To un­der­stand fi­nance, em­brace com­plex­ity

The Pak Banker - - OPINION - Mark Buchanan

Ahighly un­usual col­lab­o­ra­tion be­tween econ­o­mists and sci­en­tists of­fers an im­por­tant in­sight for those who want to fix the world's cri­sis­prone fi­nan­cial sys­tem: There's no sim­ple way to un­der­stand a com­plex net­work.

This month's is­sue of the re­search jour­nal Na­ture Physics features a hand­ful of pa­pers in which physi­cists, other nat­u­ral sci­en­tists and lead­ing ex­perts in eco­nom­ics and fi­nance -- in­clud­ing prom­i­nent bank­ing reg­u­la­tors and No­bel Prize-win­ning econ­o­mist Joseph Stiglitz -- put their minds to­gether to fig­ure out fi­nance. What the sci­en­tists bring to the ta­ble is ex­pe­ri­ence in study­ing net­works, be­wil­der­ing tan­gles of in­ter­linked and in­ter­de­pen­dent things such as an eco­log­i­cal food web or the In­ter­net.

Take a look at any di­a­gram show­ing the in­ter­con­nec­tions among the world's banks and other fi­nan­cial in­sti­tu­tions -- links es­tab­lished through or­di­nary loans, but far more ex­ten­sively through fi­nan­cial de­riv­a­tives -- and what you see will be very com­plex. We barely un­der­stand how such com­plex­ity changes the way net­works op­er­ate. What we do know sug­gests we should worry when there's too much of it.

Fifty years ago, ecol­o­gists in­ter­ested in the sta­bil­ity of food webs at first mis­tak­enly con­cluded that more com­plex­ity -- more species and a greater den­sity of links among them -- would tend to make an ecosys­tem more sta­ble. This turned out to be wrong. Later work by noted ecol­o­gist Robert May demon­strated that while healthy eco­log­i­cal net­works are rich and di­verse, too much com­plex­ity tends to make them un­sta­ble and prone to col­lapse. Loosely speak­ing, net­works with too much com­plex­ity can go wrong in too many ways.

Sound fa­mil­iar? The com­plex­ity of the fi­nan­cial sys­tem ex­ploded over the past few decades, pri­mar­ily through the pro­lif­er­a­tion of de­riv­a­tives of all kinds. Then the sys­tem it­self blew up. What the pa­pers in Na­ture Physics ar­gue is that any really deep un­der­stand­ing must fo­cus on the de­tailed pat­tern of links among in­sti­tu­tions, or the net­work "topol­ogy." (Full dis­clo- sure: I write a monthly col­umn for Na­ture Physics.)

I've touched on net­works sev­eral times in ear­lier Bloomberg View col­umns. It's worth em­pha­siz­ing again how the most ba­sic in­sights emerg­ing from this new line of work run con­trary to re­ceived wis­dom from eco­nom­ics. For ex­am­ple, banks that cre­ate and sell de­riv­a­tives of­ten ar­gue that their pro­lif­er­a­tion is ben­e­fi­cial, as it makes mar­kets more "com­plete," mean­ing that it brings us closer to a world in which es­sen­tially any kind of trade or bet can be un­der­taken at any time. Stan­dard fi­nan­cial the­ory as­sumes that such com­plete­ness is as­so­ci­ated with greater fi­nan­cial sta­bil­ity, as it al­lows any­one with in­for­ma­tion to bring it into the mar­ket­place. How can that be bad?

Well, achiev­ing com­plete­ness en­tails a vast in­crease in com­plex­ity, with con­se­quences that tra­di­tional fi­nance models fail to cap­ture. Th­ese models sup­pose that the ac­tions of in­di­vid­u­als or firms in the mar­ket are too small to af­fect its be­hav­ior in any se­ri­ous way, much as we used to think that our fish­ing the oceans would have min­i­mal in­flu­ence on the abun­dance of fish. But a net­work study from sev­eral years ago demon­strated that the seem­ingly tiny in­flu­ence that trad­ing has on the mar­ket be­comes in­creas­ingly sig­nif­i­cant as the num­ber and com­plex­ity of fi­nan­cial in­stru­ments in­creases. The generic re­sult is vi­o­lent mar­ket fluc­tu­a­tions and in­sta­bil­ity. Fi­nan­cial in­sti­tu­tions still find ways to profit by cre­at­ing and sell­ing new de­riv­a­tives, even if th­ese de­liver no ben­e­fits to the mar­ket and ac­tu­ally drive the sys­tem to­ward trou­ble. blog for fur­ther de­tails.)

Com­plex­ity also helps fi­nan­cial in­sti­tu­tions hide the risks they cre­ate. De­spite the ad­ver­tis­ing of the In­ter­na­tional Swaps and De­riv­a­tives As­so­ci­a­tion and oth­ers who cre­ate and sell de­riv­a­tives, th­ese prod­ucts are only some­times used for hedg­ing and much more fre­quently for spec­u­la­tion. In the lat­ter case, they are ex­ceed­ingly use­ful in ob­scur­ing in­for­ma­tion that would be cru­cial to the proper judg­ment of val­ues and risks. Con­sider the de­riv­a­tives that helped Italy's Banca Monte dei Paschi di Siena SpA hide hun­dreds of mil­lions of dol­lars in losses as it sought a tax­payer bailout. Any­one mak­ing deals with a bank en­meshed in a largely in­vis­i­ble web of con­tracts with far-flung counter-

(See my par­ties does so with a very in­com­plete view of the risks in­volved.

The ba­sic com­plex­ity of the mar­ket al­lows for the com­ple­tion of deals that would never get signed in a world of full trans­parency and un­der­stand­ing. Un­for­tu­nately, tra­di­tional fi­nan­cial the­ory -- which as­sumes that in­di­vid­ual ac­tors have per­fect knowl­edge and make only ra­tio­nal de­ci­sions -- ig­nores this point, blind­ing it­self to a huge source of sys­temic risk. Any sci­ence has to be­gin with ba­sic in­sights first, learn­ing which de­tails really mat­ter and which may not.

Newspapers in English

Newspapers from Pakistan

© PressReader. All rights reserved.