liebowitzIn a recent post at VoxEU (based on a recent working paper entitled “Willful Blindness”) Stan Liebowitz argues that the assignment of credit by economics departments to academic researchers is “uneconomic.” By this he means that in co-authored papers the credit shares sum to more than 1. Instead, in a survey of economics department chairs he finds that 1/3 of them do not prorate credit at all and assign each co-author a value equivalent to the entire article — the same they would receive if it was single authored. He argues that this generates poor incentives:

If the four-authored paper is not written with each author providing one-fourth or less effort compared to each author working alone, then that size of team is inefficient. But if each coauthor is given full credit, they have an incentive to coauthor even when the number of papers written by the four-author team is much lower than the number of equal quality papers they could write working alone or with smaller teams.

Basically, collaboration is subsidised over single authorship. He argues that only be fully prorating credit can appropriate incentives be given.

This may be a legitimate argument but it is incomplete. The issue is that the co-authored project has multiple authors contributing to its quality. If an article has quality Q and a research by expending their own private effort can raise quality by D, if there is proration, they will receive less than D as a result of that effort. This is a classic public goods problem (i.e., free riding) or what Bengt Holmstrom called “Moral Hazard in Teams.” The only way to solve a problem like this is to not prorate at all; precisely what 1/3 of departments are doing and others are doing to some degree. This is usually impossible to do with physical production because if you produce x units of something, you would have to give out Nx units to the N collaborators which may be impossible. But credit is free. You can assign NQ units of credit to N authors without breaking any bank. Suffice it to say, it is far from clear that a “failure to fully prorate will lead rational researchers to use more than the optimal number of authors.” Liebowitz also argues that a lack of proration may lead to “false authorship” (as it is free to add another author) and should also be applied to citations. For the fraud claims, I think we can agree although I’m not sure how prevalent that may be. For citations, the moral hazard in teams issue still arises.

A stronger argument for full proration was provided by myself and Fiona Murray in a recent paper. We argue that the alternative to co-authorship may be for each author to write up their results in two separate papers. In this case, the share of credit needs to match the share of credit apportioned through citation links or there would be a distortion. In the absence of moral hazard, this suggests full proration. But care must be taken. If papers can be “salami sliced,” then full proration may lead to a proliferation of papers and an inefficient amount of collaboration. Put simply, the interplay in the system of scientific rewards is subtler than Liebowitz is making out.

Finally, in looking at science more broadly, in a recent paper, Michael Bikard, Fiona Murray and I examine the collaboration activities of all MIT academics over the past few decades. We find that their behaviour is not consistent with full proration or zero proration but instead is more likely to be a share based on the inverse of the square root of the number of co-authors. In other words, there seems to be a balance between encouraging an efficient assignment of credit for the co-author or not decision and allowing that assignment to mitigate any issues that may arise from free-riding.

Leave a Reply

%d bloggers like this: