Inadequate Equilibria vs. Governance of the Commons

This is a cross post from http://​​​​blog:128.


In the past I’ve re­viewed Eliezer Yud­kowsky’s “Inad­e­quate Equil­ibria” book. My main com­plaint was that while it ex­plains the prob­lem of sub­op­ti­mal Nash equil­ibria very well, it doesn’t pro­pose any solu­tions. In­stead, it says that we should be aware of such co­or­di­na­tion failures and we should ex­pect our­selves to fare bet­ter than the offi­cial in­sti­tu­tions in such cases. What Yud­kowsky is say­ing (if I un­der­stand him cor­rectly) is that given that the treat­ment of short bowel syn­drome in ba­bies is stuck in an in­ad­e­quate equil­ibrium, there’s no way to fix the prob­lem on the sys­tem level. How­ever, you can spot the prob­lem and ac­quire the med­i­ca­tion needed to keep your kid al­ive from abroad your­self.

In my re­view I’ve sketched a cou­ple of ideas how to ap­proach the prob­lem, but that was just me try­ing to be clever. Some­thing backed by ev­i­dence would make me much more happy.

So I’ve de­cided to have a look at how peo­ple are get­ting out of sub­op­ti­mal equil­ibria in the real world. And that’s how I’ve got to the Elinor Ostrom’s book “Govern­ing the Com­mons (The Evolu­tion of In­sti­tu­tions for Col­lec­tive Ac­tion)”.

The book is very ex­plic­itly cov­er­ing the same prob­lem as Yud­kowsky. As Ostrom say about her re­search in her No­bel lec­ture: “One of the key ques­tions that we’ve been ad­dress­ing is ‘Are ra­tio­nal in­di­vi­d­u­als hope­lessly trapped in dilem­mas?’”

Ostrom is an economist with a game-the­o­retic bent. But she’s not a the­o­ret­i­cian. She’s a field re­searcher. I’ve ex­pected some good read and I wasn’t dis­ap­pointed.

The book fo­cuses on spe­cial sub­set of co­or­di­na­tion prob­lems, prob­lems of man­age­ment of what Ostrom calls “com­mon pool re­sources”. Com­mon pool re­source is some­thing that, un­like pri­vate prop­erty, is not ex­clu­sion­ary, some­thing that one can­not eas­ily re­strict peo­ple from tak­ing ad­van­tage of, but that, un­like pub­lic prop­erty, can be de­pleted by ex­ces­sive us­age. So, for ex­am­ple, it’s hard to ex­clude oth­ers from fish­ing in the ocean, yet, the fish stock can be de­pleted by overfish­ing. Ocean fish­ery is a com­mon pool re­source. Com­pare that with a pri­vate fish farm where oth­ers are pre­vented from fish­ing by a bar­rier. And on the other side, con­trast it with a pub­lic weather fore­cast which doesn’t get de­pleted as more peo­ple turn on the ra­dio.

On the game the­o­retic level, these sce­nar­ios can be though of as “tragedy of the com­mons” or “pris­oner’s dilemma” prob­lems. At the heart of the prob­lem is the fact that while co­op­er­at­ing and limit­ing ev­ery­one’s rate of ex­trac­tion helps to sus­tain the re­source, in­di­vi­d­ual play­ers get bet­ter short-term re­ward if they defect and overuse the re­source. For ex­am­ple, limit­ing the num­ber of cows on a com­mon graz­ing land keeps the pas­ture sus­tain­able. How­ever, each villager is in­cen­tivized to get as many cows as pos­si­ble which will in turn re­sult in over­graz­ing and de­struc­tion of the graz­ing land.

Against eco­nomic absolutism

As Ostrom notes, peo­ple read­ing about this kind of prob­lem tend to jump to one of the two con­clu­sions. One group be­lieves that the graz­ing land should be na­tion­al­ized and the laws should be en­acted to pre­vent the over­graz­ing. The other group thinks that the land should be split into small plots and pri­va­tized.

The prob­lem with the lat­ter solu­tion is that it’s of­ten not fea­si­ble. For ex­am­ple, there’s no rea­son­able way to pri­va­tize an ocean fish­ery. Fish stock is mi­gra­tory and can­not be split among stake­hold­ers. So, if you in­vest in the growth of the fish stock by tem­porar­ily fish­ing less, the fish will even­tu­ally mi­grate el­se­where and get caught by other fish­ers.

In other cases the re­source can be split but do­ing so re­sults in sub­op­ti­mal perfor­mance. For ex­am­ple, if con­di­tions on the graz­ing land vary, some parts may be dry this year, but lush the next year. Some parts may have been un­der­wa­ter the last year but are all right this year. In such cases split­ting the graz­ing land gives sub­op­ti­mal out­come for ev­ery par­ti­ci­pant. Each of them ex­pe­riences a se­quence of good and bad years. Cows die of hunger one year, but then the graz­ing land is un­der­uti­lized the next year. Each herder’s abil­ity to plan ahead is badly com­pro­mised.

As for the na­tion­al­iza­tion ap­proach, Ostrom de­mostrates the prob­lem us­ing a sim­ple game. She starts with the clas­sic tragedy of the com­mons sce­nario. Every­body is in­cen­tivized to over­graze and the pas­ture will be even­tu­ally de­stroyed.

But then the state steps in. Land is na­tion­al­ized and over­graz­ing is pre­vented by law. Par­ti­ci­pants who do over­graze are fined. And when we take the fine into ac­count in our game-the­o­retic calcu­la­tion the out­come sud­denly changes. Over­graz­ing in no longer the op­ti­mal strat­egy. The pas­ture will be sus­tained. Hooray for na­tion­al­iza­tion!

But not so fast! As Ostrom notes, law en­force­ment is not perfect. Some­times offen­ders are not caught and some­times peo­ple are fined un­justly. She as­sumes a cer­tain er­ror rate, she does the calcu­la­tion again and lo and be­hold! The in­cen­tives to over­graze are back!

And we haven’t yet con­sid­ered the cost of polic­ing. If we spent too lit­tle on polic­ing, we’ll ei­ther have less po­lice­men, re­sult­ing in higher er­ror rate, or the po­lice­men are un­der­paid and tempted to ac­cept bribes which will in turn com­pro­mise the en­tire sys­tem.

And, un­doubtly, there are many more facets of the policy that play a role and can af­fect the re­sult of the calcu­la­tion.

It turns out that think­ing about com­mon pool re­sources in ab­solutist ways is not helpful. In re­al­ity, there’s a broad con­tin­u­ous multi-di­men­sional range of policy op­tions. The villagers can split the land into pri­vate and pub­lic parts. They can spend more on law en­force­ment. They can po­lice each an­other. They can limit the us­age of com­mon re­source based on effort spent main­tain­ing it. They can use ro­ta­tional al­lo­ca­tion of land, or maybe a lot­tery. And so on and so forth.

But it’s even more com­plex than that. It not just poli­cies that de­ter­mine the out­come. The na­ture and the par­tic­u­lar­i­ties of the re­source it­self may de­ter­mine the op­ti­mal policy.

Take, for ex­am­ple, the beach seine fish­ery in Mawelle, Sri Lanka. The village beach is di­vided into two launch­ing sites, one on the har­bor side and one on the rock side. There are differ­ent rules for those two sites. The rules also vary de­pend­ing on the time in the day. The study in ques­tion notes:

The Mawelle fish­ers provide a co­her­ent ex­pla­na­tion for why they use this com­plex set of au­thor­ity rules, rather than a sim­ple ro­ta­tion sys­tem, to equal­ize the op­por­tu­nity to make a big catch. Four en­vi­ron­men­tal or tech­nolog­i­cal con­sid­er­a­tions af­fect the prob­lem of equal­iz­ing ac­cess: (1) The har­bor side pro­duces the re­ally big catches, but the rock side is more con­sis­tently pro­duc­tive when there are fewer fish. (2) The first catch of the morn­ing is most likely to be the biggest catch of the day, and the prices are high­est in the morn­ing. (3) The weather af­fects the num­ber of hauls that can be made in the day, and any sys­tem as­sign­ing a set hour of the day would be in­effi­cient. (4) Beach-sein­ing in­volves high la­bor in­puts to pre­pare a net for use and to restack it af­ter­ward, and sim­ple ro­ta­tion sys­tems al­low­ing all nets to be used only once per ro­ta­tion would in­volve higher la­bor cost.

As can be seen, the ge­og­ra­phy, the shape of the seabed or maybe the sea cur­rents, the weather pat­terns, the de­tails of the fish­ing tech­nol­ogy and so on, they all af­fect the op­ti­mal policy for gov­ern­ing the fish­ery. And it turns out that the op­ti­mal policy differs for two sites barely kilo­me­ter and a half apart.

There’s no uni­ver­sal “cor­rect” solu­tion to the prob­lem. A host of minor de­tails mat­ter. And, let’s be frank, we have yet to see an economist who in­cor­po­rates a map of sea cur­rents at a par­tic­u­lar place at the Sri Lankan shore into his eco­nomic model.

The ab­solutist po­si­tion is also un­ten­able be­cause the op­ti­mal policy may change over time. Con­sider cli­mate change. It may cause the weather pat­terns in Mawelle to change. And sud­denly, the gov­er­nance sys­tem that has worked for cen­turies stops work­ing even though none of the rules were changed.

All in all, it seems that or­gan­i­cally grown in­sti­tu­tions are a lot like Hayek’s free mar­kets. They are in­for­ma­tion-pro­cess­ing ma­chines. They ag­gre­gate countless de­tails, too small and nu­mer­ous for any cen­tral plan­ner to take into ac­count, and gen­er­ate a set of effi­cient gov­er­nance rules.

Sus­tain­able institutions

In her dis­cus­sion of the gov­er­nance sys­tems Ostrom picks sev­eral cases that proved to be sus­tain­able over a long pe­riod. She an­a­lyzes com­mu­nal tenure of high moun­tain mead­ows and forests in Switzer­land and Ja­pan, the “huerta” ir­ri­ga­tion sys­tems in Spain and the “zan­jera” ir­ri­ga­tion sys­tem in Philip­pines. All of them are prac­ticed for hun­dreds of year, in some cases even for thou­sand years. I am not go­ing to go into de­tails, but do read the book. At the end she con­cludes that sus­tain­able com­mon pool re­source gov­er­nance sys­tems tend to have the fol­low­ing fea­tures:

  1. Clearly defined com­mu­nity bound­aries, no strangers al­lowed.

  2. Poli­cies are fit­ted to lo­cal con­di­tions.

  3. Peo­ple mak­ing poli­cies have skin in the game.

  4. Sys­tem for check­ing whether par­ti­ci­pants are play­ing by the rules.

  5. Grad­u­ated sanc­tions.

  6. Con­flict-re­s­olu­tion mechanisms.

  7. No ex­ter­nal med­dling (e.g. by the gov­ern­ment).

  8. Nested en­ter­prises (ap­plies only for large-scale in­sti­tu­tions).

I am not go­ing to re­pro­duce the en­tire ar­gu­ment, but I can­not re­sist quot­ing few in­ter­est­ing para­graphs from the book and ex­press­ing some thoughts of my own (in no par­tic­u­lar or­der).

I always thought that so­cial cap­i­tal (or com­mon knowl­edge, if you will) is ex­pen­sive to cre­ate yet very easy to de­stroy. The fol­low­ing tid­bit seems to in­di­cate that it can, in fact, be rather re­silient. It would be nice to see more re­search on this as­pect of the com­mon knowl­edge. Note how it was sus­tained via a two-sided con­sen­sus about the tem­po­rary na­ture of the vi­o­la­tions of the sys­tem:

[Dur­ing the de­pres­sion in 30′s] al­most all the villagers knew that al­most all the other villagers were break­ing the rules: sneak­ing around the com­mons at night, cut­ting trees that were larger than the al­lowed size, even us­ing wood-cut­ting tools that were not per­mit­ted. This is pre­cisely the be­hav­ior that could get a tragedy of the com­mons started, but it did not hap­pen in Ya­manaka. In­stead of re­gard­ing the gen­eral break­down of rules as an op­por­tu­nity to be­come full-time free rid­ers and cast cau­tion to the winds, the vi­o­la­tors them­selves tried to ex­er­cise self-dis­ci­pline out of defer­ence to the preser­va­tion of the com­mons, and stole from the com­mons only out of des­per­a­tion. In­spec­tors or other wit­nesses who saw vi­o­la­tions main­tained silence out of sym­pa­thy for the vi­o­la­tors’ des­per­a­tion and out of con­fi­dence that the prob­lem was tem­po­rary and could not re­ally hurt the com­mons.

As for the “grad­u­ated sanc­tions” point it seems that fines for oc­ca­sional breaches of the rules are very low. So low, in fact, that it still may be prof­itable for the player to break the rule. One study that comes to mind here is the one by Gneezy and Rus­ti­chini that shows that when par­ents were fined for pick­ing their kids late from the day-care cen­ter they be­come more likely to be late. The fine have nor­mal­ized oth­er­wise so­cially un­ac­cept­able be­havi­our. One has to ask whether this effect plays a role in the sys­tems de­scribed by Ostrom.

One ob­ser­va­tion that came as a sur­prise to me was that mod­er­ate fine (as op­posed to large fine which may cause re­sent­ment) can in fact in­crease the trust in the sys­tem. The fined par­ti­ci­pant can ex­pe­rience fist-hand that break­ing the rules doesn’t go un­no­ticed and thus gains con­fi­dence that oth­ers are not break­ing the rules. Think of it as a probe mechanism: Every par­ti­ci­pant has a cheap way to test the mon­i­tor­ing and en­force­ment sys­tem and make sure that it’s still work­ing as in­tended. If the fines were very high the abil­ity to mon­i­tor the en­force­ment sys­tem would be com­pro­mised.

From the chap­ter about Zan­jera ir­ri­ga­tion sys­tem in Philip­pines:

A few parcels, lo­cated at the tail end of the sys­tem, are as­signed to the offi­cials of the as­so­ci­a­tion as pay­ment for their ser­vices. The sys­tem not only pro­vides a pos­i­tive re­ward for ser­vices ren­dered but also en­hances the in­cen­tives for those in lead­er­ship po­si­tions to try to get wa­ter to the tail end of the sys­tem.

There’s not much to say about it. We should use this kind of ap­proach in poli­tics much more of­ten.

I am in­clud­ing the fol­low­ing quote be­cause the over­pop­u­la­tion prob­lem is one of the ex­am­ples of sub­op­ti­mal equil­ibria. In Hirano, Na­gaike and Ya­manaka villages in Ja­pan:

Rights of ac­cess to the com­mu­nally held lands were ac­corded only to a house­hold unit, not to in­di­vi­d­u­als as such. Con­se­quently, house­holds with many mem­bers had no ad­van­tage, and con­sid­er­able dis­ad­van­tages, in their ac­cess to the com­mons. Pop­u­la­tion growth was ex­tremely low (0.025% for the pe­riod 1721-1846), and own­er­ship pat­terns within villages were sta­ble.

If you are in­ter­ested in the topic look also here.

Fi­nally, I’ve en­joyed find­ing out that re­sent­ment-free fair-di­vi­sion al­gorithms are ac­tu­ally be­ing used in the wild. When har­vest­ing wood from com­mu­nal forests in Tör­bel, Switzer­land:

The first step is that the village forester marks the trees ready to be har­vested. The sec­ond step is that the house­holds el­i­gible to re­ceive tim­ber from work teams and equally di­vide the work of cut­ting the trees, haul­ing the logs, and piling the logs into ap­prox­i­mately equal stacks. A lot­tery is then used to as­sign par­tic­u­lar stacks to the el­i­gible house­holds.

It seems that a similar lot­tery sys­tem is used in Zan­jera ir­ri­ga­tion sys­tem in Philip­pines.

How are the in­sti­tu­tions cre­ated?

All of that is nice and in­ter­est­ing, but hey, let’s re­mem­ber what we are af­ter here! We are not look­ing for the best in­sti­tu­tion. We are try­ing to find out how to es­cape a lo­cal max­i­mum. How to jump from an in­sti­tu­tion that sucks, but hap­pens to be a Nash equil­ibrium, to a bet­ter in­sti­tu­tion. The ques­tion thus is not how such a bet­ter in­sti­tu­tion looks like but rather how it gets cre­ated.

When dis­cussing the prob­lem Ostrom shifts the fo­cus from the long-lived in­sti­tu­tions (we have very lit­tle, if any, data on how they were cre­ated cen­turies ago) to the gov­er­nance of wa­ter bas­ins in South­ern Cal­ifor­nia. Th­ese in­sti­tu­tions were crated not that long time ago, in for­ties, fifties and six­ties, and a lot of de­tailed in­for­ma­tion is available.

The prob­lem is as fol­lows. Un­der to­day’s Los An­ge­les metropoli­tan area there are nat­u­ral un­der­ground wa­ter reser­voirs, a wa­ter-bear­ing strata made of sand and gravel. Th­ese reser­voirs get re­plen­ished at a con­stant rate by the rains that fall in the foothills and up­per valleys. They are also used by many par­ties as a source of wa­ter. This nat­u­rally leads to a tragedy-of-the-com­mons-style prob­lem. If ev­ery­one pumps as much wa­ter as they can they are go­ing to suck the reser­voir dry. Even worse, low level of wa­ter in the reser­voir means that sea wa­ter starts to seep in and will even­tu­ally de­stroy the reser­voir.

The situ­a­tion was made worse by the stand­ing law. One was en­ti­tled to spe­cific amount of wa­ter based on how much of it they were us­ing. Thus, lawyers ad­vised ev­ery­one to pump as much as they could.

The first change oc­curred in Ray­mond Basin. After the at­tempt to reach a vol­un­tary set­tle­ment failed, city of Pasadena ini­ti­ated le­gal pro­ceed­ings against city of Alham­bra and 30 other pro­duc­ers.

That changed the dy­nam­ics of the sys­tem. It was found out that the wa­ter is be­ing pumped out at a rate ex­ceed­ing the re­plen­ish­ment rate by 38%. The rul­ing would, pre­sum­ably, re­dis­tribute the wa­ter so that it was pumped out at at most the re­plen­ish­ment rate. How­ever, due to com­plex le­gal and own­er­ship ar­range­ments it was not at all clear what the rul­ing is go­ing to be and who’s go­ing to take the worst hit. Every par­ti­ci­pant had to con­sider the sce­nario where they would be the ul­ti­mate loser. That pro­vided an in­cen­tive to try to reach a ne­go­ti­ated set­tle­ment.

Within six months the par­ties had drafted a stipu­lated agree­ment to share the cut­back on pro­por­tional ba­sis signed by all but 2 of the 32 par­ti­ci­pants of the liti­ga­tion. Rather than im­pos­ing his own solu­tion the judge is­sued a fi­nal judge­ment based on the stipu­lated agree­ment.

West Basin came next. Un­like in the case of small Ray­mond Basin, the liti­ga­tion had al­most 500 par­ties. Also, when the num­bers came in it turned out that pump­ing rate was three times the nat­u­ral re­plen­ish­ment rate of the basin. Thus, each par­ti­ci­pant had to face dras­tic cut­backs if the case was set­tled by the judge. A fo­rum was cre­ated for ne­go­ti­a­tion of the set­tle­ment. Although ev­ery­one had an in­cen­tive to agree on any­thing bet­ter than two thirds cut­back, it took two years of ne­go­ti­a­tions and a threat of court to achieve an in­terim agree­ment to cut back the pump­ing back to the lev­els from the year 1949. The in­terim agree­ment was used for seven years while the pro­duc­ers pur­sued other strate­gies to en­hance the lo­cal wa­ter sup­plies, to re­plen­ish the basin, and to try to con­vince non-sig­na­to­ries to agree to the cur­tail­ment.

In 1961, af­ter 16 years a trial was held and pro­posed judge­ment was passed to the court.

Later on similar pro­cess hap­pened for Cen­tral Basin.

Now here’s a quote that I find in­ter­est­ing:

No one re­ally knows the ex­act costs in­volved in the West Basin liti­ga­tion, given the large num­ber of par­ties and the length of time in­volved, but the best available es­ti­mate is $3 mil­lion. [...] Amor­tiz­ing the costs of the liti­ga­tion over a 50-year pe­riod (as one would ex­pect to do for the con­struc­tion of a ma­jor phys­i­cal fa­cil­ity), [...] the ad­ju­di­ca­tion costs in West Basin amounted to an an­nu­al­ized cost of $2.50 per acre-foot of wa­ter rights.

I like the idea of treat­ing the cre­ation of an in­sti­tu­tion (i.e. the rules gov­ern­ing the us­age of the basin) as an in­fras­truc­ture pro­ject, similar to a dam. If you read the piece by Scott Alexan­der you’ll get an im­pres­sion that you have to at least sac­ri­fice a black goat and hire an ex­or­cist to solve a co­or­di­na­tion prob­lem. After read­ing Ostrom though, it doesn’t look any more ex­otic than hiring a cou­ple of lawyers and an ac­coun­tant.

It should be also taken into ac­count that con­struc­tion of an in­sti­tu­tion is an in­for­ma­tion good of a kind and thus the cost can be re­duced by hav­ing some prior in­for­ma­tion. For ex­am­ple, the liti­ga­tion pro­cess in Cen­tral Basin hap­pened af­ter the liti­ga­tion in West Basin and the par­ti­ci­pants were thus able to learn from the ex­ist­ing ex­pe­rience, elimi­nate un­nec­es­sary steps (for ex­am­ple, it was im­me­di­ately clear that they are shoot­ing for a ne­go­ti­ated set­tle­ment, not a court rul­ing) and the cost of the pro­ject, de­spite Cen­tral Basin be­ing big­ger and hav­ing more stake­hold­ers, were es­ti­mated at mere $450,000.

Similar pro­ject had, how­ever, failed in San Bernardino county. Ostrom lists sev­eral rea­sons for the failure, in­clud­ing that San Bernardino county is much big­ger that the other bas­ins, that it may ac­tu­ally con­sists of sev­eral phys­i­cal bas­ins and so on. But listen to this:

No vol­un­tary wa­ter as­so­ci­a­tions were cre­ated to fa­cil­i­tate dis­cus­sion of these is­sues, and no con­sen­sus emerged over time about any of them. Con­flicts emerged be­tween the large and small wa­ter pumpers, be­tween ad­vo­cates for de­vel­op­ment and ad­vo­cates for no-growth poli­cies, be­tween in­dus­try and agri­cul­ture, be­tween lo­cals and “ex­ter­nal ex­perts”, and be­tween ap­pointed per­son­nel and elected offi­cials. The lack of fun­da­men­tal agree­ment led to ac­rimo­nious poli­ti­cal con­flict, in­clud­ing sev­eral re­call elec­tions, front-page sto­ries in the lo­cal pa­pers that pushed aside sto­ries on the Water­gate scan­dal, and fi­nally the sus­pen­sion of the liti­ga­tion in 1974. No ac­tion has since been taken to limit ground­wa­ter pump­ing.

It seems to me that the real cause of the failure was that the model from West Basin was im­posed on the par­ti­ci­pants, in top-down man­ner, with­out much dis­cus­sion. The fora that ex­isted in West Basin weren’t cre­ated in San Bernardino. The re­sult was one ugly co­or­di­na­tion failure.

Which makes me won­der whether the pro­cess of cre­ation of an in­sti­tu­tion may be in­stru­men­tal in its even­tual suc­cess. What if the same set of rules in the same cir­cum­stances may suc­ceed or fail de­pend­ing on how they were con­ceived? What if the pro­cess of cre­ation, the one where peo­ple dis­cuss their op­tions, ar­gue about them and bounce ideas off one an­other serves as a mechanism to es­tab­lish com­mon knowl­edge among them? That’s a re­ally in­ter­est­ing idea and I would love to see some ex­per­i­men­tal re­sults to ei­ther con­firm it or dis­prove it.

Can this be repli­cated el­se­where?

Ob­vi­ously, Cal­ifor­nian wa­ter pro­duc­ers had many ad­van­tages on their side. They lived in a demo­cratic coun­try. They had func­tional le­gal sys­tem and law en­force­ment. Cor­rup­tion was low. The level of so­cial trust was high.

Would they be able to suc­ceed el­se­where? It’s hard to say, ob­vi­ously, but let’s have a look at what Ostrom writes about Sri Lankan ir­ri­ga­tion sys­tem.

She dis­cusses the topic in the chap­ter de­voted to co­or­di­na­tion failures. She de­scribes the situ­a­tion in Kirindi Oya. The pop­u­la­tion was het­eroge­nous, com­posed of in­di­vi­d­u­als com­ing from differ­ent re­gions, castes and kin­ship groups, all of whom are ini­tially poor. The up­stream farm­ers were us­ing more wa­ter than they needed (be­cause hav­ing rice fields flooded helps with weed con­trol) and left lit­tle, if any, for down­stream farm­ers. The cen­tral regime was un­will­ing to en­force rules im­par­tially. Po­lice treated wa­ter offences as triv­ial. Mon­i­tor­ing was non-ex­is­tent and dis­putes were some­times solved by vi­o­lent means. The in­fras­truc­ture was dam­aged:

Gates are miss­ing, struc­tures dam­aged, chan­nels tapped by en­croach­ers and oth­ers. When asked why they don’t pre­vent some of the most blatant offenses, two young tech­ni­cal as­sis­tants replied “that they were afraid to be­cause of the fear of be­ing as­saulted”.

In short, the situ­a­tion was as bad as it gets.

But then she re­counts a pos­i­tive story from the left bank of Gal Oya ir­ri­ga­tion pro­ject. The situ­a­tion was similar to that in Kirindi Oya. The fact that up­stream farm­ers were mostly Sin­halese while down­stream cul­ti­va­tors were mostly Tamil haven’t made things bet­ter.

Co­op­er­a­tion among farm­ers was min­i­mal. So­cial re­la­tions among set­tlers, who came from differ­ent ar­eas of the coun­try were strained… Re­la­tions be­tween farm­ers and Ir­ri­ga­tion Depart­ment (ID) offi­cials were marked by mis­trust and re­crim­i­na­tions. Farm­ers had no con­fi­dence in the com­pe­tence or the trust­wor­thi­ness of the ID’s staff… Many field-level offi­cials … were no­to­ri­ous for their cor­rup­tion and thug­gery. The main ob­sta­cle to effi­cient wa­ter man­age­ment, from the farm­ers’ point of view, was the lo­cal-level offi­cials, who had poli­ti­cal and bu­reau­cratic power be­hind them. On the other hand, the ID offi­cials, es­pe­cially ir­ri­ga­tion en­g­ineers, be­lieved that farm­ers could not use wa­ter re­spon­si­bly and care­fully. There­fore, they ar­gued that it was nec­es­sary to or­ga­nize, ed­u­cate, and dis­ci­pline the farm­ers to do what ID asked them to do. Thus farm­ers were con­sid­ered a part of the prob­lem while the lat­ter con­sti­tute the solu­tion.

Then an ex­per­i­ment was made. The idea was to in­tro­duce “cat­a­lysts” into the situ­a­tion of mu­tual mis­trust and un­pre­dictabil­ity. “In­sti­tu­tional or­ga­niz­ers,” (IO) mostly col­lege grad­u­ates who also had farm back­grounds were dis­patched to the area. They’ve re­ceived a six week train­ing on how to ap­proach and mo­ti­vate farm­ers and on tech­ni­cal sub­jects re­lated to ir­ri­ga­tion. Each went to a small area served by one dis­trib­u­tory canal. Their pur­pose was not to im­pose a par­tic­u­lar policy but rather to or­ga­nize the farm­ers to plan self-help strate­gies. At the same time they had sta­tus to deal effec­tively with Ir­ri­ga­tion Depart­ment offi­cials.

In­stead of es­tab­lish­ing a pre­defined or­ga­ni­za­tion, the IO tried to form a work­ing com­mit­tee to solve par­tic­u­lar prob­lems, such as re­pairing a bro­ken gate or de­silt­ing a field chan­nel. Fur­ther, IOs iden­ti­fied prob­lems be­yond those that could be solved by lo­cal farm­ers work­ing to­gether, prob­lems that had to be ar­tic­u­lated to ID offi­cials and oth­ers. Once farm­ers were used to work­ing to­gether and had achieved benefits from group ac­tion, the IO would then help form a lo­cal or­ga­ni­za­tion and se­lect, through con­sen­sus, a farmer-rep­re­sen­ta­tive. This rep­re­sen­ta­tive could ar­tic­u­late the in­ter­ests of the other farm­ers on his field chan­nel at larger meet­ings and re­port back to the oth­ers what had hap­pened in larger are­nas.

When the farm­ers started work­ing on re­ha­bil­i­ta­tion of the field chan­nel the at­ti­tude of ir­ri­ga­tion offi­cials to­ward them started to change.

In the ar­eas where the new sys­tem was in­tro­duced farm­ers started to use wa­ter ro­ta­tion pro­ce­dures quite gen­er­ally. There were even de­liber­ate efforts to make wa­ter available to the farm­ers down­stream.

On the Sin­halese-Tamil bound­ary, for ex­am­ple, main­te­nance haven’t been done in years. Farm­ers talked about pre­vi­ous mur­ders over wa­ter dis­putes. Within few months af­ter in­tro­duc­ing the new sys­tem, Sin­halese and Tamil farm­ers be­gan to work on clear­ing out the chan­nels.

That’s not to say there were no prob­lems in Gal Oya pro­ject. How­ever, the ex­am­ple shows that there are ways to get out of sub­op­ti­mal equil­ibria even in a highly dam­aged and non-func­tional en­vi­ron­ment. Some­times, though, it seems to re­quire a lit­tle nudge from the out­side.


Com­mon pool re­source prob­lems are a strict sub­set of in­ad­e­quate equil­ibria prob­lems. How­ever, I don’t see why they would be in­her­ently eas­ier to solve than other types of prob­lems. Maybe there’s a vi­able solu­tion for each such prob­lem. Maybe there’s a way to es­cape any sub­op­ti­mal Nash equil­ibrium. Maybe all we have to do is to try and when we fail to try again.