Eco­nom­ics be­comes an en­force­ment weapon

Business Day - Business Law and Tax Review - - BUSINESS LAW & TAX REVIEW - Scott Have­mann & Lizel Blig­naut

The South African com­pe­ti­tion au­thor­i­ties are ex­pected to use all tools avail­able to them, in­clud­ing eco­nomic anal­y­sis, to de­tect car­tel be­hav­iour

PRICE fix­ing, mar­ket di­vi­sion or bid-rig­ging agree­ments among com­peti­tors or car­tels are of­ten cited by the leg­is­la­ture, com­pe­ti­tion au­thor­i­ties, courts, jour­nal­ists and con­sumers as the most egre­gious form of an­ti­com­pet­i­tive con­duct. In SA, the gov­ern­ment, pol­i­cy­mak­ers and the com­pe­ti­tion au­thor­i­ties have demon­strated their com­mit­ment to the erad­i­ca­tion of car­tels.

Car­tels are by their na­ture se­cre­tive and dif­fi­cult to de­tect. Out­comes in mar­kets (uni­form prices) can as eas­ily be as­cribed to com­pet­i­tive pro­cesses as to car­tel agree­ments. Com­pe­ti­tion au­thor­i­ties have found that a most use­ful tool for the de­tec­tion of car­tels is the im­ple­men­ta­tion of a le­niency pro­gramme whereby the le­niency ap­pli­cant (whistle­blower) is granted im­mu­nity from pros­e­cu­tion if it brings to the com­pe­ti­tion au­thor­i­ties ad­e­quate ev­i­dence of car­tel be­hav­iour that al­lows the com­pe­ti­tion au­thor­ity to pros­e­cute the re­main­ing car­tel mem­bers. The South African Com­pe­ti­tion Com­mis­sion’s Cor­po­rate Le­niency Pol­icy is con­sid­ered one of its suc­cess sto­ries, with the com­mis­sion hav­ing re­ceived hun­dreds of le­niency ap­pli­ca­tions since the pol­icy’s in­cep­tion.

How­ever, le­niency pro­grammes are suc­cess­ful in de­tect­ing and prose­cut­ing hard core car­tels only where a car­tel mem­ber or mem­bers per­ceive a real threat of de­tec­tion and are will­ing to break out of the close knit car­tel. Com­pe­ti­tion au­thor­i­ties there­fore have to rely on other tools to de­tect tacit col­lu­sion or close knit hard core car­tels.

De­vel­op­ments in the field of the eco­nom­ics of in­dus­trial or­gan­i­sa­tion and econo­met­ric tech­niques have equipped econ­o­mists with tools to de­duce whether firms are in­volved in price fix­ing, mar­ket di­vi­sion or bid rig­ging ar­range­ments when ex­plicit ev­i­dence is not avail­able. These tools can be used to de­duce the con­verse.

Where car­tel-bust­ing ex­er­cises were his­tor­i­cally viewed as fac­tual and le­gal en­quiries, com­pe­ti­tion au­thor­i­ties are now more than ever be­fore us­ing eco­nomic ap­proaches and econo­met­ric tech­niques to de­tect car­tels. By us­ing eco­nomic data (amongst other things, such as price, vol­ume, mar­ket shares, de­mand shifters, cost shifters and the like) and ap­ply­ing econo­met­ric tech­niques, econ­o­mists can dis­tin­guish be­tween tacit col­lu­sion and com­pe­ti­tion, as well as be­ing able to iden­tify episodes where a firm is in­volved in ex­plicit col­lu­sion. As the tech­niques can also be used to demon­strate com­pet­i­tive out­comes, they can be used to ac­tively dis­miss claims of col­lu­sion (if sup­ported by the eco­nomic data).

Es­sen­tially, the use of eco­nom­ics in car­tel pros­e­cu­tion re­quires a test­ing whether the eco­nomic out­comes (as pre­sented by the data) sup­port the eco­nomic model and out­come of com­pe­ti­tion or col­lu­sion (whether tacit or out­right).

Eco­nomic tech­niques used to find out whether car­tels are ac­tive nor­mally run through a process of first iden­ti­fy­ing the in­dus­try which may ex­hibit anti-com­pet­i­tive char­ac­ter­is­tics (ei­ther through iden­ti­fy­ing ab­nor­mal­i­ties in pric­ing or au­thor­i­ties re­spond­ing to com­plaints of al­leged car­tels), ver­i­fy­ing that the in­dus­try may in fact be con­trolled by a car­tel (through analysing mar­ket data such as price, costs, de­mand and sup­ply shifters) and lastly, by for­mu­lat­ing an eco­nomic ar­gu­ment that can be used by the courts to ei­ther de­fend or bust firms in­volved in anti-com­pet­i­tive be­hav­iour.

Tech­niques used by econ­o­mists to an­a­lyse whether firms have been ac­tive, usu­ally start at mod­el­ling an in­dus­try which would be seen as “com­pet­i­tive” and us­ing this as a bench­mark to test whether there has been any de­vi­a­tion from the com­pet­i­tive land­scape over the years where sus­pi­cion of car­tel-like be­hav­iour has oc­curred. These mod­els tend to an­a­lyse whether a firm’s pric­ing model is cor­re­lated with de­mand and sup­ply shocks — or whether prices run in­de­pen­dently of fac­tors char­ac­ter­is­tic of the in­dus­try. Usu­ally, mod­els such as these are more equipped to deal with ho­moge­nous prod­ucts (prod­ucts that are sim­i­lar in char­ac­ter­is­tics) as op­posed to dif­fer­en­ti­ated prod­ucts.

How­ever, just as test­ing whether an in­dus­try has de­vi­ated from com­pe­ti­tion, an­other use­ful method of iden­ti­fy­ing col­lu­sion is through iden­ti­fy­ing whether there has been a no­tice­able struc­tural break in a firm’s be­hav­iour. This nor­mally oc­curs when car­tels have been formed, or as they have been bro­ken up (through firms cheating on one an­other in the car­tel). Here, pric­ing data in re­la­tion to costs can be used to iden­tify no­tice­able pat­terns in firms’ be­hav­iour or where prices have di­verged sig­nif­i­cantly from each other which may in­di­cate that co­or­di­na­tion has bro­ken down. Sim­i­larly, such an anal­y­sis is com­ple­mented if it is found that par­tic­i­pants have for­mu­lated trade as­so­ci­a­tions to be used to fa­cil­i­tate or to cover for car­tel meet­ings.

The above ap­proaches have re­cently been sup­ple­mented by econ­o­mists for­mu­lat­ing mod­els that specif­i­cally de­tail the struc­tural com­pet­i­tive and col­lu­sive mod­els of the in­dus­try — seem­ingly by es­ti­mat­ing what a firm’s be­hav­iour in re­la­tion to price would be un­der com­pe­ti­tion or un­der col­lu­sion. Test­ing these mod­els against the in­dus­try un­der in­ves­ti­ga­tion may dis­close whether firms are ac­tively in­volved in any col­lu­sive be­hav­iour and may pro­vide in­sight into which firms have been en­gag­ing in an­ti­com­pet­i­tive be­hav­iour.

Eco­nomic mod­els are also fre­quently used to show that ac­cused firms have not been in­volved in any of the al­leged anti-com­pet­i­tive be­hav­iour. Eco­nomic mod­el­ling can show that a firm’s pric­ing be­hav­iour is ra­tio­nal and com­pet­i­tively neu­tral in re­ac­tion to mar­ket dy­nam­ics, rather than be­hav­iour ex­pli­ca­ble by col­lu­sion. Whether de­fend­ing or prose­cut­ing, do­ing so through eco­nomic tech­niques is both data and time in­ten­sive and re­quires cor­rectly con­trol­ling for the many de­ter­mi­nants of a firm’s be­hav­iour.

While the eco­nomic ap­proaches such as those de­scribed above have been fre­quently used in the US and in Euro­pean ju­ris­dic­tions, South African case law re­mains rel­a­tively quiet on im­ple­men­ta­tion of screen­ing de­vices and eco­nomic tech­niques to ac­tively de­tect pros­e­cute and de­fend car­tels. To date, most (if not all) suc­cess­fully pros­e­cuted South African col­lu­sion cases have re­lied on ev­i­dence of col­lu­sion pro­vided to the com­pe­ti­tion au­thor­i­ties by a le­niency ap­pli­cant. The re­sults of an ex­plicit eco­nomic in­ves­ti­ga­tion have not fea­tured promi­nently in car­tel en­force­ment and pros­e­cu­tion.

SA, al­beit a young ju­ris­dic­tion, is con­sid­ered by its peers to be rel­a­tively so­phis­ti­cated in its use of eco­nom­ics as a com­pe­ti­tion en­force­ment tool. The South African au­thor­i­ties are ex­pected to use all tools avail­able to them and as such it is rea­son­able to ex­pect that, when le­niency ap­pli­ca­tions fail to un­earth car­tels, that these eco­nomic tools could be suc­cess­fully de­ployed to as­sist such de­tec­tion.

The field of com­pe­ti­tion eco­nom­ics and law is con­tin­u­ously con­verg­ing with the use of eco­nomic tech­niques be­ing in­creas­ingly re­quired in or­der to ren­der sound com­pe­ti­tion law com­pli­ance ad­vice to South African firms.

Scott Have­mann and Lizel Blig­naut are econ­o­mists in the com­pe­ti­tion law depart­ment at ENS.

Newspapers in English

Newspapers from South Africa

© PressReader. All rights reserved.