Focus of the CTN

The re­search ac­tiv­ity of the Coali­tion The­ory Net­work is cen­tred on the the­o­ret­i­cal and em­pir­i­cal analy­sis of socio-​economic net­works and groups. This per­tains to dif­fer­ent as­pects of so­cial and eco­nomic sys­tems, among which the man­age­ment of in­ter­na­tional pub­lic goods, the gov­er­nance of eco­nomic unions, the for­ma­tion of in­dus­trial car­tels and col­lab­o­ra­tions, the pat­terns of racial in­te­gra­tion in so­cial net­works, and the en­doge­nous evo­lu­tion and struc­ture of in­sti­tu­tions, etc.

There is now a broad con­sen­sus within the in­ter­na­tional sci­en­tific com­mu­nity that such cru­cial phe­nom­ena call for a uni­fied the­o­ret­i­cal ap­proach, based on the de­vel­op­ment of be­hav­ioural mod­els and equi­lib­rium con­cepts, and for an ex­ten­sive in­vest­ment in em­pir­i­cal re­search. The last two decades have wit­nessed a huge aca­d­e­mic com­mit­ment in this di­rec­tion, es­pe­cially within the dis­ci­plines of eco­nom­ics, game the­ory and be­hav­ioural sci­ences in gen­eral. 

More­over, a grow­ing in­ter­est in so­cial and eco­nomic net­works has been fos­tered by re­cent de­vel­op­ments in the analy­sis of strate­gic net­work for­ma­tion, and by an in­creas­ing in­tense in­te­gra­tion be­tween this new body of re­search and ear­lier work on the for­ma­tion of large net­works, de­vel­oped in the me­chan­i­cal physics lit­er­a­ture.

A re­view of the state of the art

The the­ory of coali­tion for­ma­tion fo­cuses on the analy­sis of the in­cen­tives of eco­nomic agents to com­mu­ni­cate and on how these in­cen­tives act on the de­ter­mi­na­tion of equi­lib­rium con­fig­u­ra­tions of so­cial and eco­nomic groups. It has de­vel­oped as the nat­ural ex­ten­sion of tra­di­tional co­op­er­a­tive game the­ory, which fo­cused mainly on the prob­lem of how the sur­plus of co­op­er­a­tion "should" or "will" be shared among co­op­er­at­ing agents, while the the­ory of coali­tion for­ma­tion has ex­tended the set of ques­tions in two prin­ci­pal ways. First, it in­ves­ti­gates how groups form, and what con­fig­u­ra­tion of groups will emerge in equi­lib­rium. The for­ma­tion of mul­ti­ple groups is con­sis­tent with the re­cent ex­pe­ri­ence of in­ter­na­tional ne­go­ti­a­tions on en­vi­ron­men­tal stan­dards and on trade issue, where "small" coali­tions are the rule rather than the ex­cep­tion, even when full co­op­er­a­tion would be de­sir­able. A sec­ond im­por­tant ex­ten­sion of the set of ques­tions ad­dressed by coali­tion the­ory is the study of prob­lems in which the wel­fare of a group can be af­fected by the ac­tions taken by agents who are mem­bers of other groups, that is, when there are ex­ter­nal­i­ties among groups. Such ex­ter­nal­i­ties are typ­i­cal, for ex­am­ple, of car­tel for­ma­tion in oli­gop­o­lies, co­op­er­a­tion on en­vi­ron­men­tal poli­cies, and eco­nomic unions.

In order to ac­count for these em­pir­i­cal reg­u­lar­i­ties, the­o­rists have taken new ap­proaches to the prob­lem, which can be clas­si­fied as fol­lows.

For­mu­la­tion of new game-​theory mod­els of coali­tion for­ma­tion and net­works. In these mod­els, the choice of form­ing a coali­tion is mod­elled as a strate­gic vari­able of each player emerges as the out­come of purely non-​cooperative be­hav­iour. We can fur­ther dis­tin­guish be­tween sta­tic mod­els (e.g., Hart and Kurz (1983)) and dy­namic mod­els of coali­tion for­ma­tion (e.g., Chat­ter­jee, Dutta, Ray and Sen­gupta (1993), Bloch (1996)).

For­mu­la­tion of new so­lu­tion con­cepts. Here, the ef­fort is aimed at iden­ti­fy­ing equi­lib­rium coali­tion struc­tures that sat­isfy con­sis­tency re­quire­ments both within and out­side form­ing coali­tions (e.g., Ray and Vohra (1997)). Other new so­lu­tion con­cepts ad­dress sit­u­a­tions where agents are far-​sighted (eg. Dutta, Ghosal and Ray (2005), Page, Wood­ers and Kamat (2004)). There is also a need for dy­namic mod­els and so­lu­tion con­cepts that ad­e­quately cap­ture the most salient fea­tures of real-​world sit­u­a­tions.

The in­ter­est of eco­nom­ics and other be­hav­ioural sci­ences in net­work the­ory is more re­cent, and has mainly de­vel­oped since the sem­i­nal works of My­er­son (1991) and of Jack­son and Wolin­sky (1996). This in­ter­est is mo­ti­vated by the num­ber of eco­nomic phe­nom­ena that are af­fected, in one way or an­other, by the com­mu­ni­ca­tion net­work in which agents are em­bed­ded. Job con­tact net­works and trade net­works are only two of the many ex­am­ples that may be cited. What is im­por­tant here is that both the ag­gre­gate wel­fare and its dis­tri­b­u­tion seem to de­pend on how in­for­ma­tion flows on the net­work, thus pro­vid­ing a clearer un­der­stand­ing of how net­works form a key issue in eco­nom­ics.

Mod­els of net­work for­ma­tion have been pre­vi­ously de­vel­oped within me­chan­i­cal physics lit­er­a­ture, in ad­dress­ing the issue of how large net­works form as the out­come of purely me­chan­i­cal ran­dom processes. More re­cent eco­nom­ics lit­er­a­ture fo­cuses in­stead on the in­cen­tives of ra­tio­nal agents to es­tab­lish links and form net­works as the out­come of op­ti­mal be­hav­iour, and on the wel­fare im­pli­ca­tion of such be­hav­iour. There is now a con­sen­sus within the sci­en­tific com­mu­nity that a bridge be­tween these two lit­er­a­tures needs to be built, in order to study in­cen­tives and wel­fare within com­plex ar­chi­tec­tures of the type en­coun­tered in so­cial net­works, where ran­dom­ness clearly plays a role. This ten­dency is con­firmed by the large num­ber of new con­tri­bu­tions on strate­gic mod­els of net­work for­ma­tion that bor­row tools and con­cepts from me­chan­i­cal physics lit­er­a­ture.

Re­cent re­search on net­works has had a strong em­pir­i­cal com­po­nent, par­tic­u­larly in top­ics re­lated to labour eco­nom­ics and con­sumer be­hav­iour. Peer and neigh­bour­hood ef­fects on youth have been stud­ied by Sac­er­dote (2001) and Katz et al. (2001), and so­cial in­ter­ac­tions and crime are the focus of Glaeser (1996) and Kling et al. (2005). The pos­si­bil­ity of so­cial in­ter­ac­tions in labour sup­ply is dis­cussed by Grod­ner and Knies­ner (2006), Wein­berg et al. (2004), and Woit­tiez and Kapteyn (1998), and the im­pact of cus­tom on con­tract de­sign is an­a­lyzed by Young and Burke (2001). The econo­met­ric prob­lems in iden­ti­fy­ing net­work ef­fects are com­plex, and dis­cussed in Brock and Durlauf (2001), Man­ski (1993), and Mof­fitt (2001). Glaeser and Scheinkman (2002) pro­vide an overview of some of this lit­er­a­ture, which is clearly in its in­fancy but ripe for fu­ture re­search. An area of par­tic­u­lar promise is the study of the evo­lu­tion and con­se­quences of new in­sti­tu­tional arrange­ments in the tran­si­tion economies.

Ad­vanc­ing the state of the art

The­o­ret­i­cal and em­pir­i­cal re­search on net­works and groups is mov­ing in sev­eral di­rec­tions. Some of these are nat­ural de­vel­op­ments of the ex­ist­ing state of the art, such as the for­mu­la­tion of more gen­eral mod­els and more sat­is­fac­tory so­lu­tion con­cepts that would ob­tain co­op­er­a­tive out­comes as the re­sult of in­di­vid­ual equi­lib­rium ac­tions and de­ci­sions. Other lines of re­search re­spond to the emer­gence of new and un­prece­dented so­cial, eco­nomic and po­lit­i­cal frame­works at both na­tional and in­ter­na­tional lev­els, which raise new the­o­ret­i­cal and pol­icy is­sues. From the the­o­ret­i­cal stand­point, net­works and net­work for­ma­tion pose the most new and chal­leng­ing is­sues for re­searchers. So­cial agents are in many ways or­gan­ised in a net­work of re­la­tion­ships, tak­ing the forms of friend­ship ties, pro­fes­sional re­la­tions, strate­gic in­ter­ac­tion, phys­i­cal con­nec­tions, etc… Net­work the­ory needs to be able to pro­vide a uni­fied frame­work for analysing the re­la­tion be­tween agents’ po­si­tion in the net­work and their ac­tions and wel­fare. Even more gen­er­ally, a model is needed to ex­plain how the whole struc­ture of the net­work (or the be­liefs that agents hold in this struc­ture) af­fect agents’ be­hav­iour and wel­fare. The study of net­work for­ma­tion and of games played on net­works under local and lim­ited in­for­ma­tion is in­deed one of the most chal­leng­ing and fre­quently stud­ied is­sues at present. 

An­other im­por­tant the­o­ret­i­cal prob­lem is how to model the for­ma­tion of large net­works and, in par­tic­u­lar, how to in­te­grate ran­dom net­work for­ma­tion mod­els with strate­gic mod­els in which agents form links ra­tio­nally. A bridge be­tween these two strands of mod­el­ling is needed,in order to make pre­dic­tions on ac­tual so­cial net­works, and to say some­thing about in­cen­tives within such struc­tures. 

The lack of data on so­cial net­works also sug­gests that ex­per­i­men­tal work in this area will play an im­por­tant role in em­pir­i­cal re­search. Ex­per­i­men­tal de­sign in this spe­cific topic is there­fore an­other main di­rec­tion in which re­search will make a great ef­fort. This is true also for the prob­lem of coali­tion for­ma­tion, for which a larger and more es­tab­lished body of the­o­ret­i­cal work ex­ists.

Many in­ter­est­ing ques­tions arise from more ap­plied prob­lems, in which the gen­eral mod­els of coali­tion and net­work for­ma­tion are able to pro­vide more spe­cific pre­dic­tions. For in­stance, a re­cent suc­cess­ful ap­proach to group for­ma­tion is the de­sign of pro­ce­dures to match agents to­gether in real-​life prob­lems such as aca­d­e­mic job mar­kets, ad­mis­sion of stu­dents in schools or uni­ver­si­ties (e.g., Roth (1984, 2004)). Re­cent ap­pli­ca­tions of match­ing the­ory raise new is­sues such as the prob­lem of strate­gic be­hav­iour and asym­met­ric in­for­ma­tion in large mar­kets. Also, the de­sign of ef­fi­cient and sta­ble vot­ing rules can be fruit­fully ap­plied to EU decision-​making prob­lems. More in gen­eral, the de­sign of sta­ble and ef­fi­cient self-​enforcing in­sti­tu­tions seems to be among the main ob­jec­tives of coali­tion the­ory, es­pe­cially in an evolv­ing world con­text in which po­lit­i­cal equi­lib­ria are still to be found. In this re­spect, a fruit­ful area of re­search is the in­ves­ti­ga­tion of how do­mes­tic po­lit­i­cal in­cen­tives and in­sti­tu­tions af­fect the be­hav­iour of coun­tries at the in­ter­na­tional level, and thereby the out­come of treaties and co­op­er­a­tion. Here the the­ory of coali­tion for­ma­tion can be fruit­fully in­te­grated with the the­ory of net­works, the lat­ter study­ing the for­ma­tion of bi­lat­eral in­ter­na­tional re­la­tions, while the for­mer study­ing the sta­bil­ity of eco­nomic and po­lit­i­cal unions.