Does the rest of the world subsidize research on the American economy?

When studios shoot movies or TV shows in Canada, they stick US licence plates on the cars and pretend Toronto is New York (the Rudy Giuliani biopic Rudy), Alberta is Wyoming (Brokeback Mountain), and BC is Kansas (Smallville).

Studios do it because anything explicitly Canadian tanks in the ratings. The same rule applies to blogging. Our Palgrave Econolog ranking is determined by how often other economics blogs link to WCI. Recently, a thoughtful Stephen Gordon Canadian-themed post on the federal budget generated just one link. But when Nick Rowe took Stephen's graphs, put slightly different words around them and called the post US Productivity Exceptionalism, the links poured in.

The hide-the-Canadian content rule applies to conventional academic research as well. Recently a co-author said to me, "don't mention that we're using Canadian data until towards the end of the introduction." Our probability of publishing in an American economics journal was higher, he suggested, if we could spark people's interest before we mentioned that we were using Canadian data.

Indeed, a quick survey of a tiny fraction of the research produced by Canadian economists suggests that empirical studies using Canadian data are in the minority. Of the first 26 working papers produced by the University of Toronto economics department in 2010, I counted just two using Canadian data. (The papers using Canadian data don't mention the word Canada. Instead, they employ the strategy that my co-author and I used - say "Toronto" or "Ontario" and hope for a referee with a lousy knowledge of geography). Three focus on China, two use international data, eight use US data and the remainder are theoretical papers using no data. UBC's economics department showed a similar pattern, though I found a few Canadian-oriented research papers in the UBC's Canadian Labour Market and Skills Researcher Network working paper series, and in Carleton University's working paper series about one third of the 2010 papers use Canadian data.

Almost two decades ago, Anthony Scott documented a similar phenomenon. He attributed it to the fact that many academic economists working in Canada were not born, raised or trained here, so have few ties to this country. People will want to keep their employment options in the international academic marketplace open, and so, Scott argued, will do research on topics of international interest, such as trade, as opposed to topics where specific knowledge of Canadian institutions is required, such as public finance. Scott found evidence supporting his hypothesis by looking at the reported research interests of Canadian Economics Association and American Economics Association members, and also articles published in each association's journal.

While I suspect Scott is correct about people wishing to keep employment options open, there are other equally compelling explanations for his findings.

First, it is much easier to find US than Canadian data. The Bureau of Labor Statistics and the Census Bureau provide interactive menus that make it possible to download years of data into an Excel spreadsheet. If similar data is available on the Statistics Canada website, usually at most 10 years of data will be available for free and downloading is difficult, although more data is often available for those with access to university libraries. US data sets may also be bigger, which is handy for studying low frequency phenomena, such as inter-racial marriages or same-sex couples.

Second, academic survival requires getting tenure, and getting tenure requires having research published in economics journals. The majority of economics journals are US based, with American editors and editorial boards. A study published in the Journal of the American Medical Association found some evidence that American reviewers favour American authors: "Reviewers from the United States and outside the United States evaluate non-US papers similarly and evaluate papers submitted by US authors more favorably, with US reviewers having a significant preference for US papers." However this could reflect the effect of English competence - a study in the European Journal of Epidemiology found which country's authors published in the leading medical journals could mostly be explained by the level of medical research funding and average levels language competence, as measured by country-average TOEFL scores (explaining why, for example, Northern European researchers published more than Asian ones). Yet, while suggestive, these studies do not touch directly on the issue of whether studies using American data are more likely to be published than studies using non-US data. Update: in the comments, himaginary links to this paper reporting that papers written about the US are much more likely to be published in top five journals.

Finally, academics want fame and fortune. Citations are the academic equivalent of ratings. Canadian experiences are seen, in the international marketplace of ideas, as particular, specific. American experiences are seen as general. (A particularly absurd manifestation of this occurs in international relations, where people will write "A superpower like the United States..." There is no superpower like the US. But underlying this statement is a presumption that American experience is generalizable, other experience is not.)

Again, it is hard to find hard supporting evidence. A recent paper by Cardoso, Guimaraes and Zimmermann found that the average number of citations per article published is higher for articles written by North Americans than for those written by Europeans, but again this says nothing about the content of the articles in question.

It's sad. I had a really punchy post arguing that the rest of the world subsidized research on the American economy because Canadian and European researchers used US data and wrote about the US. It was backed by anecdotes about blogs, movies and my own personal experience. But I couldn't find any data to support these anecdotes.  And when I look at the most recent issue of the B.E. Journal of Economic Analysis and Policy, and see as many articles about Sweden as about the US, I wonder if it's even true.

But trying to salvage something in terms of policy relevance:

In medical research, it probably doesn't matter whether Canadian or American data is used. A medical treatment that works south of the border will generally work north of the border as well. Indeed, drug trials are not-uncommonly carried out in low-income countries to save costs.

The generalizability of studies based on US economic data to Canada is less well-established. If I was a government person paying for economic research at Canadian universities, I would be interested in knowing how much was relevant to Canadian economic issues.

And that's the question: "Should the Canadian taxpayer subsidize research on the US and other economies?" But I didn't use that as the title of my post, because I figured that emphasizing the US angle was likely to generate more hits.

Comments

[Adam P]

brilliant post.

[himaginary]

Maybe this paper interests you (H/T Tyler Cowen).

As for medical research, this article reports interesting (and maybe infuriating) phenomenon.

[Frances Woolley]

Adam P - thank you!

himaginary, thank you so much. I was looking for something exactly like that, because I remember hearing about that research, but when I searched for "geography academic research" (the key words in the paper's title) all I got was economic geography hits.

They do support my thesis:

Second, papers written about the US are far more likely to be published in the top five economics journals, even after the quality of research has been partially controlled for through fixed-effects for the authors' institutional affiliations; the estimates suggest that papers on the US are 2.6 percentage points more likely to be published in the top-five journals. This is a large effect because only 1.5 percent of all papers written about countries other than the US are published in the top-five journals.

[Nick Rowe]

Good post Frances: a lot of it is probably just because the US is big. People are naturally more interested in their own country; and there are just a lot more Americans than Canadians, Brits, or Swedes. Even Brits are probably more interested in the US than in Sweden, just because the US is bigger.

I just checked the top 30 economics and finance blogs on the Palgrave. All 30 are (AFAIK) US-based, except WCI, BBC Stephanomics, and Free Exchange (and free Exchange is arguably as much US-based as UK-based).

Typo: "And when I look at the most *recent* issue of the"

[Frances Woolley]

Nick, here's an example one of my brother-in-laws gave me that might be persuasive. Think about the number of TV shows set in New York. (You can find a list on Wikipedia, there's dozens). Now think about the number of TV shows set in Cleveland (relatively few, again a list is on Wikipedia). The ratio of New York TV shows to New York population is much higher than the ratio of Cleveland TV shows to the Cleveland population.

I'd explain that by saying that New York experience, being dominant and hegemonic, is seen as generalizable; Cleveland experience is not. (It's not just a New York phenomenon, there's a huge number of TV shows set in LA. And it's also not to do with where the TV shows are shot, see first para of the blog post).

This is the kind of phenomenon we're talking about, the publications/active researcher ratio or publications/population ratio.

(Things might be changing though, that's the subject of the next post).

[Stephen Gordon]

I have a slightly different perspective as a time series econometrician. I don't get the impression that there's much of an advantage in using US data. If anything, US macro data has been so exhaustively studied that the bar is set pretty high for someone who claims to have found something new and interesting.

For me, the larger problem really is the first one you listed: data accessibility and lack thereof.

[Alex Plante]

I follow this blog because I'm a Canadian and want information about the Canadian economy. It took quite a bit of Googling to find this site.

[Stephen Gordon]

Frances, had you seen this post by Jeff Smith (formerly of UWO, now at U Michigan)?

What we might call the "data gap" ... has the effect of leading US researchers, at the margin, to spend their time working on non-US data

[Frances Woolley]

Stephen, no, I hadn't seen that, and it's interesting.

It reminds me, however, of a phenomenon that I see with work on gender issues. When women do work on gender issues, it's seen as somehow girly and tends not to get that much attention. When men do work on gender issues, it's respectable.

Parallel to leading US academics doing work using non-US data?

On the privacy thing: when I was in Calgary, I was quietly amused by the fact that parking meter/machines there require you to punch in your licence plate number and location in order to pass for parking.

Confidentiality concerns anyone? Better not drive your own car when you're doing illicit things....

[Chris J]

Hi all - two thoughts.

1- Stephen's post was mostly about the best short term fiscal policy for Canada - and headlined as such. Nick's post was headlined about the US data. (For the record, as a Canadian, I found Stephen's post interesting and important.)

2- It seems anything Nick writes get read carefully by Brad DeLong, who links to Nick often. Brad is a go-to blogger for discussion of US issues. So Nick's post can get serious US exposure.

[Nick Rowe]

There's a (semi) funny story behind my post "US exceptionalism".

I commented on Stephen's post, saying how he should write a second post, with a sexy title containing the word "US", with the same two diagrams, and the content "Hey, look at the US! WTF?". But Stephen didn't.

So I was having lunch with Frances, whining about how I had run out of ideas, having already posted on every idea I had ever had since undergrad. And how if Stephen had done what I suggested there would be an easy few hits for the blog. So Frances emailed me Stephen's graphs, and said I should do it.

And the hits poured in as predicted. Now the NYT is trying to set up a phone call to discuss US productivity!

I feel a bit like a tart. Put on a short skirt -- it's just so easy.

Scott Sumner linked to Stephen, noting the US graphs.

If Mark Thoma links it gets a lot of readers and other hits. I think he's the most important "aggregator" in the econoblogosphere.

[Michael D]

This Carleton undergrad also discovered WCI after it was linked to by Brad DeLong (and I'm thankful he did that - it's been near the top of my blogroll since, as it's not easy to find Canada-centric economic discussions in blog form).

[Mike Moffatt]

I'm glad I clicked on your 'Rudy' link. I thought you were making a joke, because when I think of 'Rudy' I think of Rudy.

[Scott Sumner]

Well, at least I linked to Stephen Gordon. (But then I had nothing to say about Canada.) As a macroeconomist, I tend to steer clear of Canada because I don't have a good sense of how much of the Canadian business cycle is due to Canadian shocks and how much is a spillover from the US. Perhaps that's not a good reason, but I think it's in the back of my mind.

I find Canada most interesting when it provides independent evidence on some issue, such as the US bank panics of the 1930s (which were avoided by Canada.) The Canadian monetary base fell, which shows Canadian monetary policy was partly endogenous because of the gold link.

[Frances Woolley]

Nick: "had run out of ideas, having already posted on every idea I had ever had since undergrad."

That's why some academics gravitate towards empirical work or history of thought as they age. Empirical work so that they can just keep on applying the same ideas over and over again, historians so that they can relive the intellectual debates of their youth...

[Determinant]

Scot:

Or the fact that we avoided the bank panics of the 2000s, and the foreclosure mess which looks ever so much like the 1930's.

[Kevin Milligan]

Thanks for this post, Frances. In my experience, all else equal, it is harder to publish in US journals using non-US data. Depending on the editor, it can be next to impossible to generate interest. With some practice, however, one can learn how to miminize the disadvantage through how one motivates and describes the work. But you are starting 10 metres behind the others on the starting line.

For me, working mostly on Canadian data was an extremely risky proposition. My department (as do most others) values strongly publications in top 5 journals. I would find it hard to advise a junior colleague to follow in my path. Early in the discussion of thesis topics I warn potential PhD advisees about these 'facts of life.'

Is this good for the Canadian economy to have all of these people working on non-Canadian topics? Well, if the alternative were to force them all to work on Canadian topics, that might clearly pay higher returns to Canadians. But that alternative is not on the table. The more realistic alternative is that if (for esxample) SSHRC mandated Canadian research topics then many of these researchers would simply leave. The cost to their career would be too high.

So, this means the choice is a) having better researchers resident here but not working much on Canada or b) watching many good researchers leave.

Of those two, I think we make the right choice. Why? Well, the empirical people working on US topics bring three benefits.
1) They interact with people who do work on Canada and raise their game.
2) They put out PhD students who do work on Canada.
3) They occassionally give policy advice to Canadian policy-makers.

The UBC adminstration talks now and again about a 'policy school' initiative and wants the Econ department to take a big role. But the problem is, finding enough people who work on Canadian policy AND who are top tier researchers is hard. It is a small set.

[Determinant]

Then isn't the answer to favour publication in Canada (I'm not familiar with Canadian economic publications) and/or create a Canadian refereed forum for such publications? With the Internet you could skip physical publication entirely. The cost would be the server and assembling a panel of economists to be referees while avoiding making the panel too much of an in club. Small communities and all that.

On second thought WCI could develop a subsidiary in publishing Canadian-based research papers. Worthwhile Canadian Papers, or whatever.

As a citizen I would like to see my SSHRC dollars go to support research on Canada. Partly for practical reasons and partly because for the same reason I like seeing shows that feature Canada without hiding us: it's where I live and I'm tired of having to watch American stuff all the time. Give me a lawyer in a gown!

[Kevin Milligan]

"Then isn't the answer to favour publication in Canada" We of course have the Canadian Journal of Economics and Canadian Public Policy (disclosure: which I help edit).

But that isn't the answer unless you can convince the rest of the world to treat a publication in the CJE as equivalent to one in the AER. And that will never happen because citations of CJE papers will always be lower because Canada is a smaller country.

Economics departments, and universities, want scholars of international standing. That means, more or less, publishing in things other than the CJE/CPP. That means not working on Canadian data.

If your idea is for Canadian departments to start unilaterally treating CJEs as equivalent to AERs, then they can do that, sure. But if they start paying out bigger salaries to those with lots of CJEs instead of those with lots of AERs or Econometricas, then those who publish in the AER and Econometrica will beat a quick path to the border to somewhere that appreciates their talent.

[Kevin Milligan]

"As a citizen I would like to see my SSHRC dollars go to support research on Canada. "

You could do that. But the moment you did that, hiring would become enormously difficult for us. How do I tell the bright young candidate (say, from Germany, schooled at Princeton) that if he comes to UBC he has to work on Canada issues if he wants research funding? He would tell me to go stuff myself and take the job at Ohio State instead.

So, maybe you want us to hire the best students from UofT, Queen's, or UWO instead of Mr. Princeton. We could do that, but it would be hard to argue that the students coming out of those programs are on average as good. (Disclosure: I came out of UofT.)

The talent pool we draw from would shrink. We would be a much worse department. How is that good for UBC, British Columbia, or Canada?

[Scott Sumner]

Determinant, Yes, I once did a post saying we should just throw out our entire bank regulation system, and adopt the Canadian approach.

[Kevin Milligan]

Actually, looking at Canadian phd programs to find Canadians to hire isn't much easier. Maybe 1/3rd of our phd students at UBC are Canadian. That's about typical for Canadian programs. In the US, American students are about 50% of the class in the top 10 schools. After that, it's a very high proportion of non-Americans.

North American (and British too) undergraduates typically just aren't interested in graduate work in economics. They have much better and more lucrative options.

So, your challenge is to get Italians trained at Berkeley or Chinese trained at Michigan to want to study Canadian issues, when there is no professional return for them to do so.

Not easy.

[Frances Woolley]

Kevin - as an aside - if we could just clone researchers like you, the world would be a better place.

Responding to your thoughtful and thought provoking comments.

These are very hard issues. In the next few years the mush sector (municipalities, universities, schools and hospitals) is going come under a lot of fiscal pressure.

The question that I suspect some people are going to be asking is this:

What does Canada gain from having an Italian trained at Berkeley and based at UBC publishing in the American Economic Review (AER) as compared to an Italian trained at Berkeley and based in U Washington publishing in the AER?

Yes, the department and the university gains status and prestige, but perhaps that's a vanity project we can't afford in hard times.

Sure, the person based at UBC can train PhD students in Canada, but is that cost effective? If the students are international and the faculty are international, why does it matter if the interaction takes places north or south of the 49th parallel? And for teaching of undergrads - does one really need a PhD to teach Econ 1000?

Anthony Scott seemed to be of the view that people eventually adapt to Canada and start doing more Canada-specific research. I'm not sure that this is true any more - the internet makes the costs of international collaborations so low, and it's easy to get up in the morning, read your home country newspaper, listen to your home country podcasts, etc., that there's not nearly so much of a need to adapt to local conditions.

If undergraduates aren't interested in graduate work in econ, perhaps that's a market signal worth responding to. To be quite honest, I would advise anyone who was interested in becoming an academic economist to take an undergrad degree in math with perhaps a sprinkling of econ on the side.

I've hesitated to write about these issues because I figured that any alternative to the status quo was likely to jeopardize my present extremely comfortable lifestyle.

But my impression is that the policy debate is starting to happen. So the gloves are coming off!

[Jacob Moeller]

Interesting post! I do find this paragraph to be dubious:

Canadian experiences are seen, in the international marketplace of ideas, as particular, specific. American experiences are seen as general. (A particularly absurd manifestation of this occurs in international relations, where people will write "A superpower like the United States..." There is no superpower like the US. But underlying this statement is a presumption that American experience is generalizable, other experience is not.)

It seems like I hear this sort of statement: "a mid-size economy like Canada," often enough. You can make anything generalizable if you want to sound like your claims are sweeping.

Also, isn't your blog itself counter-evidence? You have a fairly popular blog, and yet it has "Canadian" right in the title.

[Determinant]

To be a devil's advocate, why specifically should we devote Social Sciences and Humanities Research Council dollars to non-Canadian topics to the detriment of time and money devoted to Canadian topics? Tell me, as a citizen and taxpayer, why my dollars should go to boost scholarly careers and university standing instead producing results, research, advice and policy ideas based on Canadian experience?

If universities want to have "world-class" scholars who do US-centric research and who happen to live in Canada, that's totally fine, just don't ask SSHRC and taxpayers generally to pay for it.

Whatever happened to he who pays the piper calls the tune?

Jacob:

Nick and Stephen both posted substantially identical entries. Stephen used Canadian data and got comments from me and the other members of the Maple Leaf Peanut Gallery. Nick used US data and got follow-up calls from the New York Times. Therein lies the difference.

[westslope]

Is the term subsidy appropriate?

Should Canadian taxpayer dollars pay for research that uses US economic data?

Yes. The 2 economies are similar, and US trade is of the utmost importance to Canadian wellbeing.

Frankly, I always thought Canadian was in some respects over-represented in US peer-reviewed publications. I assumed it was because we once upon a time had a professional, well-financed national statistical agency. That was BSH (before Stephen Harper).

[econometrician]

Based on their comments here, it would seem that Frances and Kevin actually believe that academic research on the Canadian economy would somehow benefit the Canadian economy. As an academic economist, I would strongly disagree. To paraphrase Richard Feynman, economics is about as useful to the economy as ornithology is to birds.

[westslope]

determinant wrote: "Nick and Stephen both posted substantially identical entries. Stephen used Canadian data and got comments from me and the other members of the Maple Leaf Peanut Gallery. Nick used US data and got follow-up calls from the New York Times. Therein lies the difference."

I disagree. Let me explain the difference. No Canadian journalists telephoned Stephen Gordon concerning his blog post about the Canadian economy. US journalists telephoned Nick Rowe concerning his blog post about the American economy.

If Stephen wrote the blog in French, Le Devoir might have called. As for the Globe&Mail, the Financial Post, and the Canwest papers formally owned by Asper family, would anybody consider them to be in the same category as the agenda-setting elitist New York Times?

[Frances Woolley]

Econometrician - "it would seem that Frances and Kevin actually believe that academic research on the Canadian economy would somehow benefit the Canadian economy."

What I wrote was: "If I was a government person paying for economic research at Canadian universities, I would be interested in knowing how much was relevant to Canadian economic issues."

Lots of non-Canadian research is relevant to Canadian economic issues.

Being relevant to Canadian economic issues isn't the same thing as benefiting the Canadian economy.

Moreover, economic research is not just about studying the economy. It's about studying people, firms, institutions, etc etc.

At the same time, there are several hundred academic economists in Canada. Typically they devote approximately 30 to 40 % + of their time to research. Do the math: # economists * proportion of time in research * average academic salary.

That's a good chunk of change to just throw at people and say 'work on whatever you feel like working on.'

(On the other hand, is it worse than the alternatives? I'm not at all convinced of the ability of governments to anticipate upcoming economic issues, or their desire for non-partisan analysis of such).

[Determinant]

Why should the Government of Canada and the provincial governments pay for research that isn't of interest or benefit to them or to the public?

If academic economists can't get funding for their research outside of government, i.e. through private donations, corporate chairs or whatever, then is it of value? This reminds me of Nick's post on the lack of a free market when he was carrying out course scheduling duties as chair of the department.

How many publicly-funded economists would take a hard line on a Crown corporation that without government funding would be a dismal market failure and promptly go bankrupt? No, I'm not being over-the-top, change economists to "nuclear scientists and engineers" and you've got AECL.

What do funding arrangement say about the credibility, integrity, honesty and relevance of Economics research in Canada?

[Greg Tkacz]

I don't think I'd attribute the lack of work on Canada to a lack of data -- last time I checked CANSIM contained 41 million series, which should keep many time series econometricians gainfully employed for a while.

Perhaps it's related to my upbringing at the BoC, which encourages and rewards work with a strong Canadian focus, but Canadian data just tends to be a lot more interesting. A lot of funny stuff is happening in there which might not appear in U.S. data, so this forces one to be innovative.

For publishing purposes one can add some U.S. results to make a paper more palatable to a U.S. audience, but by no means is this mandatory (and, hopefully, not every paper is targeted at the AER).

[Jacques René Giguère]

Numerous analysis of Science Citations Index showed that low scientific productivity in no-western countries ( adjusted for GDP) was an artifact.
SCI indexed only the subjects studied in the U.S, If you were into soil sciences or the taxonomy of plant eating bugs, sorry you were as good as dead.
So continue to plant U.S.data in your articles, the way movie producers bring garbage to Toronto so it looks like New York.

[Jacques René Giguère]

And I discovered WCI through Krugman's NYT blog..