Marginally Zero-Sum Efforts

Bostrom re­cently noted the prob­lem of the com­mons in la­bel­ing efforts “im­por­tant”; each man­age­rial player has an in­cen­tive to la­bel their pro­ject world-shak­ingly im­por­tant, even though this de­val­ues the pri­or­ity la­bel as used at other times or other pro­jects, cre­at­ing pos­i­tive feed­back in in­flated la­bels.

This re­minds me of how my grand­father, a pi­o­neer in quan­ti­ta­tive ge­net­ics, reg­u­larly be­moans the need to write more and more grant pro­pos­als to main­tain a con­stant level of fund­ing. It’s not that the fund­ing is dry­ing up in his field. But sup­pose there’s money for 20 grants, and 21 sci­en­tists in need of grants—or one sci­en­tist who’d like to run two pro­jects, or re­ceive more fund­ing for one pro­ject… One sci­en­tist doesn’t get his first grant pro­posal funded, so he writes an­other one. His sec­ond grant pro­posal does get funded, which uses up a grant that could have gone to an­other sci­en­tist, who now also has his first grant pro­posal de­nied, and has to write and send off a sec­ond grant pro­posal too...

The prob­lem here is that, while some ini­tial level of effort is benefi­cial, all effort be­yond that is marginally zero-sum; there’s a marginal re­turn to the in­di­vi­d­ual on ad­di­tional efforts, but no marginal re­turn to the group. If there are 20 grants, then ul­ti­mately only 20 grant pro­pos­als are go­ing to be funded. No mat­ter how many grant pro­pos­als any­one writes, the to­tal fund­ing available re­mains the same. Every­one would be bet­ter off if ev­ery­one agreed to write only one grant pro­posal. But in this case, there wouldn’t be much com­pe­ti­tion for any given grant, and the re­wards for writ­ing an­other two or three grant pro­pos­als would be huge… un­til ev­ery­one else started do­ing the same thing.

There’s no ob­vi­ous limit to this pro­cess; the 21 sci­en­tists could write 1,000 grant pro­pos­als apiece, and still get only 20 grants be­tween them. They’d all be bet­ter off if they only wrote one grant pro­posal apiece; but any­one who cuts back unilat­er­ally will be snowed un­der.

In a way, this is even worse than the clas­sic prob­lem of the com­mons. A com­mon graz­ing field even­tu­ally gets eaten down to bedrock and the farm­ers find some­thing else to do with their herds. When pro­fes­sional efforts are marginally zero-sum, but yield pos­i­tive re­turns to the in­di­vi­d­ual, the re­sult­ing cy­cle of busy-work can ex­pand to the limits of in­di­vi­d­ual en­durance.

I’ve of­ten sus­pected that a similar effect gov­erns bu­reau­cra­cies (both gov­ern­ment and cor­po­rate); the longer you stay at your desk each day, the more you are per­ceived as a hard worker and get pro­moted. But there’s only a limited num­ber of pro­mo­tions to go around… and only a limited amount of gen­uinely im­por­tant work to do.

So­cial ap­pro­ba­tion is the usual method for deal­ing with non-pos­i­tive-sum ac­tions. Theft has pos­i­tive re­turns to the in­di­vi­d­ual, but not pos­i­tive re­turns to so­ciety, so we put thieves in jail. But in this case, the so­cial dilemma is that nei­ther writ­ing grant pro­pos­als, nor show­ing up at your office desk, is in­her­ently an evil deed. Some grant pro­pos­als do need to get writ­ten. It’s not in­her­ently a zero-sum ac­tivity. It’s just marginally zero-sum be­yond a cer­tain point.