How many World Bank reports are downloaded or cited at all?

About 13 percent of policy reports were downloaded at least 250 times while more than 31 percent of policy reports are never downloaded. Almost 87 percent of policy reports were never cited. More expensive, complex, multi-sector, core diagnostics reports on middle-income countries with larger populations tend to be downloaded more frequently. Multi-sector reports also tend to be cited more frequently. Internal knowledge sharing matters as cross support provided by the World Bank’s Research Department consistently increases downloads and citations.

By the way, about 49 percent of these reports have the stated objective of informing the public debate.  There is more here, from the World Bank itself, by Doerte Doemeland and James Trevino.

A prediction: from MR alone, this will be one of the Bank’s most widely downloaded reports.

Hat tip goes to Justin Sandefur.


Good one. I recall trying to get some publication from the World Bank in DC, in paperback (back in the days) and finding their book store that was open to the public was pretty lame. This is another example of some NGO existing only to talk to itself and its sponsors enough to get funding. Typical parasitic organization.

The World Bank is a mess indeed--particularly the one in DC. However, I found the one in Angeles City to be absolutely useful. The benefit to that branch is that it's just down the street from a number of brothels where 18 year old Filipino girls start working as early as 1pm. So you have a coffee, jump down for a report on timber exports in central Asia in the 1980s, and then flip through it while drinking tequila out of a third-world teenage girl's belly button.

It's great being king.

The World Bank is an NGO?

Does this speak to the low quality of reports, or to the World Bank's failure in promoting their work to its intended audience?

Or the infintesimally tiny size of the intended audience?

Or that sometimes, no one really cares about the fiscal policy of Latvia.

Wait a minute! Latvia's fiscal policy is important. Latvia's free market, low tax policies have resulted in year after year of 6-8 percent GDP growth. That's a number Americans would love to see in the U.S.

Its a simple power law.


do hardcopies count same as downloads? i have a shelf of them from the 80s

While I agree that the World Bank is probably not necessary, in that anyone really interested in an analysis could probably justify paying for the analysis, I do wonder what would be an optimum downloading amount?. Let's assume that the World Bank exists to supply reports that add more in social value on average compared to what they cost. Lets also say as an extreme case, the least popular of the World Banks reports was downloaded a large number of times and created social value way in excess of what it cost to produce. You might say in that scenario they were not optimizing, in that they were not producing enough reports to satisfy demand because clearly the marginal return was very high, which must mean they are leaving money on the table. In other words they should add more analysts and do more reports. But this would mean that the marginal report would get fewer and fewer readers as they expanded their number of reports. So the case where many of their reports are not read may actually be optimum.

Maybe the World Bank should concentrate on lending (donating?) money than generating reports.

Because this blog is so much more popular, obviously we should only have Koch-funded studies produced at all, and all governments should cease funding any research into the World Bank and all other thinktanks.

They can then use the saved cash to provide a tax cut to the Koch brothers so they can spend even more on the Mercatus Institute.

I agree, not only should rich people not be allowed to do research of any kind, government should take over this public good and create thousands of non-partisan research institutes as part of the administrative branch, responsible only to Obama, in an executive order. It would generate thousands of jobs and even if no one knows or cares about the research produced, basic research is an excellent investment that can only truly be undertaken by government.

Since government research has given us horrible things like the Internet and space travel, I think that's an absolutely horrible idea!

Absurd. Space travel is the most important thing to happen, ever. And probably one day, the World Bank will discover space travel 2, if we just give it enough money.

Space travel?

I'll grant government funded research has given us a few things for it is high spending, but not really those.

It may have primarily kind of solved some coordination problems with the internet, and now even space boosters are accepting that a space program is mostly a feel good ad for interest in science and exploration.

You are obviously not in IT nor do you read /.

While grant funding to Stanford, UCLA and others did create key foundational blocks it would of gone nowhere without the help of Xerox, Cisco, etc. The internet would not be possible without Ethernet and the gov had no part in its development.

Real life is rarely black and white.

This is famously true of academic writing as well. I suspect that most US government publications are the same way; see the recent WaPo article tracing the construction of the "dog and cat fur" report as an example. Still and all, most books published in the private market have few readers, most music fails to reach an audience, and most movies fail.

Yes. A friend, who is a rather successful literary agent and author, recently explained the economics of non-fiction books to me. I was shocked how few had to sell in order to be considered a "best seller," and even more surprised that authors still publish given the likely outcomes.

Probably the World Bank authors have similar motivations.

How few books do you have to sell, to be considered a "bestseller" ?

Downloads are a bad predictor for how many reports have been read, an even worse predictor for how many of these reports have have been understood, and probably not related at all to the IMPACT these reports have. If one report is read by a few people only and has a positive impact on policy it can make real change. I am not saying any of these reports had any impact, I am only saying we do not know it. A useful discussion on this blog may be how to measure impact of research and knowledge product.

Measuring the impact of knowledge? That's a really difficult metric. For example, I'm trying to code into numerical models math knowledge developed more than a hundred years ago. If the people in charge of money a century ago have called the mathematician for a performance evaluation, he must have been fired and starved to death, no doubt.

Another issue, knowledge is not only created. Since human life is limited, knowledge also needs to be transferred from one generation to other. Of course there is some knowledge, like how steam engines work, that can be discarded and left in the hand of historians. But fields like medicine, biology, math, physics......they keep accumulating it everyday.

So, knowledge development should be treated like an educated guess, almost a blind bet on the future. Knowledge transfer can be better evaluated: education.

That is actually a pretty good record. Ex post, one frequently finds reports, data, and research that is a dead-end. There is likely some routine research that for many years is never used, but one year is used for somthing very important.

I have seen private corporations hire consultants to do expensive projects, and the reports end up just gathering dust on a shelf. An ex-consultant you worked for me for a while used to joke about this fact.

A Marginal Revolution link is worth around 1k hits if it's a hyperlink in the text and maybe double that if it's in an assorted link. Most of the hits come that day, and about half spread out over the following year. I'd imagine this link would be lower, since people are likely not going to want to read a dry 30 page World Bank report.

I think the policy implication of this paper is clear: the World Bank should stop meeting directly with government ministers and their staff to present headline findings of reports in PPTs. It should also stop emailing copies of the PDFs out. Instead, finance ministers should be obliged to go to the Bank's institutional repository ( and dig out the relevant report. I am sure that McKinsey, Bain, etc. deal with their clients in the same way, right?

"Almost 87 percent of policy reports were never cited." This ignores there use, perhaps even widespread use, in informing policy papers, backgrounds notes, and intelligence assessments in the government space where their is little public disclosure as to their value (these products are typically not publicly available) and the norms of attribution/citation are informal and less rigorous/comprhensive.

Comments for this post are closed