capita.

If dividing GDP by population were sufficient to ‘control’ for the differences between a large and a small population, that would be a claim (and a surprising one) that there were no effects from scale.

If you’re not literally saying that, but merely saying that while every country has economic factors peculiar to itself, Norway’s dependence on oil renders its economy too peculiar to be compared to other economies, but other economies can be usefully compared to each other, I’d say that’s a strong claim to make without a full analysis of the peculiarities of the other economies you’re discussing, but it wouldn’t be as surprising as the sentence of yours I quoted.

52

Kieran Healy 07.17.07 at 2:30 am

There’s a reason we dominated mass production and mass marketing so early—the first truly united continental-scale mass consumer market.

See, e.g., Alfred Chandler’s _The Visible Hand_, _Scale and Scope_ etc.

53

Megan McArdle 07.17.07 at 4:33 am

Why would an economics blogger find it slightly risible to be criticised by a sociologist on a technical economics question? For the same reason a biology writer might find it odd to be criticised by an economist: each field has its little quirks that those who spend their time following it know about, and those who don’t, er, don’t. Mr Healy’s degree in sociology is irrelevant, except as it may have given him excess confidence in critiquing economics; “The Question of Norway” is not a new one, nor is it an example of some general tendency to throw out outliers. Any proof which relies heavily on Norway to vindicate one’s claims is suspect, because, as Sebastian and I and the Economist’s blog have pointed out, it is a small country dominated by oil wealth. This is not some unique failing to sociologists; almost every field seems to feel that a PhD in it leaves one uniquely qualified to comment on the rest of the academic universe. However, economics, because of its political implications, is perhaps nearly uniquely regarded as ripe for critique by anyone with a degree in anything, from Comp Lit to Physics. Almost no one decides that they can develop “their own ideas” about Unified Field Theory, but almost everyone thinks they can do so about rent control. My point was not that Mr Healy is a sociologist, and sociologists are stupid; it was that Mr Healy is not an economist, and that among economists, the debate about Norway (and Luxembourg, and a few other annoying small countries) has advanced rather farther than generally sound platitudes about not smoothing your data too aggressively.

Incidentally, I agree with one of the commenters that Norway’s smallness makes it special, but not in the way that the commenter seems to mean. Small countries are prone to be dominated by a few sectors, in a way that large countries almost never are: the most oil-dominated economies are small, as are the diamond- and natural gas economies. Bigger economies are more diverse, and hence their GDP does not get swung by movements in the price of a single commodity. In general, scale is a problem when making cross-country comparisons, one that economists have not really solved, except to the generalisation that once you cross a threshold of about 50 million, it probably ceases to matter much. But that doesn’t really help the “keep Norway” argument.

Henry, I refer to the random speculations about my authorship of Free Exchange posts, at least one of which has been in error; and more broadly, to a deeply angry take on even well-meant posts–as when you attributed to me a nasty insinuation about your intelligence, when I had been making what I thought was a rather genial remark about the abuse you had been taking from your left. Given that I am, as you all seem to believe, a complete moron; and also dishonest and generally venial and nasty–well, I find it hard to understand why you dignify my tripe with a response. I don’t think that I have ever posted about any of you with the dripping contempt to which you all treat me pretty much every time my name comes up on your blog; I disagree with you, vigorously, but if I do choose to write about you, I don’t try to convey the impression that really, I’d rather be squashing roaches. I mean, if I, and the publication I work for, are really so awful, why do you spend the time? Personally, I don’t trawl the gutter depths of the internet looking for excuses to be offended.

The various sociological explanations I have been offered are: 1) I work for The Economist and 2) I am female and 3) I have the extreme bad taste not to have a PhD, and to disagree with Your Eminences–none of which seems entirely satisfactory, but leaves me wondering what, exactly, it is about me that seems to so uniquely rub you the wrong way.

Don’t get me wrong; y’all are perfectly free to waste your free time in anyway you choose, and if I devoted any emotional energy to caring what random strangers on the internet thought of me, I’d have shuttered the blog long ago. I am just puzzled, and in the end, curiously flattered, by the apparent depth of your distaste for me. And since I like your blog, even as I disagree with most of its conclusions, I will continue to read it regardless of the invective hurled my way. I’ll just duck in every time one of you declares unilateral victory to point out that this wasn’t quite agreed upon by all the spectators.

54

Kieran Healy 07.17.07 at 4:46 am

Mr Healy’s degree in sociology is irrelevant, except as it may have given him excess confidence in critiquing economics

I have no idea which post of mine you were reading, but the one on Norway had nothing to do with “critiquing economics,” and made no claim nor mention whatever about data analysis as practiced by economists in particular. Indeed, it explicitly introduced itself as a making a point subsidiary to the main discussion about the economics of tax rates and revenue, in response to some commenters (intellectual affiliation unknown) and their requests for more quantitative analysis. The anti-economics spin is entirely your invention. And frankly, the tone of your discussion — “his own special category of wrong,” “sociologists rush in” and all that crap — doesn’t sit very well with your claim to Henry that “I disagree with you, vigorously, but if I do choose to write about you, I don’t try to convey the impression that really, I’d rather be squashing roaches.”

55

alex 07.17.07 at 5:17 am

An interesting discussion. I don’t buy Megan McArdle’s response that

natural-resource based economies are special in a fairly regular way, and a way which most economists agree makes them difficult-ish to compare to economies that are not dominated by a single globally traded commodity.

This argument would hold up equally well if you replaced countries strongly dependent on natural resources with countries that are catholic, protestant, located in africa, have english as their primary langugage, etc etc. Each of those categories shares some regularity and makes cross country comparisons problematic.

At the same time, I wonder about Kieren’s suggestion to use an estimator more robust than OLS. Wouldn’t this be exactly the same as a person looking through the data and throwing out the outliers by hand? Perhaps not mathematically the same, but wouldn’t the results quite similar a vast majority of the time? The only difference is that the outliers would be effectively omitted by a mechanical algorithm, rather than a human being with a possible agenda. Isn’t this simply a commitment to omit outliers in any data sample you run a regression on? Such a commitment might make sense, but I do not buy that it is fundamentally different than the knee-jerk response to throw out Norway because it seems to lie far from most of the other data.

56

observer 07.17.07 at 5:22 am

Someone should also explain to Ms McArdle the difference between nominal and real GDP before she launches into another embarrassing spiel about how changes in the price of oil drive up Norway’s GDP Obviously the task should fall to an economist, because sociologists like Kieran may not have grasped these sophisticated improvements that economists have made to isolate changes in Q from changes in P in terms of how they affect P * Q. This In a way it is sad to see that this is the best the Economist can do in adapting to blogging.

57

walt 07.17.07 at 5:27 am

What’s odd about this is what would be the effect of Kieran’s proposal on the results? It would eliminate most of the effect of Norway on the data. Kieran is making a purely methodological point that would change the result by a very small amount of anything.

58

walt 07.17.07 at 5:28 am

Pwned by Alex, as the kids say.

59

Tim Worstall 07.17.07 at 8:40 am

“This argument would hold up equally well if you replaced countries strongly dependent on natural resources with countries that are catholic, protestant, located in africa, have english as their primary langugage, etc etc. Each of those categories shares some regularity and makes cross country comparisons problematic.”

I believe that at least one sociologist looked at precisely that catholic/protestant (and also at Confucianism and Taoism) split and drew economic conclusions from it.

60

Troll 07.17.07 at 11:17 am

Well, I find it hard to understand why you dignify my tripe with a response.

That’s pretty much my point, except that the institutional importance given Ms. Galt’s tripe by “The Economist” means that the responses she gets will be out of proportion to the intrinsic interest of her writing.

And as we know, responding to tripe is one of the major internet procrastination strategies for those of us who are not hard-charging, up-and-coming young professionals. The tripier the better, I say!

Mmmmmm. Tripe. Menudo. Chitlins. Hong You Du Si. You have to develop a taste for it.

61

Glorious Godfrey 07.17.07 at 12:18 pm

Hmmm, I was making no sexist insinuations. I was actually mocking the plausible deniability conceit of the preceding post (#27). My dislike for the Economist is certainly not so intense as to warrant sleazy potshots at its staffers.

I should have been less cute. I’m probably an idiot, all things considered.

A particularly abject apology goes to Jane, of course.

62

Megan McArdle 07.17.07 at 12:43 pm

SInce I can’t decipher the Dread Insinuation, no apologies needed, I suppose.

Kieran, certainly the fact that you’re a sociologist makes you feel qualified to comment on basic statistics, no? Except this is less a basic statistical question.

And Alex, you’ll be surprised to know that in fact, there is resesarch about whether different national religions affect the economy (answer: the evidence is murky at best), and that Africa is often thrown out of data sets: when we decide, for example, to concentrate on rich countries because their taxation and other public policies, and the results therefrom, differ substantially from those of poorer countries. For example, no one would try to develop some sort of generaliseable rule about pension schemes by looking at a global dataset, because the main thing that makes OECD schemes relatively successful is that they started with a lovely, large, thoroughly endogenous pool of GDP to draw from. A social security scheme that was parsimonious by European standards would bankrupt Burundi. So if you’re looking at pension schemes for the developing world, you throw out the top of your global data set, because you know in advance that it will screw up your numbers.

Development economics is constantly wrestling with the “omit Africa” problem, because for reasons we don’t really understand, the impact of development projects varies substantially, and surprisingly, by continent. Economics suffers badly from having a maximum of 186 N, which differ wildly in everything from size to culture to natural resource endowments. IT is far from ideal, but that’s what we’ve got to work with.

63

Megan McArdle 07.17.07 at 12:52 pm

Also, Alex, by the way: no it wouldn’t work equally well. A “non-anglo” set would include Japan and Luxembourg. A “non-protestant” set would include both Ireland and Portugal. A “non-catholic” set would include both Britain and Greece. Etc. Moreover, these aren’t tenuous theoretical relationships. You can take any natural-resource-heavy economy you want and observe exactly the same, really not very remarkable thing: the economy, and government revenue, varies pretty much in lockstep with the price of the resource. That is not true of the other variables you named–both Jamaica and America have English as their national language. This is why there is a robust economics devoted to natural resources, and much smaller or nonexistant research body devoted to the other issues you named. Even the Africa problem has been belied by a few success stories.

64

albert 07.17.07 at 2:47 pm

My lurking stops here. I can’t quite parse the statement, “Except this is less a basic statistical question” but I think it gets to the whole point of Kieran’s comment and the root of the disagreement.

It’s not really a matter of how sophisticated a statistical approach is, but where how certain assumptions behind statistical analyses produce certain conclusions. The “economics” analysis and the “sociology” analysis are trying to do different things (or develop different kinds of knowledge). Each thinks that the kind of knowledge it’s developing is the best kind, and therefore handles the data in a particular way (each being legitimate within the separate disciplinary fields.) It seems to me that Megan thinks the kinds of claims/theories that economics tends to aim for are the better/the best/the only legitimate ones. I disagree substantively and find rhetorical statements of their superiority dubious. Again though, it’s about what is the best kind of knowledge to have. I haven’t seen much debate on that here.

65

s.e. 07.17.07 at 3:15 pm

Sociology to economics
Pot to kettle.

66

walt 07.17.07 at 3:32 pm

Megan: Kieran’s point is statistical! Let me repeat it for dramatic effect. Kieran’s point is statistical! Kieran’s point is statistical!

I’m beginning to suspect you don’t know enough statistics to understand the point he’s making; otherwise, I can’t make sense of what you’re saying other than marking your territory. Robust regression is a method for removing the influence of outliers, one that has the advantage of being immune to cherry-picking what you leave out. If economists did it Kieran’s way, they would very likely get the same answer.

67

Francis 07.17.07 at 3:33 pm

In his own special category of wrong is Kieran Healy, who vindicates an old professor’s precept that “sociologists rush in where angels fear to tread.” In this case, he lectures a bunch of economists (and economics journalists) on their temerity for throwing Norway out of the set … contrary to Mr Healy’s querulous assumption

later, this:

I am just puzzled, and in the end, curiously flattered, by the apparent depth of your distaste for me.

Nah, there’s NO reason for Keiran to be aggravated, Megan. You weren’t the least bit insulting. Or, to be more blunt, if you really are puzzled about the depth of Kieran’s distaste, then you’ve established my long-standing suspicion that you are a classic example of the Peter Principle in action.

68

Megan McArdle 07.17.07 at 3:36 pm

Albert, no one’s talking about any sort of contested claim of economics. The claim that a nation which has a majority, or large fraction, of its income provided by a single commodity, will therefore see its income vary as the price of that commodity, is basically tautological. Sociology and Economics have different methods largely because they attack different questions. In the areas where they intersect (and where serious work, rather than petty name calling, is being done), they start looking a lot more like each other. Steven Levitt’s reliance on a basically sociological approach to studying drug dealer finance is a good example.

This is a quarrel about domain knowledge, not method or approach. Kieran made a perfectly valid general statistical point–don’t just massage your data until the outliers disappear–but it wasn’t particularly applicable in this case. Sociologists also throw stuff out of their data set when there’s a good reason to believe in advance that it will be an outlier–you might, for example, remove Down’s syndrome children from a study of how income is correlated with parental social class, because for Downs’ kids, unless they’re among the tiny number with vast trust funds, the effect of profound retardation will pretty much overwhelm any other factor. There’s all sorts of interesting ways in which it might screw up your data set: are older women more likely to be high SES, and have Down’s kids? Or are they less likely, because they get sonograms and abort? But whichever way it goes, the fact is that they obscure, rather than illuminate what you’re trying to study, because the unyielding reality is that parents can’t transmit their SES or earning potential to a kid with a 40 IQ. On the one hand, yes, you’re portraying reality a little less accurately in some sense. But on the other hand, the discovery that Down’s kids don’t become investment bankers is not a surprising finding that needs heavy statistical proof, and it muddies the question we’re interested in, which is how much of their SES do parents transmit to kids without disabilities?

The trick is knowing which stuff to throw out. Those in the profession (whichever profession studies the particular question) are more likely to know the answer to that then outsiders, because they’ve already parsed the arguments. There’s no reason that an outsider should know that Norway’s economy is so oil dependent, or that Luxembourg is a banking haven, or what have you. But one should tread a little carefully with those who might have a reason beyond “this data set looks so *messy*”.

69

walt 07.17.07 at 3:45 pm

Oh I see. It was territorial marking after all.

70

Kieran Healy 07.17.07 at 3:52 pm

Megan, look: how is what you’ve just said substantively different from what I said in my original post regarding the importance of scope conditions (for theory) and robust estimators (for methods)? To repeat for the fifth or sixth time, despite your insistence to the contrary my post had nothing to do with any alleged difference between economics and sociology, and everything to do with making a straightforward point to those commenters on blogs who seemed to believe that deleting cases was the way to manage outlying observations. If you read any serious bit of economics or sociology working with this kind of data you will see very similar efforts to worry about structural restrictions, do robustness checks, worry about endogeneity, implement procedures to try to reduce a priori bias on behalf of the researcher, and so on. Because of this lack of substantive difference, I can only conclude that your belief that I have some axe to grind against economics tout court, and your feeling that you are a defender of the forces of economic science against competition from other fields, is what led you to go off on your pointless and insulting rant against me in the first place.

71

albert 07.17.07 at 4:05 pm

Megan-

I don’t think my point was terribly clear above and your response doesn’t address what I’m trying to argue.

I want to avoid making this about your income/parental class analogy, because I think reasoning by metaphor is slippery territory (especially across such a difference of N’s and topics) but I do want to say I don’t particularly agree with how you address the definition of an outlier in that case.

The trick is not, as you say, “knowing what to throw out”, because that is not “knowable” in the prima facie way that you contend. Here’s my main point: I understand there are conventions that insiders in particular disciplines follow, but I those aren’t “answers” or “tricks” so much as projections of the kind of analytical/theoretical work that those disciplines consider important. There’s no reason to assume that something like an ideal type of “resource-rich countries” or “banking havens” exists in the way removing them from analysis assumes they do. The real work is consideration of whether or not statistical validity and logical validity are equivalent when attempting to generate theories of international development using states as cases.

72

Troll 07.17.07 at 4:08 pm

See? Abuse is equally effective and more fun.

73

Walt 07.17.07 at 4:31 pm

Quicker, too.

74

eric 07.17.07 at 4:48 pm

Since no one has mentioned it yet, there was a big debate in the comparative political economy literature (political science flavor) on whether or not to drop or include Norway in models of economic growth.

In Google scholar, try:

garrett lange jackman norway

and a treasure trove reveals itself. The bottom line, as I take it, is that collecting data on the “specialness” of the countries of interest should be preferred. In this case, measuring oil dependence and working that into the model improves the substantive undertstanding that we can take away from the exercise.

75

engels 07.17.07 at 5:14 pm

This is a quarrel about domain knowledge, not method or approach.

This is the nub of it, isn’t it? Kieran was making a rather basic point about statistical methods. Megan otoh is determined to portray this dispute as a turf war between economists (of which she apparently counts herself as one on the basis of her possession of an MBA and a degree in English Literature) and sociologists. In this respect she is being true to the agenda of her publication and its readers who insist that there is an enlightened “economic” way of looking at the world which anyone who reads the The Economist and broadly shares their right-wing pro-market, pro-business ideological assumptions can grasp (whether they happen to economists, or just businessmen who don’t actually know the first thing about economics or even quantitative methods in general) but which must remain forever elusive to the rest of us uncomprehending dullards even if like, to name just a few of that illustrious “newspaper”‘s recent bête noires, Joseph Stiglitz, Paul Krugman or now, apparently, Mark Thoma, they actually are economists and very highly respected ones at that. Needless to say, this is one reason why many of us find The Economist so bloody irritating.

76

Henry 07.17.07 at 6:02 pm

ah but eric, Garrett, Lange and Jackman are all _political scientists_ and ipso facto unrigorous, and to be “consigned to the same circle of pseudo-scientific hell”:http://www.janegalt.net/blog/archives/004125.html as the sociologists. Like Dives, the best we pseuds can hope for is occasionally to be able to glance up from our pit of anguish to the scientific empyrean where Bona Fide Real Economists(tm)like Megan pronounce confidently on who is, and who isn’t, worthy of salvation.

77

Sebastian Holsclaw 07.17.07 at 6:12 pm

“This is the nub of it, isn’t it? Kieran was making a rather basic point about statistical methods.”

Except he made the point by invoking a particular country with particular characteristics which he would would have been aware of if he were an economist or if he had done a bit research. His general point was valid, but his specific illustration was not–and for reasons which are well understood.

78

Walt 07.17.07 at 6:24 pm

Sebastian, are you so tribal that you are actually required to choose the wrong side in this argument?

79

Kieran Healy 07.17.07 at 6:53 pm

Hey, looks like I’m off the hook: Megan now has bigger (much, much bigger) fish to fry and is taking “Cosma Shalizi”:http://en.wikipedia.org/wiki/Cosma_Shalizi to “school”:http://janegalt.net/cgi-bin/MT/mt-comments.cgi?entry_id=9901 on the topic of quantitative methods, endogeneity, the Flynn effect and statistical artefacts.

80

LizardBreath 07.17.07 at 7:03 pm

Sorry about that, chief.

81

albert 07.17.07 at 7:23 pm

#78 His general point was valid, but his specific illustration was not…

Are you paying attention to the discussion at all? The whole issue is that the specific illustration is necessarily valid when his general point is valid.

82

jw@jw.com 07.17.07 at 7:56 pm

I think Megan is going to need to move the goalposts somewhere other than “This a quarrel about domain knowledge.” Brad Delong qualifies as a member in good standing of the cool-kids Econ domain—indeed, much more so than Megan—and he has already opined that there would be no need to omit Norway from the results.

83

Watson Aname 07.17.07 at 8:31 pm

Megan, about half of what your saying here seems to me to boil down to `It’s economics, therefore the economists must be doing the right thing (or at least, something sensible)’. We can generalize it by replacing economics with X; either way it simply isn’t true. There is a smaller part which seems to say that it follows that the lay understanding of X-journalists is bound to be insightful, which is even worse. Besides the broader questions here, the particular plot in question is either inept or intentionally misleading. Shameful either way.

84

alex 07.17.07 at 10:36 pm

Megan,

The set of countries heavily dependent on natural resources is not as regular as you make it out to be; it includes Norway, Russia, Saudi Arabia, and Sierra Leone, the economies of which are quite dissimilar in many ways.

Regardless, nothing of what you wrote really grapples with the fundamental issue discussed here (as I see it): for any country in any dataset, one can always come up with some unique properties that this country possesses and no one else does. Further, you can also give reasons why these properties should have significant economic impacts. So, if this sort of justification is sufficient to leave out Norway, then you ought to be consistent and throw out every country in your dataset.

I can’t see, in anything you wrote since Kieran made this argument in his post, any kind of answer on your part. Mostly, you seem to be giving examples when throwing out data is standard practice; appealing to the practical experience of economists; and repeatedly emphasizing the large effects of natural resources. None of this is any kind of answer to the methodological problem.

85

Megan McArdle 07.17.07 at 11:37 pm

Ah, yes . . . Brad resolved the problem by doing something even more interesting than throwing out his outliers; he altered the composition of a standardised statistic for one, and only one, of the data points, using, ahem, domain specific knowlege of the Norwegian economy.

86

Megan McArdle 07.17.07 at 11:39 pm

And look, I’m not being cute here: my employer won’t let me confirm or deny which posts I’ve written. So baiting me by attributing Free Exchange posts to me is pointless; I’m not allowed to respond.

87

Henry 07.18.07 at 1:56 am

Megan:

Brad resolved the problem by doing something even more interesting than throwing out his outliers; he altered the composition of a standardised statistic for one, and only one, of the data points, using, ahem, domain specific knowlege of the Norwegian economy.

Brad:

Personally I see no need to omit Norway. I do see a need to plot the Norway point on the graph correctly. The revenues plotted on the vertical scale include oil excise taxes levied on corporations. The tax rates plotted on the horizontal scale do not–hence the Norway “tax rate” of 28% rather than the correct 52%. Move Norway out to its proper position–with the same tax concept on both axes–and everything is fine.

For “altered the composition of a standardised statistic” we might more accurately read “corrected a fundamental error in the WSJ graph.” If you can’t see the difference between this and tossing out inconvenient outliers, well … you really shouldn’t be engaging in this kind of debate, should you. If this isn’t actively dishonest, it is pathetic.

88

Jon H 07.18.07 at 2:04 am

“of which she apparently counts herself as one on the basis of her possession of an MBA and a degree in English Literature”

Of course she’s an economist. It says “Economist” right there on her business card.

89

engels 07.18.07 at 3:27 am

Megan #53: “My point [in the linked Free Exchange post] was not that Mr Healy is a sociologist, and sociologists are stupid; it was that Mr Healy is not an economist…” [emphasis added]

Megan #88: “And look, I’m not being cute here: my employer won’t let me confirm or deny which posts I’ve written. So baiting me by attributing Free Exchange posts to me is pointless; I’m not allowed to respond.”

It looks like you already have.

Comments on this entry are closed.