World Bank Discussion Papers

Institutional Adjustment and Adjusting to Institutions

Robert Klitgaard

The World Bank Washington, D. C.

Copyright © 1995 The International Bank for Reconstruction

and Development/THE

WORlD

BANK

1818 H Street, N.W. Washington, uc 20433, U.S.A. All rights reserved Manufactured in the United States of America First printing September 1995 Discussion Papers present results of country analysis or research that are circulated to encourage discu-sion and comment within the development community. To present these results with the least possible delay, the typescript of this paper has not been prepared in accordance with the procedures appropriate to formal printed texts, and the World Bank accepts no responsibility for errors. Some sources cited in rhis paper may be informal documents that are not readily available. The findings, interpretations, and conclusions expressed in this paper are entirely those of the author(s) and should not be attributed in any manner to the World Bank, to its affiliated organizations, or to members of its Board of Executive Directors or the countries they represent. The World Bank: does not guarantee the accuracy of the data included in this publication and accepts no responsibility whatsoever for any consequence of their use. The boundaries, colors, denominations, and other information shown on any map in this volume do not imply on the part of the World Bank Group any judgment on the legal status of any territory or the endorsement or acceptance of such boundaries. The material in this publication is copyrighted. Requests for permission to reproduce portions of it should be sent to the Office of the Publisher at the address shown in the copyright notice above. The World Bank encourages dissemination of its work and will normally give permission promptly and, when the reproduction is for noncommercial purposes, without asking a fee. Permission to copy portions for classroom use is granted through the Copyright Clearance Center, Inc., Suite 910,222 Rosewood Drive, Danvers, Massachusetts 01923, u.s..A.

The complete backlist of publications from the World Bank is shown in the annual Index if Publications, which contains an alphabetical title list (with full ordering information) and indexes of subjects, authors, and countries and regions. The latest edition is available free of charge from the Distribution Unit, Office of the Publisher, The World Bank, 1818 H Street, N.W, Washington, D.C. 20433, U.S.A, or from Publications, The Work' Bank, 66, avenue d'Iena, 75116 Paris, France. ISSN: 0259-210X

Robert Klirgaard, a professor of economics rt the University of Natal, Dalbridge, World Bank's Operations Evaluation Department.

South Africa, was a consultant to the

Library of Congress Cataloging-in-Publication

Data

Klitgaard, Robert E. Institutional adjustment and adjusting to institutions I Robert Khtgaard. p. cm.-(World Bank discussion papers, rSSN 0259-Z10X;

303) Includes bibliographical references (p. ). ISBN 0-8213-3450-6 1. Institutional economics. 2. Africa-Econonllc 3. Economic development. 1. Title. II. Series. HB99.5.K59 1995 338.9~.odc20

policy.

95-36053

crp

Contents
Foreword Abstract 1. Applying Institutional 2. Market Institutions 3. Nonmarket Institutions The Incentives Problem: The Case of The Gambia Overcoming Constraints on Incentive Problems 4. Indigenous Institutions Guidelines for Adjusting to Institutions
5. Institutional

v vi Economics to Development Problems
1

3
4

4 7 8
9

Adjustment and Adjusting to Institutions Healing the Aid Relationship

13

14
17

Appendices Table 3.1 Two Approaches to Incentive Reforms
Box

5

3.1 Designing an Experiment with Incentives: Outline for a Model Memorandum

6

iii

Foreword
This paper applies institutional economics to development practice. Experience shows that the qualitative insights of the new economics of information can yield practical guidance for addressing such classic problems as:
iii

markets that seem not to work well for poor people; government agencies that seem corrupt, inefficient, and overcentralized:

and finality threaten too many vested interests. The r;ilpaperoffers a selective, experimental or "learning ,.',i'byoing" approach in which government officials d '(define performance measures and set goals, and plan cooperatively how to meet them. Experiments that are limited in time and scope, and in which people with vested interests can help design the ways to measure success, give reforms a chance. The paper also outlines a framework for taking indigenous instituti~ns into account in development efforts. AdjUsting to indigenous institutions will have many different forms; sometimes it is appropriate, sometimes not. But systematic attention to them in program and project design, through frameworks such as the one proposed in this paper, will generate practical ideas we might not otherwise reach.

!II

IIl!I

how to "take account" of indigenous institutions in development interventions.

Drawing on experience as well as theory, the author outlines frameworks for analyzing institutions, both governmental and market, as a prelude to improving their functioning. !i.tThe heart of the paper deals with the crisis in f\ipublic sector incentives. It recognizes that at:tempts to craft once-for-all reforms-though often prescribed-usually fail, because their complexity

Robert Picciotto Director General Operations Evaluation

v

Abstract
Institutional economics can make a positive therapeutic approach for addressing the issue of contribution when applied to such classic develop- . corruption. The paper also advocates that developmerit problems as markets t1wtreguladyJaiLthe ment institutions no.~w2.f1ly.J14fu.stgQy.emm~flt agencies:1Jutaajust to them. This means giving the PO()fl.corrupt and inefficientgovemmentagencies, and how to take account of indigenous institutions. instituti6nSfes6i.ircesaridrespect, empowering Wen·functioning markets often depend on wellstakeholders, and working with them. Outside institutions can sometimes help government ftiQd~(5ntngstafes:lri'ordertoiiriprove li6Wt'leU govemmenHnsfiwtions are funchoning,ofie-m'ust officials address sensitive issues and make credible unCierstand ••here?sQl1s.forpoor·performance= I commitments. The World Bank can playa catalytic corruption.poor incentives, and,offen;overcenrole in this regard. The paper suggests that.the aid relationship is itself in need of adjustment, and that ttaHzat1on. r_h~paperpE?E()~~~~!"1,(lppr()a~h.for overcorningcOnsrrarntS·to ..ncentiveferi5rms~ahd i international development institutions may themselves benefit from the types of measures being .proVides amOd€{g~9riQmic6:4mewoi:kforE~solvproposed. ing the.incentivep:r9lJiem . .It.recommends a

vi

••.•••••••.••••••

~~

•••••.•.•...............

_

_

w

1.. Applying Institutional Development Problems

Economics to

Six years ago at the Annual World Bank Conference made contributions that are not generically differonDevelopment Economics, Brian Van Arkadie ent from the kinds of studies that might be made of (lamented the disconnected approach to the study of political or social institutions. 1:Jl<ilhistor-y,and iinstitutions in development and the lack of practical Ina~ag~l11~nt.di9?yncragrmatter,.is a common i Wead hihning"thr~ugh· these sttl(lies. a~S.§9fl§_and .. Pplif<ltiQns.! John Nellis commented a that the subject, though "profoundly important," Ii! "lacks an analytical method, a conceptual frame'\A thIr~group indudescritif2soK~c()fl()l1}ic:J)Wory ,ii work capable of rendering it coherent" and said the 'wh6ti~~.i!1.~~it-ii~-~risa~~m~?n!i.to,.a.ttack,pre.Y<liling paradigms, "Institutional economists," writes P "next and most crucial step ... is to specify precise operational methods- and tools by whichto il!1:prQve Warren J. SamuelS~--uaIso·conaud protests against both the established market economy and the performance in institutions.'? Today, looking over established body of orthodox theory which they a-dl1tef"ehfiritellectuallandscape, one may be more see as too closely tied to existing institutional optimistic. arrangements."! NQ doubttheJield,oriields,_arestilldiffuse. On Finally, therear~peoplejnterested .inwhat ~~e side we have theoreticaL_~S9~?~jsts such as might becalledindigenous economic institutions Bengt Holmstrom, Paul Milgram, John Roberts, ranging from loc<l;Lsy~temsofpf()perty rights and and Jean Tirole, some of whom--such as Pranab exchange through to cultural normsandcOrivenBardhan and Joseph Stiglitz-iA19XkQ!'ldevelop...!1"l~l1t.3 In variousvvaysth~Yctr~xeYQlt!t{oilizlng the tions unders!29.cli[lPa.:Etas soIuti9Il§JQ~<::onomic the0ly of the Hrnl., industrial organization, and the problems .. )fl'isk~sharin&.~?~~~llic.<ltiQI1 ... ( and studygfl1)~~rkefiristitutiOnsby'sysferriahcruly"""'" credible commitment.' inserting .. roblerns'Ofirrf~effecrahd··.asymirietrical1y p Given this diversity-and these four categories fail held informatron. Tfieiit("jQl$.<iI'e athematical, but m to encompass the richness of the work, for example, many of them produce models that ofterfresh, of Douglass North, Elinor Ostrom, and Oliver qualitative explanations for phenomena that are Williamson, and of the many superb institutional hard to explain using elementary economists who work at the Bank, such as Ed microeconomics-for example, share-cropping Campos, Brian Levy, Mary Shirley, among othersarrangements, wages in firms that are below it is not surprising that institutional economists may workers' marginal products early in their careers find themselves at odds. And this may also explain and above their marginal products later on, and a why development practitioners still judge that the variety of institutional mechanisms that trade off subject has little to offer of practical use. risk and incentives. On the other hand, quantitative testing has been limited, since key parameters of ,. the models are difficult to define and measure. ...... j;; In what follows, I summarize briefly some of my own forays intoinstitutionaleconomics, .which.have Another brand of institutional economics includes as.theiIQPjec!iyethej.mP.mY~m~ntQf12QH<:Yanalysis .!'l-nQ..rnan<igementin . d.eveloping.countri.es, empirical studies of@<irl<et~t:z:ldc:_ty,E~~~ndWilY the economic organiz§lliQnswork. Here poHtical ,. Drawing on the contributions of others, particularly the economists of the first camp mentioned above, I sclennsts·a:rid businessadffiinistration experts have
1

2
have tried to show that the qualitative insights of the new economics of information can yield practical leverage when applied to classic problems such as:
II!i

2. Comment on "The Role of Institutions in Development," pp. 177,179. 3. Two recent books contain many references and helpful bibliographical essays. Paul Milgram and John Roberts, Economics, Organization and Management (Englewood Cliffs, NJ: Prentice Hall, 1992), and J-J. Laffont and Jean Tirole, A Theory of Incentives in Procurement and Regulation (Cambridge: MIT Press, 1993). For applications to developing countries see The Economic Theory of Agrarian Institutions, ed. Pranab Bardhan (New York: Oxford University Press, 1989); The Economics of Rural Organization, ed. Karla Hoff, Avishay Braverman, and Joseph Stiglitz (New York: Oxford University Press, 1992); and J.Y. Lin andJ.B. Nugent, "Institutions and Economic Development," in Handbook of Development Economics, Vol. 3, ed. Jere R. Behrman and T.N. Srinivasan (Amsterdam: North-Holland, 1995). 4. Warren 1. Samuels, "Institutional Economics," in Companion to Contemporary Economic Thought, ed. David Greenaway, Michael Bleaney, and Ian M.T. Stewart (London
and New York: Reatledge, 1991), p, 106.

markets that seem systematically not to work well for poor people, including even highly competitive markets, such as those for agricultural products but also credit, labor, and land;
II!

governmentagencies that seem mired in inefficiency and corruption, with tendencies toward overcentralization: and how to "take account" of indigenous institutions. Notes
I. Brian Van Arkadie, "The Role of Institutions in Development," Proceedings of the World Bank Annual Conference Oil Development Economics 1989, Supplement to The World Bank Economic Review and The World Bank Research Observer, pp, 153-175.

!Ii!

5. For example, Economic Anthropology, ed, Stuart Plattner (Stanford, CA: Stanford University Press, 1989).

2. Market Institutions
':Ma~k~t~i~d~~;i~pi~~g~o~ntriescanheliberanied'" i in the sense of freeing prices and yet not work well iJw 12<?.2L£~£ple.Conventional economics teaches that there are various reasons \Vhyconlpetil:ive markets milY nof-yTdd Pareto 'optimal results, such as decreasing cost industi'lesl=md'monopoly, externalfties,publicg06ds.· The economics of .ID<?ralhazard, and incomplete contracting," These'" institutions can participate in and even facilitate economic modernization.) In this perspective, one emphasizes the importance for)na~Dg.IJli,1rkets Y"9rk~'.Qt~etter legal systems and th~llnCOrI"upted 9nq~fficierlf<;l~ministratiort of justice .. One stresses such informaticnalaspects as better weights and measures and, more generally, better institutions for gauging, certifying, processing, and diffusing information about the quality of products, people, risks, and services.

iriforrnanonaddsOth:eis;:Such.as ag~Y5:r~~ctection,
information is scarce, difficult to process, and asymmetrically held; where property rights are unclear; and where legal systems do a poor job of defining and enforcing commercially relevant rules, such as loan repayment and contracts-in other words, under conditions that characterize many developing countries, especially in rural areas.

last probleii1swiIllie'especlaHy"t,everewhen

Are there success stories we can learn from? And are there usable frameworks that might help us in practice analyze market institutions and suggest ways to improve them? The answers are yes and yes. Several examples and an analytical framework can be found in my book, Adjusting to Reality.1Two general points deserve emphasis. /Indigenous societies develop their own institutions :Iheiirst!,is that w~U~6JJ;!£tigrri!lgmar~~l:? ften o !Fto address problems of adverse selection, moral dep~!:<i9r:"~~11-furt~!11:?~illgst<ite~.For xample, e 1 hazard, incomplete cO!1tracts,a,nd the rest. Econoproperty rights and contract laws must be well mists have analyzedsherecroppjng as a rational specified and enforced-so must rules about the solution to the need to balanceincentives and risk repayment of debts. Since information has aspects in a principal-agent relationship, The prevalence of of a public good, states playa central role in sharing arrangements, indeed the relative strength helping markets overcomeproblems of informaof extended family and of clan, can be partially tion asymmetry . Thesecondl. point iSJh?t states are understood as a response to environments of great often ineffective in playing their roles of enforcer risk. Rural people in Africa may "save" in the form and inforlJ:la#oneI1hqD£~r:... t:ndeed, the private of cattle so as not to be subjected to harassment for sector, especially in developing countries, is less not sharing a more divisible resource such as grain constrained by states that intervene than by states or money. Rotating savings and credit associations that intervene ineffectively, corruptly, or not at all. may reflect a similar need to save in a way that one And so our attention turns to the question, what"]: is credibly able to refuse kin and dan who wish to can be done to make the institutions of govern- 1:: borrow. ment=-and more generally nonmarket institutions-work better?...:, Economic development can be understood as supplementing or even replacing these traditional Note institutions and practices with more modern institutions for dealing with risk, imperfect inforL Adjusting to Reality; Beyond "State vs, Market" in mation, credible commitment, and contract Economic Development (San Francisco: ICS Press and ment. (Below Iwill discuss how indigenous International Center for Economic Growth, 1991).
3

3e

Nonmarket Institutions
I have picked an issue that I believe is of central importance and also susceptible to new approaches by organizations such as the World ,i.jBank It concerns the crisis in public sector ';::incentives. Some "success stories" are outlined in Adjusting to Reality. In the first appendix to this paper, I provide a qualitative outline of a model of the incentives problem" which goes beyond the framework of Adjusting to Reality. For now, consider an example of current interest, one that is not yet a success story: public sector incentives in . The Gambia.

"Better states" does not only mean states that spend modestly or states that respect rights. Nor it only mean states that adopt sensible macroeconomic policies. "1?ett~r':i'lJs()~~~ns overcoming.chronic problemsofpoorjnc~ritiyes! systematic corruption, al1~oyercenj:r<lli?<ltion. CiProblems with publiC agencies should not be seen :;as the simple result of evil or megalomaniac \!politicians nor,!?f public servants who lack adminiIH~trative,~~U_$,,\\An approach more in the ecoiWnoinfcspirit says that government agencies and ~~eir employees are.reac~ng rationally to the ~Clrcumstances they inhabit,

does

Part of the problem is the nature of the outputs and production processes in the public sector. 1r marlY cases:'pr?ducts':,aff..'hard to In~<lsurei1D.9., ittney have aspects of publi<:$i;i>ds;h<itd to ¢h~tg~aptiCfi fOT:J>roduction technologies-that is, the way'tnatinputs combine to produce public goods-arediW£1lltt9specify. For these and .other reasons, It will often be more difficult to !improve internal structures of information and !incentives in public agencies compared with private agencies. Institutional economics predicts I"weak" incentive systems, long-term employment •• contracts or understandings, merit systems based Hon credentials but not performance, and a "pro:'cess culture."

The Incentives Problem: The Case of The Gambia
As an enclave surrounded by Senegal, The Gambia is in many ways atypical. But many of the problems I learned about during a recent visit are generic.

I

But improvement is possible. Both the new theory iof the firm and contemporary examples suggest '1 that once the problems are identified and adi, dressed, improvements (be they first-best or second-best) can be devised. These will in turn be subject to gaming, ratchet effects, influence activities, and other phenomena; but these aspects, too, can be creatively and productively addressed.

Low pay. A January 1994 government document notes that in 1992-93 government "salaries in real terms were still well below their 1984 levels .... In addition, official public service salaries for high positions are also extremely low when compared with those in the parastatals and the private sector." A systematic study of salaries is planned for 1994-95. But even in the absence of comprehensive data about relative wages and the brain drain out of public service, virtually every docu.• ment I saw and every interviewee made the point .•• that salaries were too low to attract and retain /'qualified professionals in many key areas of ..government.' In the just-completed May 1994 high-level meetings between the government and ... Bank, it was estimated that government the -.salaries have fallen by 40 percent in real terms over the past decade.

4

5
.~,

.... - ..-.-... .',

An example is the Accountant General's (AG) blueprints.Tbtrdi.because such studies are technidomain. Currently, two members of the AG's cally complicated, they must be undertaken by approximately 175 central staff have a professional ..;'~xpensive foreign technical assistants. (Is it cynical credential in accounting; three are away on train'Wko note that this prevailing approach is pushed by ing. None of the approximately 175 accountants in T'k number of foreign technical assistants") the various ministries and departments, which report to the AG, has a professional qualification. The number of studies planned and underway in The Gambia is remarkable. For example, in the Poor pay helps explain these problems. Even at Ministry of Finance's efforts to reform taxes, many the technician level, employees can earn 100 recent meetings have had as a principal outcome percent more in the parastatals and 200 to 500 the definition of a long string of studies to be percent more in the private sector, plus perks. carried out by foreign experts. Government "The biggest headache is pay," lamented the . documents in preparation for the World Bank's Accountant General. "The government pays for thE! sectoral adjustment loan-documents themselves training and then when people come back, they go' drafted in.large part by foreign technical assisoff to greener pastures." . tants-foresee many other new studies. Weak pay-performance linkages. It is becoming widely recognized in The Gambia as in many other countries that there is something economically undesirable about pay that remains the same whether one achieves much or little. How should one deal with the problems of low pay and weak linkages between pay and performance? The prevailing and mistaken approach ,TJ:te Gambia seems to have three components. \Firsi::flny reform should Involve-massive changes ,y :,.-" across the entire civil service. :Se~()Illi)such reforms require comprehensive studies and
\. _"._O, ..

The phenomenon of studies is not confined to economic matters. In the course of a conversation with the Solicitor General, she told me that a large sector study of the legal system was to be undertaken with support from the US Agency for International Development. I noted that I had read a review of the justice system that was done a few years ago. Was it possible that the Solicitor General and her colleagues already knew what was wrong with the legal system. without any new studies, and that they could in effect do a sector review themselves, working weekends for a month?

TABLE

3.1; Two

APPROACHES TO INCENTIVE REFORMS

Prevalent Approach Ends of reform Across-the-board pay increases; horizontal equity across jobs Long-term studies leading to systemwide reform; foreign technical assistants do the studies; learning by planning Budgetary austerity; donor pressure to reduce wage bill

Proposed Approach Selective pay increases that eventually spread; incentives linked to performance targets Experiments with a few key elements of the civil service: public officials define measures of success; learning by doing Begin with revenue-raising and cost-saving experiments that can pay for themselves; use aid to fund experiments "Institutional adjustment," including better information, more client participation, competition

Means

Constraints

Facilitating conditions

Studies; technical assistance; political will to reduce the size of the public service

6

Box 3.1:

DESIGNING

AN

ExpERlMENf

WIlli INffiVl1VES:

Our

LINE FOR A MODEL MEMORANDUM

1. Quantitative summary of the current unsatisfactory situation. Because of X, Y, Z shortcomings (resources, incentives, capabilities), we are currently able to process only A% of the cases we should, and of those, only B% are processed adequately. As a result the government and citizens forgo a, ~,y benefits and incur t.., <.p, e costs.
2. Examples. Here are three recent examples of things we were unable to do that clearly led to forgone benefits or additional social costs. 3. Measures of success. After considering our objectives and our organization's key tasks, here are the measures of performance along which we believe it is fair that we be assessed. Some are quantitative indicators of the tasks we perform and the results we obtain. Others are qualitative ratings, constrained by a curve so that not everyone can be rated "top one-third." Still others may be based on the detailed evaluation of a sample of particular cases.

If we had x, y, z (additional resources, incentives, capabilities), we will in K time-period be able to achieve (even if qualitatively) the following measurable benefits and reductions in costs: 1, 2,3,4, and so on. We are willing to make such-and-such of the incentives conditional on the attainment of so-andso performance targets, which will be monitored in the following transparent ways: i, ii, iii, iv, and so on.
4. Proposition.

She shook her head and smiled. "You're right. We could do a study quickly that would identify six areas that need fixing. But that's a problem here; it's always more studies and more studies. Most of the aid received goes for studies, and then the studies say what we already know." Policy research of course does have its place. It is the reflex of calling for a comprehensive study that deserves noting and criticizing. In contexts like The Gambia's, this approach generally takes too long, costs too much (especially the salaries of the foreigners who do the studies), runs into political •roadblocks because of its "grand reform" nature, "and, in the case of the civil service, probably cannot f:engender enough savings to make a big enough "difference to civil service pay to matter. The different approach recommended here is selective and experimental, It is at once humbler and bolder than the current approach. It is humbler because it recognizes that comprehensive studies are limited in their ability to blueprint change. It is bolder because it calls for learning from experiments, even from their negative lessons. It hopes to shock the entire civil service system into change by experimenting with indicators of performance and pay. It would involve Gambians centrally in the design and evaluation of the experiments, as opposed to their usually peripheral involvement in grand studies.

Specifically, the government would select a few key functions, such as revenue raising, auditing, and the justice system. In each area, the government would challenge Gambian officials to define qualitative and quantitative performance measures. These could include:
Ii

quantitative measures of (a) activities undertaken and (b) results achieved;

~ contests designed by employees;
I!l! estimates of the quality of a sample of activities by peer groups, outsiders, or clients, on the proviso that ratings also include "grades on a curve" so that not every person and activity is deemed "excellent" ; II tIii

staff morale and turnover; and

statistical controls that" adjust" measures of performance by the relative difficulty of the target group one is working with (for example, for tax collectors, which region, type of economic activity, type of tax, and so on, all of which affect the amount earned). These ideas would be broached in participatory seminars. Then experiments would make rewards of various types contingent on the achievement of

~gIlitiOt"l,'

X1 'f'Ch/..;D1· •••. ·
C:C:f,,,nrnPl"\t t'l1't'llm(1,1"I.n,n

fU,the

' .. ,

working conditions, more independence, and so forth. Some of the incentives could be for individulals, but many would be for teams (offices, bureaus, !)even entire ministries). As an example, Teneng [aitah of the Ministry of Finance and I worked with several groups of Gambian officials on the outline shown in Box 3.1. Each agency is challenged to work through this outline, as a kind of preface to experiments with incentive reforms. Overcoming Constraints on Incentive Reforms The current approach to incentive reforms rightly feels constrained by budget exigencies and pressures from foreign donors to reduce the overall public wage bill. The approach recommended must also countenance these realities. But it has several decisive advantages. ·FI~st, it proposes experiments, not wholesale reforms. Experiments are less expensive. Some technical assistance (TA) money now used for studies could instead fund learning-by-doing with . better performance measures and new incentive schemes. Second, because it is selective, this approach can begin with sectors where the experiments are likely to be self-financing, such as revenue-raising. Cost savings can be documented rather readily in certain areas, such as tendering, public works, and legal costs. In other areas, such as auditing, accounting, and investigating! the immediate benefits may be less financially tangible but essential to a campaign to improve government management in general. Third, this approach can be built into a broader new strategy for institutional adjustment:
!il!I

.. , .andpoHticalparties.. procedures'thataliow these different interests and voices to make a difference in policy and management.

.1IIiI~hhcfl
tion.
iiIII

~ §ystefuatic

campaignagmnstrorrup.

Harden the budget constraint. One possibility is to reduce foreign assistance. Another is to make aid contingent on progress in institutional adjustment. .

This approach contrasts\vith traditWnal preaches to institutional development, which are based on the premise that what is needed is more: more training, more personnel and equipment more coordination, more studies and central planning, and more TA. My argument is that without improvements in information and incen·tives-and more broadly, without institutional .•.... adjustment-the usual approaches can be expected ':to fail in the difficult environments now facing ···'many developing countries. Do we have analytical frameworks to guide US, and success: stories to give us ideas? Again, I believe the answers are yes and yes, although there is still much to learn. Adjusting to Realitf presents details on the incentives problem (see also the appendices, below) and ideas about economic approaches to two other chronic problems of nonmarket institutions: corruption and determining the right level of centralization and integration of public agencies, Notes
1. For example, a recent government document estimates that salaries in Inland Revenue are about half as great as those for similarly qualified people in the private sector. The average monthly wage of a customs officer is so low that it is "extremely difficult for staff to manage on their monthly wages. This may well be a facto! behind the widely publi.dud allegations of bribery and corruption within the Customs and Excise Department." ("The Gambia, Administrative Reform Program," January 1994, draft, pp. 27,311) 20 Especially chapters 6-10. Also relevant is Robert Klitgaard, Controlling Corruption (Berkeley and Los Angeles: University of Cali fomi a Press, 1988). Both hooks include case studies of success as well as analytical frameworks,

a~

Improve incentives. Link incentives to information about the attainment of agreed-upon . objectives.

II

Enhance information and evaluation. Put it in the hands of clients, legislators, and those with official oversight (regulators, auditors, judges, and so on).

4. .lndigenous Institutions
Recently, tnY research has turned to a third. area that involves InStitutional economics, although other disciplines quickly enter. How can one understand and "take into account" what might be called indigenous institutions in economic and political development? Concern for indigenous institutions is increasing, and we have seen in recent years remarkable examples of relevant research. For example:
II The

,t study, such as rotating~il,!ings aI\dc~it ~a;;,''! tioJ}$Qrag~gr6upsor water~users.associ.?ti<:l~'~

'These

mal'~~ude

organizations we can see~d

rigorous documentation of the high correlation between the density of civic institutions to the quality of government in Italy.1 The combination of Q\l!ilJtEitiveMil qu.antitative methods to show that "fi(betweentraditional institutions and modem structures distinguishes "successful" and "unsuccessful" Native American reservations.s

.•.... as ofuetlmistitutions" that are tess tangibIe;·: well such as rules of enforcement of repayment and agreement norms for aiding people in distresS"· (when risks become realities), and belief patterns," .The economics of informatioll, helps (or should help) anthropologists get beyond •their debate over whether primitive economiesean .be understood as haVing markets, and instead analyze many institutions in terms of risk-sharing, overcoming problems of imperfect information;' .,and the like.'

negative

eVen
,..

j . .. .--::,:: -

III

me

New studies of successful African management which emphasize the roles of indigenous institutions.3
I!lII

economics can help us to see that in4!g. enous if!§.ti.w,tiQTI$arepartofa complica~~stem o£relationshipsl:wherein the institutions are· .. .. themselves a function of such things as the socioeconomic setting and conditions, the particular paths of local history, and public policies, and the same time those institutions change with some lag and themselves condition the "production function" for development.

)~or;d(

at

Studies suggesting that indigenous institutions are crucial to the success of many rural credit programs:'
II

Indigenous instituticns « f (socioeconomic setting, history, policies, ... ) (2) Development outcomes == g.fsocioecenomic setting, history, policies, indigenous institutions,

One of the first applications of the crosscultural information about local institUtions from a~~anthropol()gica1ata bank to a pra(iicru d problem of development polky} Two sorts of economic insights are relevant, is the idea that. in the absence of formal legal systems and well-deve:loped markets, people solve problems of incomplete contracts, adverse selection and moral hazard, and risk-sharing through a variety of indigenous institutions.
8

...)

(3) Utility function for development outcomes :;;; h (socioeconomic setting, history, indigenous . institutions, ... )

The third equation states that the utility function

i~tittltio'!1S'

:~~t~~~Sa~;~a:;~~v~~~~~ii~1r:;~r·
arrtlJi\gothet variables. Thus, indig-

formulate useful Q' uesnons anthropologists/and .. , . ......;~o . '. .. equations look likeinfacl?What sorts of interactions there amongpoiid~ a'{ld indigenous institutions? How (to suchinstif!1tions mang~ and how might local peopl~.riiist those changes or accelerate them, if they sO desire"?! .

or the'·
autonomy of .from materiii!causality, culture provid~ andindigenousinstitations embody.;±ig,tf!?incidentally;this text Can also be a CridtLcoe'lly,fri,mnoneconomists and nongovemm~t organizations vis~~-vis economists, quantitative approaches/and large formal institutions.

are

mat

·sinipre;'howet'·~r:!h1l1eY....

3. Enabletlum. Aninsight of texfis institutions, even traditional onesfreighteit with meaning as well as functions, are not exogenous or static. They respond to, amongether-things, conditions of risk and opportunity, and to both market and nonmarket failures. They.will be optimally enabled=-in.seme cases strengthened, '; in others perhaps replaced-by creating a wellfunctioning democracy, the rule of law, incentive~~~~:;~j~g~:~:~n~~ds:ffi1t~~~~~~~r' driven and output-oriented government ~!I~~~.~~!2~!I .. gyl'l~.!l'bureaucracies, and v~gom~~,~m~tjEtx~_tn,~l;~ .. .;with ample information. In other words, after ics of meaningancf~r1i{lcp~@l~gfi.r:!~f:§£i~!l0' structural adjustment, democratic reform, and ahdmatket EoV/~r. Thus, they potentially have both benefits imdcosts, and both should be "taken "institutional adjustment" indigenous institutions into account." will be optimally "taken into account" via correct incentives for the public and private sector to do But what exactly do the words in quotes mean? I .i ; so. Therefore, says this text we don!t have to . figure out how to take them into account; we just discern very different answers to this questionwhat might be called different "texts" in the sense have to create a system with lots of information of literary criticism, namely something akin to and correct incentives so that local actors will intellectual reflexes or templates, Let me distinthemselves do so. guish, artificially but I hope recognizably, four such texts. 4. Connect with them, reconcile with them. Dials work hypothesizes that at the core of Africa's ;,' 1. Cite them resources. To some, taking indigedevelopment problems is a 4!~r&1}!~E!:: ..p~tween '. nous institutions into account means giving them modernandtT§<!itionalinstitutions .. The coronary resources and agreeing to go along with whatever Is Inaf aiire(:ongU~ii9n~~gt~heseinstitutions the is §,Hsyyer.Twofeatures of these proposlt1011S .!; they come up with. If one asks why such institutions should not simply be left alone, the answer is text seem important. that they have legitimacy but lack "capacity." A danger here is a kind of "traditional fundamental, One is akin to "I'm okay and you're okay," ism," and what many Africans apparently perj.nd~~n{)~~i:1I\qrl.JQ4?rninstiWJiOO£halle.SQme:: ceive as inefficiency and unfairness in many local thing;lsigif~:they must be reconciled, merged, institutions, even when "legitimate." For excreating something new. As a consequence, to ample, some possible costs of relying on indigthose relying on this text. it is uncomfortable to ask enous institutions come to the fore when a gender whether for some tasks and in some settings the perspective is taken. problem is too much merging of traditional practices into modern institutions, or that in other' 2. Give them respect. To some observers, both circumstances merging may lead to co-optation donors and governments systematically bypass and even demise . How might indigenous African institutions be "taken into account" in the improvement of governments, markets, and management? The Bank's ongoing research under the label" Africa's Management in the 90s" is exploring this question through a series of fascinating studies," As We have see", indigenous institutions can among
U

Guidelines

for Adjusting t().~tlsti~tion8

as's.' ;';,
"'·%'·N.%·~ ..••

.................•......•

,

_ .•......-- -

.. .•...

--

-----~---.-~,,~-~-

The read;r ~m recog~l;e that these four texts go wen beyond issues of indigenous institutions and development Indeed, I suspect that one's choice among these prototypical texts is influencedby such' things as one's academic discipline, "personality type," and "culture," perhaps as much as by the evidence on the specific subject under discussion, such as indigenous institutions and development. But the point here is not the origins of these "texts" or even their validity, rather that we need a response that is akin to literary criticism. It is worthwhile to recognize that multiple texts exist; to analyze them and deconstruct them-e-especially our own texts. We must be ready to hear these differing responses and to double check our own reactions to make sure they are not simply texts. to How might we push such discussions beyond texts? One. s?aI9fgevel()pme,nt iSJoreplaceil1cfficient iri'ithOcis;)fprq<1u(;t1on (technologies in the amplest sense) with m9rE:eJf1ciel}ton~. Since productivity is often higher in other parts of the world than in many underdeveloped regions, it would seem easy to develop" simply by borrowing better production technologies. But we have learned that 1J?~rowing is often nots:uccessfulbecause (among other things):
U

···~~~~;~ia~~!:~.:~ ~ttings the same
IMY ... . . and institutions. Here are

of problems

.

tinpaCltffiisnia:trtxofproblem,· ~oCiocu1tur£11context,
..

or no useful role.·~Weneed to.

sorri~id~al(6foing d

so:

Ta.kE!

rnar~~ts~san example. They exhibit predidai)fe'prObieiTlSofinfonnation. oo:mQmIesof scale, andeIlf()rceIll~nt: I~i!.~f;;~!}()H~in~~tut:i0ns may help OveITOI11,£th~epr()blems.Forrnstance, the riskiness of specific potential clients and more efficient screeningtt:chniques. They may provide alQ~~st and credible way of pooling trapsil,£P(}Il£Osts and risks, takingadvantage of economies oEcale. They may possess both techniques of €flforcement and (if carefully designed) appropriate incentives fur applying them.

theymaY'Q~Y~~~~:!fl·&q£ffi~~2non

00

R~sourg:~~ilr€!i11St1ffident(capital, human capital). ..... Incentives are skewed (public sector, private

Ii

This is only One example, The general heading for project designers and policy makers is this: What are the market failures we see? How might ind igenous insti ttl tions provi de. theinf(jTf[\;)tiOn" economies of scale, enforcerilent:,incelltives., and so on" that could help mollify those failures? What .new problems would using such institutions entail ..(for example, their managerial inefficiency and their own implicit market power) and how might these be ameliorated (for example, management training, accounting systems, oversight, and so on)? Bureaucratic Failures

sedor).
II!i! Ilil

Pro peTty

rights are weak

.. TypkaUyit is difficult to measure outputs in public agi~Ci~ a~~ tl,1eri:fe:.rediffkuIt to·create appropriate incentives \Vithinthe bureaucracy. Controls, disCipline, profeSsionalisn1;. and exhortation attemptto substitutefor links between pay and performance. But they are imperfect substitutes, monopoly plus discretion minus accountability is also a recipe.t9tfQP"tlption. In many countries, bureaucracies underoerform and even become c~~~tneswith low

Monopoly power is seized for private ends and not public puqmses, in part because countervailing institutions are weak or absent.' Responding to these problems, development

strategies have moved from the simple increase of
physical and human capital to the adjustment of economic policies to governance to institutional adjustment. Now we raise a further issue, "Might

and

pr~d;t~ry,perh;;;ps~~Pea;1fyi~

create a o'f ...."" .. oii"l,; ,ci1"llTiimliiilli'ot-k overcoming these failures? policy analytical analysis,which .. task is.to examine the sources of the failure one-byone, Might indigenous institutions be util~ t:I:?" " situations: .r.: provide better informatiohontheoutputs of <,\i' ,,<.' '"iiii'" , """"",' government agencies? Howt"Might they provide These headings are rh~pe~tgg~ti;;~utthey are incentives to government agents? Might they be" ,,' , "certainlynot exhaustive. They try to exemplify efficient mechanisms for the delivery of pub He three points, services, perhaps because they ,~/intemalizeHsome"" ,," oJhde~tand whet-J'j of the costs of predatory behavior and therefore mIght have a role·ft) pM , . have better incentives to deliver the goods? Might specific sector orPfo~lema they be used to provide "competitiontfn service delivery? Working through such questions might n<?l"!~~t~$~j~· . . . generate a host of ideas that otherwise would not mod~lJpr analyZing market or nonmarket failures. The one Ihave been using is economic; ..... have occurred to us. there may well be other useful approaches. Third,ii Once again, we must also anticipate the problems ~, in this analytical effpr1:itwiltofQ')ursebeesSentiai that using such institutions might entail--<.:oto hZIVe localknowledgf®~involvec local ex~d:s in optarion, inefficiency, the creation of market poweq indigenous institt,lt19D§.Through this process one and others-s-end consider ways to ameliorate or !; would analyze the benefits and costs of various ways of taking indigenous institutions into account. prevent these problems in advance.
-I'- ...

il

The Rule of Law

1

to loe would likely have orms, epending on the problem, the The administration of justice is often an obstacle to institutions, and the sociocultural setting~.SoI1'le: econ6mlcaswelt··-as··poUtinl·oevelnpnrechC·····MoW times i~di~~n()us ins~.!mg£mSW;QuJ&iJ:,ei~ortant. t, Sometimes the result might t:tadltioni:11itlStltutinnsheIpremedy'some of m!ghf»€toi~~~l~t" nous i.nstitutions alto- .. " the failures of the current system? As in the previgether, The point is that by working systematically ous headings, we would analyze v~ltiQ·us:.dimen~ sions ofinformatiqnan,dincentives. We would add in project and policy design through frameworks like these, we might be able to go beyond "texts" £urtherhe~ding~.'C~~ traditional means of dispute and generate practical ideas for taking indigenous resolution be exploited to replace or supplement institutions into account-ideas that through our more formal mechanisms? customary modes of policy analysis might otherwise Again, the c;~~~i~!J£L~gt",}!~iugj.nsi.&~.;:t8~~.", never surface, institutions·~O;:l~!?!>~£l,itJX~~~'k.~~eIL What sorts of trainirigano accountability might render these Notes "informal" mechanisms more efficient and less subject to monopoly-creation and arbitrariness? 1. Robert Putnam, with R. Leonardi and R.Y. Nanetti, Making Democracy Work: Civic Traditions in Modem Italy Would these "improvements" themselves threaten (princeton, NJ: Princeton University Press, 1993). the indigenous institution in any way?

Adiusti

attdS()metime~~m ..

Governance How might indigenous institutions helpirnprove representation, opeIU1€:ss, transparency, and legitimacy? Consider a Bolivian initiative-the use oflocal organizations not only to define needs and in some cases deliver local public services but also (in concert with this) to create a new, set of

2. Among many papers, see Steven Cornel! and Joseph Kalt, "Reloading the Dice: Improving the Chances for Economic Development on American Indian Reservations" (Cambridge, MA: Kennedy School of Government, March 1992). 3. for example, the many studies described in Africa-Technical Department, Indigenous Management Practices: Lessons /01' Africa 'sManagemeni in the '90s, Work Program and Methoow ology (Washington, D.C.: The World Bank. May. 1993);

appears to be the introduction of a social mechanism that lowers transaction costs, while supplying effective peer i '. pressure fotscreening loansppllcants and collecting i6anS'" .....

to success

(p.
Beyond: '.. '. 'Transplant" to IFlStiUtttCaaj 5. S. Rom~~ff, S.Cart~,andJ.Lynam, "Cassava Production ReconelliatJon{Wishingmrt, D.C.: TIle World B_<" . and Pm~s~j9gi~a¢rg~;9u'ttJl:alS;llIlPk".{New Haven, cr~ i fortheorrting}>'~<;F

HumanRelati(lnsAreaFiles~peWnper195)1).
'j" ."
'j .• ,./ .• -. •••••• ', •• .,.... ..... ".'."" .....:

.

'

6. Many atlliB~ hav'~ii~i~thiii~iri~iMi~~s"means b~Vol. "organizationS" and "riohn~ beU6fs,even patterns oft:onven·
' ..'

1O. Robert·1(Jitgaard,t~Poisonous

" .. ,

'.

,§,j

Texts," NegotiatiUn :hurrwl, .

g,Np;3 (JulyJ992).

A.:ti us~il1g···o "Inst!lutions t
institutions fha:

5. . Jnstitutional,A,dju~tment

_
worried .about

lnstit~iil)~i ~djustment means irm rovin the ...
.

(Let us take "we" in the most generous possible sense to include especially people themselves as they decide to resist or accelerate various brands of

-~~!l~~f
"developmenL';l.. ..
Long ago Albert 0. Hirschman warned that no matter which choice we undertake we will be criticized for it. If we are what he calls "trait takers" we decide to work with the institutions we find. We will later be criticized for perpetuating inefficient, hierarchical, sexist, urban-biased institutions-seven if our efforts succeed, If we are "trait makers," trying to alter underlying institutions, we will be criticized as naive, imperialistic, culturally biased social engineers-s-even if our goals are met.' And this is one reason why dealing with institutions-a9jL\~!i~g:Jh~.}?[~~j~~ ... · is precarious. It is compoi.mdedb act t we seem to know relatively little about how to do either. True, when thinking about institutional adjustment, there are appealing analogies to structural adjustment. In the case of structural adjustment, we worried about "getting prices right" for private sector agents. Now we worry about "getting incentives right" for public sector agents. Ther~"~~. tried to create opt"n and markets. Now we try to create open transparentg()yem~ rfi.etits.There we worried about property rights and privatization in order to decentralize economic decision-making efficiently. Now we worry about empowering stakeholders and optimally decentral-

fQur

economic competitl(m,.,Now~e'Y()r:rx,abOut administrative and political competition~·
" ",: '-'-:-'~:- .'-:':=? .: ;~: .:.: --";: .

izing public services. There we

There are, however,important'differenceS that affect the process of designing adjustment strategies in the two domains. I will exaggerate to make a point.

Devising a structural adjustment cel1terS., on a few crucial decisions at the top: devaluation, setting prices free, lowering tariffs and removing quantitative restrictions, deregulating, and cutting spending. A few people can make and dictate these decisions in a relatively short time. In Turkey, for example, fewer than ten technocrats under Turgut Ozal knew the content of reforms •.before they were announced.s If the'political . leadership wants it done, it is done. (Aside to readers who are political leaders: I did. say I was exaggerating.) In contrast, institutional adjustment requires extensive tailoring to specific circumstances, and it will require ownership not only by top leaders but by the rank-and-file officials who will implement institutional reforms. Thus, in most cases institutional adjustment will require extensive consultation, participation, and joint learning. Process is . crucial. difference implies, I believe, a greater emphasis on experimentation. The word experiment may carry unwanted connotations, but it captures. an important dement of successful institutional adjustment. A lesson from many past efforts at administrative reform is that attempts to craft once-and-for-all, systemwide changes fail. For one thing, the very complexity and finality of such reforms offer too many chances for vested interests 13

prograrn

nus

.

.to leanj.~they
FramlIl&

:::.. -t:,'.'_::;:::·:,·':'·

adjustment efforts . ..•. ..••... . empha~es . the.learning. It also sl.lggestsmeasureme,lJtMt~t ~. specified period of timeiwhich is also welcome. There are exceptions these generalizations. Sometimes, in order to be credible and ""H'~lut> changerqH~tl?e.· e.i'ih'Cl .sudden, the annotil1~p .. experimental P~99~q%;~t'l;':/ But with~eg~;d'~i:~~'~~~~djUstment, my experience is that it isWise to begin in carefully defined sectors with efforts that may be criticized for being "too simple," but from which lessons can soon be drawn and improvements made. Doing too many sectors with too.complicated a system is theoretically preferable but practically perilous. But aren't these institutional issues too sensitive and too contextual for the World Bank to consider? Do the Bank's comparadveadvantages include such things aslocal knowledge, politics and history and culture, and projects with large requirements for participation and partnership in the sense I····· have been describing? 1\These questions raise serious issues. But in some cases I believe the World Bank's entrance into 'uenalad'ustm 'c. Somees outsi e assistance can help legitimate sensitive issues, help leaders make credible commitments. Incidentally, with regard to institutional adjustment the dividing line between "willing" and "unwilling" leaders is not as dear as journalistic accounts may make it appear. For example, in my experience many leaders are, so to speak, schizop~11l<:i'.lQ()tjtSQl'1J!ption. They may sincerely loathe it and wish to eraciicatt':;. t, while at .i the same time participating initoraUowing it to occur. It is also true that in my workshops on corruption, after some time people are remarkably frank about the corruption that exists, how it works, and how it might be prevented-s-even when these analyses belie an incriminatingly intimate knowledge. I do not have time to describe these several-daylong workshops here, but I do want to cite certain As in pants then self-diagrtosis;ands~lf~preScription. corm ptiori'iil' thei~riwn ministry or coun causes/3rtdlfsp0sSibierernedies. assists in several ways: by asking questio helping combi~eseemingly different phet1()Il1enaor separate seemingly similar ones, by pushing when the group avoids wor~ of escapes into relativism or cynicism. Out of suchan experience emerge both a deeper understanding of general phenomena and specific manifestations of corruption, and a plan of action to begin to fight them. ..
I

goo~

..".. ..... .

. This plan may r;~:ir~'aSsistance, and this where external actors carl help. Beyond money and ideas, outsiders can help in surprising ways by Tmpos-ing" conditions and deadlines that fortify local . decisionmaking and self-discipline. An aura of international respectability-"part of a worldwide program tofighfthe universal plague of corruption" not just here in [country x]"-may help nervous actors coalesce around reform. As part of a strategy of institutional development, : the World Bank and other agencies may be able to "lever" such workshops. A project that follows such an event=-cr several such events at:different levels of the pubHcarid private sectors-may use . the workshop's recommendations, co-opt key actors as managers and monitors, and via carrots and sticks improve the chances that the adjustment process will succeed, Healing the: Aid Relationship The therapeutic metaphor may be carried one step further-the need to adjust the institution of international aid itself. The relationship between

is

can be analyzed using inStifu~tio:ruH OO)IlOmiltS/.\ Consider three incentive n.····rN'>!~J ...'mt:; relationship. and trani "scend certain historldl.l roles. Listen and' see if his words do not also ring true for the aid giver, the foreign adviser, the expatriate activist:.
....",.rl'n"T'

Se(;Q~4~·\.staff members of dono~iristi~ti~~~ften

arid promotion depends hardly at a110n .. .. . tq~'" eventual success of "their'" projects, but fI1stead on shorter run measures of "craft" quality and even the amounts of money or number of projects moved through the donor bureaucracy. If officials are forced to operate through narrow-gauged projects, it can inhibit their ability to tackle crosscutting institutional issues such as information and incentives.'

Ia~'R.·aEl5.rqprI~re·Jn{:~lifi~t;;S;·jheirrom.j5~~s~pon

thir~/ len~iflga.gendes .. dono!&p*imallysharethe··· risks that their conditional aidentaila..In principle, the amount they are repaid should depend in part on the success of their conditionality, given that the recipient follows it Currently, the risks are disproportionately borne by recipients. These three incentive problems help explain some of the chronic problems of the aid relationship. three will be difficult to remedy, Like some of the problems analyzedabove, their amelioration may require an experimental attitude, the participation of staff and clients in pilot efforts, and an openness to learn with other donors and with recipients.

An

"In~moreenllghtel};4~~'QrI~/~:~R~n~l'ote in 1950, comparing thesHuation to that in Freud's time," and under muchIn0re ~o lic~~histori~ cal conditions the the whole problem of jUJicWuspartners1iipwruch expresses the spirit of analytic work more creatively than does apathetic tolerance or autocratic guidance. The various identities which at first lent themselves to a fusion with the. new identity of the analyse=identities based on talmudic argument, on messianic zeal, on punitive orthodoxy, on faddist sensationalism. on professional and social ambiticn=-ail these identities and their cultural origins must now become part of the analyst's analysis, so that he may be able to discard archaic rituals of control and learn to identify with the lasting value of his job of enlightenment Only thus can he set free in himself and in his patient that remnant of judicious indignation without which a cure is but a straw in the changeable wind of history ."S

analyst~ltst··.o'n~iMore
<

Notes
1. Albert O. Hirschman, Development Projects Observed (Washington: The Brookings Institution, 1962). 2. Anne 0, Krueger, Political Economy of Policy Reform in Developing Countries (Cambridge: MIT Press, 1993). p, 131.
3. Samuel Paul, Institutional Development in World Bank Projects: A Cross-Sectoral Review. WPS 392, (Washington, World Bank, Country Economics Department, 1990). esp, p, 45, 4. Thomas Bucaille, "Metamorphoses du probleme africain:

So, the economic perspective suggests that aid resembles a transaction involving risk-sharing and appropriate incentives.~~!.~is!.i~!ng~~.th.g.!l a transa,':~~()f.l: .... i~l?.~..~l<t~()l1Shi12~,,£!j?"~!!)~~b!r. r (as the Bank's Katherine Marshall among others has emphasized], The aid relationship is itself an institution in need of adjustment Beyond economics, might other metaphors lead us to the right kinds of institutional reforms in this domain? Thomas Bucaille has usefully noted that aid is, or should be, a pedagogical relationship in the best sense of that dangerous word.' Understanding our individually and mutually inappropriate incentives

n.c.

L'economie africaine et la cooperation francais depuis 1945," Etudes, Vol. 373, Nos. 1~2(1990). 5. Erik H. Erikson, Childhood and Society, 2nd ed., revised and enlarged (New York: W.W. Norton, 1963 [1950]), p,
424. emphasis added.

This appendix tries to c6ri~~y ili~'~~~~pe of an
economic approach to the}l1ceptive problem, using a variation of the prindpal-agentmodel. central idea is that since thea~ent'~performance

A

is only imperfectly measured, paying wage based on what the principal can observe creates a risk for the agent. Because of risk-aversion, the agent would prefer a fixed wagethatdo€s not depend on performance. But the principal dislikes fixed wages, because then the agent has no economic incentive for higher performance. The optimal wage agreement strikes a balance between risk-sharing and incentives. Usually this combines a fixed payment that does not depend on a measure of performance with a variable payment that does. The appendix then tries to derive, schematically, some policy implications of this approach. The Model Suppose you are the principal and I am your agent Your profits ate a function of my efforts: Profit ",.Ple). But my efforts are costly to me: Cost = C(e). Ideally, you would pay me a wage equal to the value of my marginal product. But you don't know my effort e, at least not perfectly. You can observe an indicator of my effort, which is
z""e+x,

a

B is a measure of the incentive intensUy of our wage contract. When it is zero, you pay me a fixed wage a. The parameter y measures how much weight you give to y in rela~on.t'?:z.
What contract wou]d~ ~iany optimal (for the two of us?) Assume expected values of x and yare norrnedto zero (for convenience), and assume no wealth effects. My certain equivalent is equal to expected iricome=-costof effort-a risk premium for the income risk I bear because (x + yy) is a random variable. Milgram and Roberts (1992: 247) show that the risk premium turns out to be approximated by 1/2r pZVar (x + yy). Your certain ...equivalent as a risk-neutral employer is the expected profit-s-expected wage. Both of our certain equivalents depend on the four variables C1., e, ~~. nd '(. By the assumption of "no a wealth effects," an efficient contract will maximize the sum of the certain equivalents. But there is also an incentive constraint, which says that I as your .. agent will set my effort level such that the T!.1!!;8i£!~ benefit» marginal cost So an employment (ontt'acfis'effiCie'n'rrfarid only if the choices a €, P, and y maximize the total certain equivalent among all "incentive compatible" contracts where p. - C'(e) xz 0, The solution generally has some fixed pay and some variable pay that is a function of the two measures z and y ~
f

"

where x is a random variable representing the measurement error. You can also observe a variable Yr which is not correlated to e but is correlated to x. For example, suppose z: is sales of the product for this month. It is some function of my effort but also depends on

17

Addiqo~ efforts ~ ".. ' £Pll\:llit se~}~f5f big gains in effectiveness r
= agent's
_lr'V;\nfl':

6rts by pUblic

constraints, ..·· ....

B·/ .
•••• , .. "_ "C -;,-

yield no gains in eff~v;~.~s 't

'~~_-li

risk-aversion

Employees are almost riskneutral, perhaps because ....plentiful qpl?<'~ti~ exist .... and thef~~~r~qfWen-off
".;. " c', '-. ~: ..,:

:_'-i~'"

C",_,' ..

,'co,

_'

.-

to measure . ,)~.> ...,
C'(e) = responsiveness of agent's effort to incentives

Effort and reStd~~reeasy

Effort is very responsive to incentives (for example, high discretion)

..<'Hffoi!: is not responsive to .ince.l:'ltives (for example, fixed.'paceactivity)

From this model we can derive the incentive

intensity principle. Th(;:§tr~l:'lgiligfjD£~l:'l!!.y~. should be an increasil:'lgft1I1~2!tgL!h~JnflIgYm1 with which " reffiITiSffiffiet:iiSk;lli~itccura.,.." r;y,,,""., ,",.,."""..,,,,,,.,.,. "' .•................ , , ....,..•.." ..,.,."." .. •........ "., , " ,.,",.,', ..
performance is measured, the responsiveness the agent's effort to incentives, and the agent's risk-tolerance.

Insufficient funds available many poor.countries.

for evaluation

in

of

.'.Th.: j,~s~~.1!!..~~~ensity,pri.ndplesuggeststhat
conditions, trying to raise ~Quld .have meager and indeed perhaps: Deg?tiye

iric~~tfv~·:WMI~9~Y1.~hi~ifjg~ni.kYr~,!~gacies·
Uridertnose

tilla~~s?me. uf\faV'0!,<!~I~.com1iti()ns!

.iDtense

p

~ "" P'(e)/

U + r[Var(x

.pyy)]C'(e))

effects.
. .But there is a third point !£~r",~IJ.1"~!.~"rsof the incentive intensity princiPle are: not necessar.'

where C''(e) is the slope of the marginal effort curve. This formula is computed by maximizing the total certain equivalent of principal and agent with respect to e, 1

~b~

..·i1y·lrITfffITtal51e:···lnparncul"ar;"flottce·'t'l1aTwnenthe measUrelnent of performance improves, a wage

The incentive intensity

under··oome'·cor1"dfffons·Wiso"ptimartohave"'c
and not favoring performance-based pay,

principle suggests that

"hl~~i •• i~!~~~:1.lj~,n~Y~~; .• QgT!i"ng~£·!?ther

.{n!oiiiiatw:,tlS"linrrerreait'"

package can be constructed that both e~h:§!.n~ incentives and reduce:f> risk.3 This is why better

refo:;;;;:"""'"

shows sOme 5l til"eextrernircondinons favoring

(:s>,,~itionsa Hat wage .is.tne right choice. Table Al

Coping with Problems with Incentive Reforms
Box Al summarizes some of the difficulties facing performance-based incentives. Beyond the categories suggested by the incentive intensity principle, it adds considerations of incentive dynamics, political economy, and layers of

In many public bureaucracies, especially in developing countries, the conditions seem to resemble those in the right-hand column or Table Al. 0t,£,~~E!:lL' ortance is the difficul of

meas

g~ ....

Va

'

hierarchy.
In particular, agents and principals may take dynamic steps that undermine incentives and information. Consider agents first:

1m

Agents may performance bonuses

l:ipt;::Ui;U nl"<>'Tt>nn ,< ..... , .. .' ..... .' .

". . . . ..... ..

cannot
"

optima!incentive intensity any longer, even in theory.
'..._~.:::>,}_ _;"_" __
c

No~eth~l~s,thev~rycateg()ries s~ggested by
II

Ratchet effects: after learning more production function, principals :t:ll0y~th~goal posts, leavin£~g~~"~,,,:Y~1"S=,.lJft~RE!tf£!e.

a6~~tth~'

g

Intermediate layers ofthe'b~~:j';acy may simply lack incentives to undertake performance

theory as undermining perf()rmance incentives provide a framework for considering how in practice these problems might be mitigated. Box A1 contains some examples. And, building on the' incentive intensity principle, the framework for policy analysis in Box A2 suggests ways to make incentive reforms more likely to succeed"

Box AI:

S~iARY
P'(e)
T=
=

OF

SotAE CONDmoNS

A~~G

THE DEsIRABILITY OF PERFORMANCE-BASED

INCENTIVES

Varix +yy) '"
C''(e) Additional
=

Marginal social benefits of more effort by agent Agent' 5 risk-aversion How accurately agent's effort can be measured
Responsiveness-of-agent's effort to incentives

complications:

measurement techniques when possible; involve agents and clients in performance fees-for-service: watch out for bonuses 1:11,a1: become standard),

L How to afford incentive schemes (use non ..monetary incentives as wen; use samples; borrow appraisal; use partial

2, Extraneous factors, determine pee) (control for them statistically; use tournaments, rankings, but these create side-effects), . 3. Teamwork (may heed group incentives, but then free-rider problems; collusion).

contests, relative

4. Dynamics and political economy (skewing agents toward the measurable to the detriment of the less measurable; "influence activities," including dissimulation, gaming, sabotage, and corruption; ratchet effect; creating disincentives for transfers, learning new skills, and so on; answers include a richer informational environment andprocesses that build transparency and credibility),

5. Layers of bureaucracy

{evaluators may not have correct incentives};

••

'.

c.

Fore~unin~itemt~ov'~ manae'ement help appraise progry~ and set new incentive schemes, along with.fi guarantee to return t",status III' Incentive reforms feCful10lhe.p?trti.qpatipnor~ tjU£l ante under agret?d~upon conditions, thismay employeesthernselve~ inth~speci.D,c~tioriof cash engendt¥th~ cg.r..a4~t;tceto enable an experiment agency S objectivesJ peltprmance fu~~~/81ld. with incentiye.~Yj;Qbegin. ., ):<#;';%'f'< ., incentives'Thishelps educateeveryorieon~e . plans for all linkS befv,teen effort and val tleo-addeq ... '.••••...•....••....... II! Avoidi~centiy~~ru;ter J III Helpernployees improve·tl1eqU ify ofthcir and a11.til11e~,.L(;am6ydoing. Make sure affected efforts (training, feedback-ort achievementS). parties take partin.theevaluation of the incentive experiments. '< mr Sometimes p'(e) is close to~ofor anYi . individual but is large for gioupsof employees. • FaciHtateernp19yeeself-seIection. Introducing In such cases team incentives ate more feasible performance-based pay can be expected to lead and desirable than individualincentives, (Freeworkers with lower risk-aversion to prefer public rider problems may thenernerge, which demand sector jobs. another iteration of solutions.] 3. Reduce the of measures of performance. 2. Reduce the risk-aversion of employees. ~ Include information from clients. II1II Empower clients. Seek analogies to market II!! Raise the level of the pay. power Of joint management. Experiment with user E Help remove employees' uncertainties about charges and analogies to them such as in-kind -. pay-for-performance by running transparent experiments where employees {and clients} help to contributions, sharing them with employees. design quantitative and qualitative measures of II! Quantitative and qualitative outcome measares " performance, appropriate incentive schedules, and can be used. So can peer ratings, as long as ratings
<,"

'.'.. it is?eing~ught. _WMt ofthe ()rgaruzation?;V€h~tj does it take to perform them be.tter?: .. ' .: ...... ..... '" are Hie .

",n>t"'m"""

~d~i~~~:~har~e ..... J•

a1

agertdes

variana')

The problems raised by dynamic considerations and the political economy of incentive reforms do not yield ready solutions. Probably, however, part of the answer concerns the processes through which incenHves' are!;9nsti::uCii.:d ..and,.test.efi ..and.reas. ~~~~~~~g.t changes in process can be under.. hese T stood, I believe, as a rational response to the possible dynamic and political economy problems with performance-based incentives.

(1) perforp}iH2£~2I\ea~¥r~g,JlJ:tz~lil~~kl!~~.~nsi. (2)

perfQfll1ance •. <1114.high.leyels.·of. corruptiol.l .. iUbe w chronic becaUSe,.QUKIUodels say, theyare.:m;tipnal" resp?nsest() a miserable organi?i1tiQ[@l envirQl.Jfl)~;)!~.. For organizations to work better, this environ'ment must change.

The last point opens new horizons

for thinking about institutional development. From the perspective of a given manager or minister in a given system, refo . incentives rna be im IDle. Not 0111y'(15 servjc;;f~j~s riot the

Implications

or

re-

sources to generate the measures
We have seen that the desirability and design of pay-for-performance schemes depend in predictable ways on aspects of the task environment. Given the task environments found in many public bureaucracies indeve!oping, cOl1l1tri~s( Eerfor~. rnance paYlNill·fa.it ahacoriseqlJer;ily iow"jevels of
....·<:;:i.;"·~::' ,..:.,:.:.:.:::::::.:.,.:~:";.~~~(,~,,;~:~,;=;m-r"'=-;:;-y
.. . .., .....

pe~f()~nce on which an effective inc€DtiYe5:ystfm ~~~s. What is needed is analogous to structural adJust. ment: a chaZ\g~~in the ...c'~r'j'%~ enablin .',.' '. key feature of theiieeded ment agencies do and what results they achieve .

retor:rrtfs'better

Tntormation

about what gowrn,.

IIiIi

. ..

. ..• " .' .....

•. .'. .',

takenitlto .:

.;:i.p

account in the design ofinCeti.tive schemes (the .in the incelluveiritensity prlndplerotildtry .tt>.. imeasUfesilch exttane6u.svru:iJillles;wTIichare"

y •..

itPl'1101tior

givenweightg) .. ExarrtPlesdtt'~l'Itromrtg(ofi'" students' socialbackgroiIDffiHrieS£mates of
school contributions to leartting, and in the Philippines Bureau of Internal Revenue controlling for the tax base of a local di~trict in . estimating the efficiency ofthedlstrlct office in. raising revenues. Also, an Incentive scheme may employ measures of relative performance, analogous to tournaments, which help "control for" the extraneous variables that affect everyone's performance up or down, 4. Reduce the costs to employees of additional effort. Begin with the easiest cases. In particular, try reforms in areas where performance is relatively easy to measure objectively arid where the revenues raised or costs saved can make the experiment self-financing. II!I Through training and better equipment, shift the cost-of-effort curve.
II!I

5. Reduce the costsof providing

incentives.

worries them in design and evaluation. '. ,... .., .' 1& Challenge technical assistance ITA) by··.·· ..... foreigners. For example, learn by doing rather thanattempting comprehensive studies: that often end up being inconclusive or unsatisfactory. For example, use TA funds to finance experiments where local expects and even government officials carry out the required "studies" based on the participatory diagnosis of what is already known about problems and possible solutions. I!!i Privatize creatively. This can mean experimenting with hybrids of public and private sectors working together to provide services. Information about performance may incidentally be enhanced.

The task is daunting, precisely because public agencies take on the provision of the kinds of goods that private markets will not optimally supply, But it is not impossible, as is testified by an exciting array of efforts around the world,

effective management is elmosttmpossible" [p, t 75].) As· prospects for performance pay, the two other comers of the table are in-between.

----

..

--~-----.--~---~'1lS,

Cart measure

Notes
J. Paul Milgram and John Roberts, Economics, Organization and Management (Englewood Cliffs, Nl: Prentice-Hall, 1992), p. 247. 2. James Q. Wilson distinguishes four types. of public organizations, depending 011 the measurability ofwhat might be called their efforts and their outputs. In "production" organizations, both efforts and outputs can be measured; here, the prospects for performance-based incentives are strong. At the other extreme, "coping" crganizanons, in which neither efforts nor outputs can be gauged, are weak candidates for incentive pay. (Indeed, Wilson says, "In coping organizations

ou#cmne.t

P,0duct1on (e.g., ;gtemal revenue. =is! :rerun!)',
post. office, FBI) ~ (e,g., I!mlOO forces in peseetirne]

t-:raft (mf~¢'.'!lt

rorpsof""~~
fl!hgm)

Conm« measure
outcomes

Coping (foreign ~
oome ~oo!s. activities)
!!OOl6

p!lIiee

StWrt:e; Based on James Q, Wilsoo, Bureoucracy CN- Y Olio:; Sll£ic lI«lb,
!989), Chap, 9.

3. Milgrom and Roberts show that there is an associated monitoring intensity principle. More resources should be spent on monitoring when it is desirable to give strong incentives. Measuring performance carefully and providing intense incentives are complements. It is also possible that policy changes exogenous to any particular manager improve

The f91lowing points hirltSor lessons from the

shidi~d.

• reforms employees themselves in tnesoecmca agency's objectives, performancemeasures, incentives. ... ...
II .In

and

designing performance Inea~ures, it is helpful to define "key tasks" -in other words, to analyze the organization's "production function" carefully.

• easiest cases,I~particularf try reforms inareas where performanceis relatively easY~rrleasure objectivelyand where the revenues raised or costs saved can make the experiment self-financing.
II

IIIii

Quantitative and qualitative outcome measures can be used. So can peer ratings, as long as ratings are forced to be "on a curve" (that is, not everyone can be rated "excellent").
!II
IIIiII

Remember the principle of the sample: incentives can be based on samples of performance, Especially in an experiment, there is no need for the comprehensive measurement of each and every outcome of each and every action.

Include information from clients.

Empower clients. Seek analogies to market power or joint management In pursuing such reforms, continually think "information and incentives."

• Cultivate political support particularly from unions and foreign donors. The idea of an experiment reduces their worries and involves them in design and evaluation.
II Challenge technical assistance by foreigners. For example, learn by doing rather than attempting comprehensive studies that often end up being inconclusive or unsatisfactory. For example, use TA funds to finance experiments where local experts and even government officials carry out the required "studies" based on the participatory diagnosis of what is already known about problems and possible solutions. II

IilI!

Experiment with user charges and analogies to them such as in-kind contributions, sharing them with employees.
II

Team incentives are often more feasible and desirable than individual incentives,

Incentives include money but also other things: promotions, training, travel, special assignments, transfers, awards, favorable recognition, and simple praise. Even information about how well one is doing turns out to function as an incentive.
!l!I

Privatize creatively. This can mean experimenting with hybrids of public and private sectors working together to provide services.

23

J-v.,r-...oy
Av. d".RoI202 1060~

1Ia.ct!.lM

H.!JNc. :KCINC,
,....2000Ud.
~W~

MACAO

1mA.ZD.

~T~~ R.oo.~~209
CIlW9s... P",""" liP CANADA t",~ lSlA&uL de~ lk~~ ~35E6

~

W~~

1tIo.li1ooi'

~~~

l'WNGAIlY ~bM.rket~ ~Ut17.19

:a.ll11~
ll'mIA M1lod~ ?S1~Ra.r.d h!v.o.b Ud.

NOlIitWAY

w....-~c.-.
i'loo!<~ 1'.0. !b<612S~

~~a:.,

tm~Roo:.cI.
~~K1B3Wll

~.600002
Pt Indl...t.lMii:llid J.do:n So.'<.1bud"" 21)
1'.0.&.<181
~1Q3.20

~00I0'

~

i'~AN ~&o.1<Apr.q

l".o.bN<>.7'.19

m,~

W-54!m ~l.1Mv-tr
S~T_

F D. ~ T~
!..~~ COSTA EICA..,liiW:n;.CUA'm

ntAI!i ~PW~

p._

1957S-S11·

~~ l".o.lbtI.am ~
llIIl\lIDAl) ~

~~ m~A,:rANAMJI. ~~
So.nJ",,~7

~""""""""ts..ppJJ..mAf>"I"OY
4,.5~~

S,_~Ud

TOfIACO,

r~

'OuI:>!ift 2

il9w_~

75~<II~&dHotml~

~

~ ~
~i"I<x>&~ $ull:$ 72C!, Cily~

.fi~W

... d~

y""""",tf..lw..I:Ut'<iUd. P.O. l'l<»i 561155 1"..1 A"", 61560 ~ lilO.Y.t~ P.O.bl3Q56 TdAYiv61130

a:rJ:EU'l.Vom:!E c...tl\e d1'Ai_«t&" ~(CEDA) 04ll:.P.541

10 ~""'mT_2 Al"'" A~ KV. deitr CooII...&_
~ MirtroMmii9

UGAN'DA ~Od. 1Ml'l<>ot,Room4,~~ £'£1.&.<9997/. Piot(~~!I.oo.:l

.

~~~ ~ ~~

~

C-cfAI'f&d~ 6,~Sttw~~ F£). lkoc 1006 N"","", CZfiJCB~

p.o.a

~Ini'oiI,,_~
l!lOO'2J~

~~&M

~ ~~~ ~

m.~31m
OO4!!l7W"""""""1!

rtAI.ll

u.:.-~s..r-uSPA
v;., 0.- D! ~
c.....u"l''''"~lI5:I

1/1

N~W~~

Sim~F"_

Rwt o..~ro.14
lm~ Mtmf ~lA. )$m'!!ook~ P A&.::3t!16 QA'rAlt

~p~

~ ~~

F.Qm...66Il

~U:Ur~l

i~~<
~WI! ~~(I"II4tA

c.-...t~Rood~

~H~

T.-t..~~

PAbST125

Recent World Bank Discussion Papers
No, 237 No. 238 No. 239 No. 240 No. 241 No. 242 No. 243 No. 244 No. 245 No. 246 No. 247 No. 248 No. 249 No. 250 No. 251 No. 252 No. 253 No. 254 No. 255 No, 256 No. 257 No. 258 No. 259 No. 260 No. 261 No. 262 No. 263 No. 264 No. 265 No. 266 No. 267 No. 268 Indonesian Experience with Pinonaal Sector Refon«, Donald P. Hanna Pestidde Poliae: in Developing Countries: Do TIley Encourage Excessive Use? Juman.h Farah Intergovernmental Fiscal Relations in Indonesia: Issues and Riform Options, Anwar Shah and Zia Qureshi in Overexploited Fisheries. Joshua John

Managing Redundancy lnstitutionai

CIJange and the Public Sector in Tiansitional Economies. Salvatore Schiavo-Campo

Africa Can Compete': Export Opportunities and CIUlllengesfor Garments and Home Products in Ihe US. Market. Tyler Biggs, Gail R. Moody, Jan-Hendrik van Leeuwen, and E. Diane White Lberalizing Trade in Services. Bernard Hockman and Pierre Sauve

Women in Higher Education: Progress, Constraints, and Promising Initiatives. K. Subbarao, Laura Raney, Halil Dundar, and Jennifer Haworth VV1lal We Know About Acquisition

if Adult

Literacy: Is There Hope? Helen Abadzi

FQrmuUlting a National Strategy on Information Tedmology: A Case Study oj India. Nagy Hanna Improving the Tranifer and Use

of Agricultural

Information: A Guide to If!formation Technology. Willem Zijp in Sub-Saharan Africa. Marc

Outreach and Sustainability of Six Rural Pinance Institutions Gurgand, Glenn Pederson, and Jacob Yaron

Population and Income Change: Recent Evidence. Allen C. Kelley and Robert M. Schmidt Submission and Evaluation of Proposals for Private Power Generation Projects in Developing CouI1/ries. Edited by Peter A. Cordukes Supply arId Demandfor Finance of Small Enterprises in Ghana. Ernest Aryeetey,Amoah Baah-Nuakoh, Tamara Duggleby, Hemarnala Heuige, and William F. Steel Projedizing Adjustment the Governance Approach 10 Civil Service Refonn: An Environment Assessment for 'preparing a Seaotal Loan in the Cambia. Rogerio F. Pinto with assistance from Angelous J. Mrope

Small Firms Informally Financed: Studies from Banglade~h. Edited by Reazul Islam.]. D.Von Pischke, and J. M. de Waard Indicators for Monitoring P01Jerty Reduction. Soniya Carvalho and Howard White Vio/ence Against Women: The Hidde« Health Burden. Lori L. Heise with Jacqueline Pitanguy and Adrienne Germain
WOfl1m's Health and Nutrition; Making a Difference. Anne Tinker, Patricia Daly, Cynthia Green, Helen Saxenian, Rarna Lakshminarayanan, and Kinin Gill

Impr01ling the Quality of Primary Education in Latin America and the Caribbean: Toward the 21st Century. Lawrence Wolff, Ernesro Schiefelbein, and Jorge Valenzuela How Fast is Fertility Declining in Botswana and Zimbabwe? Policies Ajfeding Fertility and Contraceptive Duncan Thomas and Ityai Muvandi ofTivelvc SJ.lb·Sa/:aran Countries. Susan Scribner

Use: An Assessment

Finandal Systems in Sub-Saharan

Africa: A Comparative

Study. Paul A. Popiel

POFerty Alleviation and Sociai lnvestment Funds: The Latin American Experience. Philip]. Claessner, Kye Woo Lee, Anna Maria Sant' Anna, and Jean-Jacques de St. Antoine Public Policy for the Promotion of Family Farms in Italy: T1Je Experience of the Fund for the Formation of Peasant Pmperty. Eric B. Shearer and Giuseppe Barbero Se!fEmploymel1tjor Arvil V Adams the Unemployed: Experience in OECD and Transitional Ecol1omies. Sandra WilSOll and

Schooling and Cognitive Achicl'ements of Children in Morocco: Can the covernmCl11 Itnpr01le Outcomes? Shahidur R. Khandker.Victor Lavy, and Deon Filmer World Bonk-Financed Projects wilh Community Gita Gopal and Alexandre Marc Seed Systems in Sub-Saharan Participation: Procurement and Disinmemmt Issues.

Africa; Issues and OptioMs. V.Venkatesan

Trade Policy Riform in Developing Countries Since 1985: Review of the Evidence. Judith M. Dean, Seems Desai, and James Riedel Farm Restructuring and Land Tenure in Riforming Socialist Economies: Europe. EUroCOl1SUlt and Centre for World Food Studies Comparative Analysis of Eastern and Central

(Continued on the inside back cover)

Sign up to vote on this title
UsefulNot useful