Policy Directions for the New Economic Model in Latin America

Policy Directions for the New Economic Model in

Latin America

JOSEPH RAMOSUniversity of Chile, Santiago, Chile

Summary. Ð The new economic model (NEM) has been extremely successful in reducing in¯ationand in increasing exports, but quite disappointing in terms of the growth of output andproductivity. It is argued that this calls for a ``second generation'' of reforms, especially centered atthe plant level and its surrounding factor markets, to close the huge productivity gap separating®rms from best practice and so enable the accelerated growth rates typical of successful ``latestarters.'' Policies should emphasize the systematic identi®cation of those technologies andpractices best suited to local conditions and their rapid transfer, adaptation and di�usionthroughout the economy. Ó 2000 Elsevier Science Ltd. All rights reserved.

Key words Ð industrial policy, growth, Latin America

1. INTRODUCTION

As has been well documented in the papers ofthis special issue, in the last 10±15 years a ``neweconomic model'' (NEM) has been put intoplace in Latin America. A major strategicreversal has taken place, from an inward-looking model of industrialization based on thesubstitution of imports (ISI) to a strategyaimed at penetrating large and growing inter-national markets on the basis of specializationand comparative advantage; from an economywith strong government intervention inmarkets to one where resource allocation islargely determined by the interplay of free andunregulated prices; and from one where thepublic sector played the leading role in devel-opment to one where the private sector is thekey agent of dynamism in the economy.

In the earlier model, `ìndustrial policies''played a prominent role in development strat-egy. On the one hand, high and variable tari�protection led to marked di�erences in thetreatment and e�ective priority accordeddi�erent productive activities. On the otherhand, given the control of interest andexchange rates, credit and foreign exchange hadto be rationed administratively. Thus criteriafor such rationing were requiredÐbased oneconomic considerations, such as the expectedimpact on employment, foreign exchange orgrowth generally, or on less noble and moremundane, rent-seeking considerations. Since at

the heart of the ``new economic model'' was thefreeing of markets and the liberalization oftrade, the scope for such traditional `ìndustrialpolicies'' has therefore been necessarily reducedif not fully eliminated.

Hence, two closely related questions come tothe fore. First, and most fundamentally, iswhether indeed there is any longer a role for`ìndustrial policies'' in free and open econo-mies, and if so what type of policy? In the spiritof the papers in this issue, I argue that there is arole for industrial policy though with a markedshift in the choice of instruments and in thefocus of such policy (now termed ``competi-tiveness policies'' to denote this shift in focustoward enhancing productivity growth).

Second, just how important are suchcompetitiveness policies for high growth? Arethey a complement, but a secondary one, to themajor structural reforms currently underway,or, as I argue, are they indeed indispensable ifthe markedly higher growth rates, typical of the``catch-up'' phase of late-starter developingcountries, are to be achieved?

2. FROM THE `ÌNDUSTRIAL POLICIES''OF THE ISI STRATEGY TO THE

``COMPETITIVENESS POLICIES'' OF THEPRESENT 1

In the heyday of Import-Substitution Indus-trialization (ISI), four sets of policies made up

World Development Vol. 28, No. 9, pp. 1703±1717, 2000Ó 2000 Elsevier Science Ltd. All rights reserved

Printed in Great Britain0305-750X/00/$ - see front matter

PII: S0305-750X(00)00042-5www.elsevier.com/locate/worlddev

1703

what was known as ``industrial policy.'' First,at the heart of program design, were ``verticalpolicies,'' that is, those aimed at promotingsectors considered to be of high priority.Typically, manufacturing was favored oversectors with traditional comparative advantage,such as agriculture and mining. Second,preferred sectors bene®ted from a wide varietyof instruments: (a) tari� protection or importquotas on output, while their imported inputscame in with low or no tari�s; (b) preferentialcredits; (c) purchases from public sector ®rms;and (d) the supply of utilities and key inputsfrom public sector ®rms at below market prices.Third, there was a tendency to see the propersequence of sectoral priorities as going from®rst, relatively easy to substitute ®nal consumergoods, to the later substitution of intermediategoods (steel, petrochemicals, paper, etc.), whichwere more technologically demanding,normally requiring continuous productionprocesses (as well as large-scale production),and ®nally, to the domestic production ofcapital goods far more technologically sophis-ticated and demanding. Fourth, there were also``horizontal'' policiesÐfunding for training,preferential credits for small and medium-sizedenterprises (SMEs), and technology policies(aimed especially at building up the techno-logical infrastructure of research and technol-ogy institutes), but these were largelycomplementary to ``vertical'' or sectoral poli-cies, and normally involved signi®cantly fewerresources.

While under ISI the region as a whole grew ata solid rate of 5.6% per year (1945±80), wellabove that of any period before (or after), thestrategy ran into increasing problems as itmoved from consumer goods to intermediategoods and then on to capital goods: the grow-ing limitation of domestic markets too small forminimum e�cient scales of production; insuf-®cient competitive pressure to induce produc-tivity increases; the tendency to overdiversifyproduction rather than to specialize; the ``dead-weight'' costs of administering all of theseprograms. In addition, the ISI strategy tends toact as a disincentive to exporting, since tari�sare, in e�ect, a tax on exports, especiallynontraditional exports.

For all of these reasons by the early 1970s thestrategy of ISI had reached a stage of rapidlydiminishing returns. If this was not immedi-ately noted at the time, it was no doubt due tothe fact that GNP continued to grow at a goodpace in the 1970s, but on the basis of everincreasing, foreign debt-®nanced capitalformation. The growth of total factor produc-tivity in the region came to a halt in 1973 andstagnated thereafter (see Figure 1). 2 Of course,all of this became patently clear with the onsetof the debt crisis, when the collapse of capitalin¯ows put an end to growth in the region inthe 1980s.

It is thus not surprising that the debt crisiswould bring into question if not disrepute theISI model followed since the Great Depressionand would give rise to the NEM based on price

Figure 1. Total factor productivity, 1950±98, Latin America, Best Practice and Asian tigers (index 1950� 100).(Source: ECLAC.)

WORLD DEVELOPMENT1704

stabilization, liberalization, privatization andorientation to international markets. Theresulting freeing of prices and interest rates, theelimination of most subsidies and the sharpreduction in tari� as well as nontari� barriersenormously reduced the scope of ``vertical'' orsectoral industrial policies. Moreover, instru-ments such as public sector purchases of (moreexpensive) domestic products or public sector®rms' sales of key inputs to selected sectors atbelow-market prices were virtually eliminatedboth because of the disrepute in which theseinstruments had fallen and because of the crit-ical need to reduce ®scal de®cits.

Thus the NEM brought two importantchanges in industrial policy. First, and by farthe most important, sectoral or vertical policieshave been downplayed, leaving only horizontalpolicies, such as training, credit to small andmedium-sized enterprises (SMEs), and buildingup the scienti®c and technological infrastruc-ture. 3 Moreover, given the stringent needs ofmost public sectors since the debt crisis, theresources devoted to such programs havetended to decline and at most maintain them-selves. 4 Second, technology policy has movedaway from an almost exclusive reliance onsupply-driven instruments and increasinglytoward market and demand-driven instru-ments. In e�ect, in the past, e�orts to raisedomestic research and development (R&D) to1% of GNP were largely concentrated in direct,public sector funding of R&D at university andtechnological research institutes, with littlerequirement that such work be demanded oreven considered relevant by the private sector.In this sense, then, past technology policieswere supply-driven. To correct this relativeindependence of public sector-®nanced R&Dfrom the needs of the productive sectors, tech-nology policies are becoming increasingly ``de-mand-driven,'' public sector support for suchresearch now being conditioned to the expres-sed interest of private ®rms in the projects inquestion, if not requiring partial ®nancing aswell. 5

These programs are all consistent both withthe old and new model, the di�erence being that``horizontal'' policies tended to be a secondarycomplement of the ISI model, whereas they arethe principal form of industrial policy justi®edunder the new model.

3. THE RESULTS OF THE NEWECONOMIC MODEL

While it is still possibly too early to make afair evaluation of the new economic model, todate the results are mixed (see Table 1). Therehave been two notable successes. First, andmost noteworthy, thanks to the sharp reductionin ®scal de®cits, in¯ation has been broughtdown from an average of 400% per year duringthe 1980s and an unheard of 1200% in 1990 tosome 10% in the present. This latter is but halfof the in¯ation rate during our ``golden era,''1950±80.

Second, thanks to trade liberalization, andnotwithstanding the exchange rate appreciationbrought on by ®nancial liberalization andensuing capital in¯ows, the volume of exportshas grown close to 9% per year in the 1990s, arate four times that of 1950±80 under ISI.Moreover, the growth of nontraditionalexports has been even greater, so that todaysome 50% of Latin American exports arenontraditional products, whereas in 1980 thiswas but 25%.

Yet the results in terms of growth andproductivity have been modest if not mediocre.Economic growth in the 1990s has averaged3.5% per year, above that of the ``lost decade''of the 1980s, but well below the 5.6% per yearof the ISI period which the new model wasmeant to surpass, and far below the above 7.5%annual growth of the late-starting ``Asiantigers.'' Moreover, overall labor productivity inthe region has grown at a most sluggish rate(1% per year), less than one-third the rate of theISI period.

Table 1. Economic evolution of Latin America: 1945±98, Annual rates of growtha

1945±80 1980±90 1990±98

In¯ation rate (%) 20 400� 1200! 10Export growth (volume/year) (%) 2.1 4.4 8.8Growth of GNP (annual) (%) 5.6 1.2 3.5Productivity growth (GNP/worker) (%) 3.1 )1.8 1.0% Poor 50� ! 35 35! 41 41! 38

a Source: ECLAC.

POLICY DIRECTIONS FOR THE NEM 1705

To be sure, the new economic model has notbeen fully implemented in all countries, andseveral countries did not put it into e�ect at alluntil well into the 1990s. Moreover, the modelhad simultaneously to cope with gallopingin¯ation. A fairer assessment would thus haveto discount those disappointing results due tomacroeconomic instability and stabilizationand then take into account the necessary timelags before desired results could be expected.Yet it is especially here in the issue of theappropriate time frame that the issue of themicroeconomic adjustments possibly needed toaccompany structural reforms is critical. Yet isit reasonable to expect 10 years of virtual stag-nation to pass before steadily positive resultsbegin to show up? Does success or failuredepend on just how long we must wait for ®rmsto respond favorably to structural reforms? Is acritical determinant of ®rmsÕ speed of response,the industrial or competitiveness policies thataccompany structural reforms?

The response of ®rms to structural reformshas been uneven. Thus privatized activities(telecommunications, utilities and ®nancialservices) have been among the sectors with themost rapid response to the reforms in terms ofgrowth and productivity. Processed orcommodity manufactures have also beenamong the most rapidly growing export activ-ities, as have export processing zone manufac-tures, especially apparel and electronicassembly products, and automobiles (theselatter bene®ted with sector speci®c measuresand trade policies).

Prominent in all of the above activities havebeen transnational ®rms and domesticconglomerates. On the other hand, far slower interms of growth in output or productivity hasbeen the typical response of the bulk ofdomestic, family-owned ®rms, large or other-wise, especially those outside of the areas oftraditional comparative advantage (Peres &Stumpo, 2000). While many of them haveintroduced important increases in productivity,these have been generally of an organizationalsort, but with the same basic plantÐreducingproduct lines, inventories and layers ofmanagement hierarchies. Important as suchchanges are, they are limited by the fact thatcapital shortages, among other factors, havekept down outlays for new plant. Such mattersare worse, of course, in the SMEs of the region,whose production and productivity have notonly lagged but at times even declined. 6

Thus the response of ®rms to structuralreforms in recent years can best be described interms of the same S shape typical of ®rmsÕspeed in adopting and di�using new technolo-gies (see Figure 2). Given the segmented natureof agentsÕ response, however, it would seemthat not all ®rms are on the same S curve.Transnationals and domestic conglomerates inLatin America are on a rapidly growing one,whereas most other ®rms in the region, not onlySMEs, are on a far more slowly rising one, thusproducing a kink, if you will, in the economy-wide S curve. This would suggest that theproblem of accelerating the favorable responseto reforms (and so, shifting the S curve to the

Figure 2. Percentage of ®rms that modernize over time and speed of response by type of ®rm and sector.


left) is acute, requiring more, not less, work atthe microeconomic level to raise productivityand competitiveness.

The productivity gap separating countries ofthe region with developed countries on thefrontier of technology and best practiceproduction is strong and generalized throughoutLatin American economies, and not restrictedto a few sectors. Total factor productivity levelsin the region are on the order of 40% of those inthe United States (ECLAC, 1995). Laborproductivity di�erentials are, of course, evengreater. For example, in the case of Brazil,these average 27% for all sectors, varying fromlows of 14% and 18% in food retailing and foodprocessing, respectively to a maximum of 88%(steel); and even important modern sectors,such as airlines, telecommunications, bankingand auto assembly register less than half of USproductivity levels (McKinsey Global Institute,1998).

This fact casts doubt as to the relativeimportance of sectoral policies. For much ofthe thrust in favor of sectoral policies in thepast derived from the conviction that the regionwas poor because its output was concentratedin activities characterized by low productivity(e.g., agriculture) or with prospects of slowgrowth in demand in international markets(most raw materials). Yet, to the extent thatsectoral productivity di�erentials are large andwidespread in our economies, sectoral policiesought to play a lesser role than in the past, andso, for some time to come, will be secondary topolicies aimed at more speedily narrowing theproductivity gap throughout Latin Americaneconomies. It is this change in focus that isbehind the move from industrial to competi-tiveness policies, and not merely a new name forthe same old instruments. The issue, in a word,is not so much what is best to produce, but howbest to produce it.

4. NEW POLICY DIRECTIONS: TOWARDA ``SECOND GENERATION'' OF

REFORMS

If the foregoing analysis has been anywherenear the mark, the principal new directions forenhancing accelerated growth in the regioncenter on raising productivity and so, compet-itiveness. Such competitiveness policies, in myestimation, go well beyond the dominantneoliberal prescription of: more privatization,liberalization (now of labor markets) and lower

®scal burdens on the private sector. For stabi-lization, privatization and trade opening up arenecessary but not su�cient conditions forspeedy growth. Moreover, rapid ®nancialliberalization, with insu�cient regulation, hasproven to be one of neoliberalismÕs greatestfailures, resulting in the major banking and®nancial crises that beset the region in recentyears (and now Southeast Asia). Credit is aspecial type of ``good,'' one that must berationed not just by price, as most other goods,but by quantity as well. Credit involves theexchange of a good now for a promise to pay ata future date, thus requiring great con®dence,and so an appropriate regulatory framework if®nancial crises are to be averted. 7 The otherneoliberal proposals are, in my estimation,showing rapidly diminishing returns. For, aswill be presently discussed, though freermarkets may be required in other factormarkets, this is generally secondary to the moreserious failures besetting those marketsÐfail-ures generally not due, in my estimation, togovernment imposed regulations (the standardneoliberal culprit).

If this is the case, then new policy directionsmust emphasize two sets of policies, muchneglected or heretofore played down: ``meso-economic policies,'' or the systematic correctionand strengthening of key markets, especially,factor markets, but at an adequate resourcelevel; and ``microeconomic policies,'' that is,those especially aimed at raising productivity atthe level of the ®rmÐto assist in the systematicidenti®cation of those ``hard'' and ``soft'' tech-nologies and practices best suited to localconditions (factor scarcities and resourceavailability), and their rapid transfer, adapta-tion, assimilation and di�usion throughout theeconomy. 8

(a) Mesoeconomic policies

Four key factor markets require deepeningand strengthening: the markets for physicalcapital, both savings and investment; forhuman capital; for labor; and for foreignexchange. A sample of ®ve policies to improvethese markets follows.

First, raising savings is important for growth,and this can be done by raising private and/orpublic sector savings. A necessary condition forraising private savings is that interest rates bepositive in real terms (a condition that wasoften violated in the past). Hence, recentreforms have eliminated those restrictions


which often led to negative real rates of interest.Nonetheless, experience indicates that once realinterest rates are positive, overall privatesavings (as opposed to ®nancial savings) havenot shown themselves to be signi®cantly sensi-tive to further interest rate hikes. Othermeasures that have proved successful in raisingprivate sector savings include the following two(ECLAC, 1995). One is conditioning publicsector credits, or even subsidies, for housing tofamilies that have complied with a minimumsavings of their own. It has been demonstratedthat low and middle to low-income families arewilling to save far more than they currently doon their own, if such savings can lead to aqualitative jump in their standard of living, aswould be the case if it could provide them withaccess to decent housing otherwise not avail-able for people in their income bracket. Second,there is evidence that private sector savings riseif taxes are shifted from undistributed pro®tson ®rms (thus inducing reinvestment) todistributed pro®ts in the form of dividends toindividuals, for ®rms, it would seem (though itis not unambiguously clear) have a higherpropensity to invest than do individuals.

Savings can also be increased by reducingpublic sector de®cits and/or increasing surplu-ses. How then can public sector savings beincreased? One answer (a neoliberal favorite) isto cut spending. But after a decade of relative®scal austerity, aside from military spending(not always easy politically, above all in aregion, believe it or not, where military spend-ing, of however dubious utility, is in fact low byinternational standards), there is little else ofsigni®cance left to cut. The alternative is,therefore, to raise public sector savings byincreasing taxes.

Despite neoliberal misgivings, there wouldseem to be considerable room for improvementhere (ECLAC, 1998). For the fact is that theaverage tax burden on the private sector is but16% of GNP (compared to over 25% in theOECD). Indeed, the country in Latin Americawith the highest tax burden (increased by twopercentage points by the democratic govern-ment in 1990 and with no harm to growth) is infact, the regionÕs greatest success storyÐChile.There would thus seem to be ample margin forincreasing tax earnings for critical socialinvestments without hindering growth. To besure, emphasis should be, not so much onincreasing tax rates as in increasing the taxbase, by eliminating innumerable exclusions ofa regional or sectoral sort, uniforming rates and

including all forms of income in the tax base(not only salaried income but capital gains,professional incomes, interest and dividendincomes and incomes from rentals). This notonly leads to greater equity among taxpayersbut greatly simpli®es the costs of anti-taxevasion campaigns. 9

A second important area of reducing ®scalde®cits or increasing surpluses is in introducing``actuarial'' accounting into the social securitysystems of the region, and balancing thebudget on actuarial terms. Even those fewcountries that enjoy a current account surplusin their social security systems are in fact inactuarial de®cit. For the combination ofreduced fertility ratesÐalready in placeÐandof the continued extension in life expectancyafter 65, means that by the year 2020 the ratioof pensioners to economically active laborforce will rise by 30%. Thus today we areincurring 30% greater obligations for tomor-row, requiring an increase in contributions orreduction of bene®ts of 30% or some combi-nation of the two. Thus on actuarial account-ing we are already in de®cit right now, and sounder obligation now to increase contributionsand/or reduce bene®ts as of now. Doing sowould lead to current account surpluses(savings) for some time to come, but surplusesneeded to o�set the de®cits on the horizon, inwhose obligation we have already incurred.This would be another indispensable step toraise savings coe�cients.

Second, it is not just a question of increasingsavings, and so investment, but of assuring thatthe savings available for investment is allocatedto the best uses. Only an e�cient system of®nancial intermediation can assure this, andsuch a system is sorely lacking in the region.Because of the costs and uncertainties ofproject evaluation and the problems of seem-ingly inadequate guarantees, the market forlong-term capital in the region is virtuallynonexistent, not just for risk or venture capital,but for restructuring and even expansion.Hence, except for the very few ®rms traded onthe stock exchange, most long-term invest-mentÐfor restructuring or expansionÐis self-®nanced, so that investment funds are allocatedin accordance with past and not expected ratesof return. 10 How can capital move at a fasterpace 11 into those sectors with promiseÐagri-culture, mining, and activities supplying,processing or servicing our key naturalresourcesÐwhen these ®rms are, for the mostpart, regardless of size, family-run ®rms, not


listed on the stock exchange? How can theseeconomies increase the amounts of ®nanceavailable for the expansion of those promising®rms, already capable of competing interna-tionally, as well as ®nance the restructuring ofthose ®rms in ``challenged'' import competingactivities, whose restructuring would renderthem competitive?

It would seem that these activities are idealfor the massive introduction of innovative®nancial instruments, still incipient in theregion, such as leasing, factoring and securiti-zation (in this latter case of bank loans and/orsupplier credits and their subsequent ``packag-ing'' and resale in specially developed second-ary markets so as to diversify risk andmassively draw resources into such activities).It is thus not a question of freeing capitalmarkets, but of developing them. For in theabsence of mechanisms to intermediateresources into such promising activities, theirrequired expansion or modernization will bedangerously put o�.

Third, though investment in human capital isrecognized as a key to growth (and socialequity), notwithstanding the annual social ratesof return of 15±20% characteristic of thisactivity, there is virtually no private lending forinvestment in education or training. Thus, whatinvestment takes place is subject to the ups anddowns of the public sectorÕs budgetary situationor the income of individual families. Therefore,talent and the prospects of the investment playmuch too little a role. Given the importanceof human capital in growth (and equity), animportant source of growth is thus beingsquandered. Important as non-negative realinterest rates are for savings and investment,freeing interest rates does not su�ce. The realproblem lies elsewhere. What seems to bedecisive is the guarantee such a person canprovide, which, in the absence of slavery (!) istenuous at best.

A proposal to overcome this problem(ECLAC, 1992) is to take advantage of the factthat all workers will end up being creditors oftheir social security programs (be they old style``pay as you go'' pension programs or individ-ual capitalisation programs, increasingly morefrequent in this day and age). The idea is thatany prospective student could use his futurepension or pension rights as a guarantee for atraining loan from a bank, in exchange forwhich his interest and principal will be deduc-ted automatically from his future earnings inaddition to his regular social security payments

until the loan is fully repaid. To cover the casesof unemployment, labor force withdrawal,informality or leaving the country, all recipientswould be required to have a guarantor (typi-cally a relative) in the social security system,from whom these monthly payments would bewithdrawn in case the debtor was not repayingthe loan. It is neat, tidy, and virtually risk free,and so, would be quite attractive to privatelenders. It would have the merit of renderinguseful today, as a guarantee, today's obligationof forced savings, which all salaried workersmust now incur, and which is currently of noworth until retirement. It thus redresses a majormarket failure, opening access to capitalmarkets to the mass of the population withpotential, and it does so in a truly market-friendly way. 12

Fourth, much is made, and correctly, of theneed to reform the labor market. But, thefavored neoliberal reform, increased ¯exibilityand deregulation (lower minimum wages,weaker unions, reduced entry barriers, lowerseverance pay), while reasonable in severalcontexts 13 overlooks the principal source ofunderemployment and low labor productivityin Latin American economies. The latter stemsfrom a nonpublicly imposed practice, butnonetheless a real institution: paying wages forthe most part according to hours worked. It hasbeen proposed that ``participatory salaries,'' beintroduced where an important component ofwage income (say 15±20%) is variable inaccordance with the performance of the ®rm orthe department or section in which the workerbe found. This tends to raise productivity sinceit more closely ties the workerÕs interests withthe destiny of the ®rm, reduces labor±manage-ment con¯icts, and inverts the process of socialcontrol (workers who go beyond minimumrequirements are today criticized by their peersfor tending to pressure up standards whereasunder participatory wages, these would bepraised for providing a bene®t for them all, notjust management).

Even more important, participatory wagestend to increase productive employment.During a recession, such ®rms are more apt tolower their prices to maintain sales andproduction (and so employment), for theyknow that their labor costs, at least the variablecomponent thereof, will automatically decline.Fixed wage ®rms, on the contrary, under theobligation to maintain their contracted wage,have far less leeway to face such a drop indemand by lowering pricesÐfor their costs


remain ®xedÐand thus such ®rms tend to holdprices in the face of recession, while cuttingback production, and so, employment. Indeed,this is not just theory. 14 The success of Japa-nese ®rms, and more recently, those of Taiwanand South KoreaÐall of whom tend to payparticipatory wages to hold down unemploy-ment in the face of serious prolonged stagna-tion or sharp recession, speaks of the practicalmerits of this form of wage payment. Forsimilar reasons, incidentally, in periods ofboom, participatory wage ®rms will tend toexpand output and employment rather thanprices and pro®t margins, to avoid givingunnecessary cash ``windfalls'' to workers, thustending to stabilize prices while expandingemployment.

If they are so good, why does not everyoneuse participatory wages? On the one hand, aswith all innovations, the ®rst to introduce sucha contract will probably need to pay a premiumto overcome the uncertainties and suspicion itwill engender. Thus the innovator is likely to bepenalized rather than rewarded. On the otherhand, this type of contract generates externali-ties: above all, it bene®ts workers that all other®rms pay participatory wages, and one have a®xed wage contract, for the others will assure``full employment,'' while oneÕs own ®xed wagecontract will reduce income variability and risk.Thus there is a public policy case to be madefor promoting the introduction of suchcontracts.

Fifth, although neoliberals argue that ISIwas always a mistake, and structuralists arguethat ISI made sense in its day to start o� aneeded process of industrialization, theprotection of domestic markets was nevermeant to be but a ®rst stage in the developmentof an internationally competitive industrialbase. There is a wide consensus today that thefuture lies in the penetration of foreignmarkets, and that consequently, trade openingis needed.

What is at issue today is not so much howfast trade should have been opened upÐthat iswater under the bridge nowÐbut whether tradepolicy should be neutral, as neoliberals arewont to favor, or whether it should have atransitory pro-export bias as many neostructu-ralists argue (Agosin, 1993; Macario, 1995,2000). For although tari�s were the instrumentsuited to the ISI strategy, many neostructural-ists argue that export promotion is now thepolicy suited to a strategy of internationaliza-tion.

The infant industry of today is no longer, asin the past, a matter of the acquisition of theskills required to produce e�ciently, but of thepenetration of international markets. Herein liethe skills we are most lacking, and these have a``public nature'' to them. In e�ect, the ®rstproducers to discover that a new (nontradi-tional) product can be successfully produced inour country and sold abroad competitively atinternational standards of quality areSchumpeterian innovators, who bene®t otherproducers with this critical information, whichthey cannot themselves fully appropriate. Toovercome the otherwise resulting underinvest-ment in such a critical activity, policy shouldbene®t these pioneer nontraditional exportersfor the costs they incurred and the bene®ts theyare bringing to other domestic producers. Acase for transitory subsidies can thus be easilymade to encourage such penetration of foreignmarkets with nontraditional exports, until the``herd'' comes in (who, of course, should notreceive it), and all the more so in this criticaljuncture of globalization and internationaliza-tion. A simple instrument and e�ective way toprovide such an incentive, without having toprede®ne what is a nontraditional export, is toprovide a special drawback or subsidy, sayequivalent to 10% of sales, to all exporters ofproducts whose total national exports are lessthan say $10 million (and diminishing to zeroonce the exports exceed $20 million, when the``herd'' can be presumed to be doing the bulk ofthe exporting). 15

Similarly, the case can be made that a ®rmthat develops a new market for a traditionalexportÐsay, manages to penetrate the Japa-nese market with this productÐeven thoughthis export by now is traditional, warrants ananalogous subsidy, for the penetration of a newmarket is quite demanding, and will bringenormous bene®t to all who follow thereafter.Again this is a Schumpeterian innovatorwarranting a special prize for his e�ort and``public'' bene®t. To encourage this sort ofcritical activity again a transitory subsidy can bejusti®ed; 16 say 5±10% of sales to that market,until that market imports more than say US$5million of the product in question.

While according to World Trade Organisa-tion (WTO) regulations most export subsidiesmust be eliminated as of the year 2003, thesame e�ect as above can be achieved withWTO-friendly instruments. For example, andsimply for illustrative purposes, such pioneerexporting activities can be provided with credits


at some one or two points above LIBOR, andit will be WTO friendly, for it simply putsexportersÕ ®nancial inputs at internationalprices, a permissible policy. Yet in all LatinAmerican countries a credit that is only a pointor two above LIBOR is a huge incentive, giventhat domestic rates are typically double that.

Finally, the region must increase severaltimes over its expenditures on trade missionsand attendance in international trade fairs. Forexample, Japan spends US$5 per capita insupporting trade missions. This implies that aLatin American country of intermediate size,such as Chile, should spend $75 million peryear in this function, and, given the fact thatexports weigh in at twice the proportion ofJapan (30% of GNP versus Japan's 15%),possibly up to $150 million! Whereas Brazil,using the same parameters, should spend closeto $1 billion on this alone. The region'sexpenditures in this area are, in fact, but afraction of this, notwithstanding the fact thattoday we consider exporting central.

(b) Microeconomic policies

A second set of policies, even more neglectedthan the former and hence warranting all themore attention, are microeconomic policiesaimed at raising ®rm-level productivityÐthrough more up-to-date ``hard'' technology(more and better equipment) or ``soft'' tech-nology (better management practices, leanerproduction, greater quality control, ``just-in-time'' production, etc.). Yet, since ®rm-levelproductivity would seem to be the soleresponsibility of the entrepreneur, policyemphases currently as well as in the past havefocused on the macroeconomic and ``mesoeco-nomic'' levels rather than on the ``microeco-nomic'' or ®rm level.

Public policy at this level has thus been largelylimited to the building up of the scienti®c andtechnological infrastructure, but this is only partof what needs to be done. Moreover, it tends todraw attention to raising own R&D, which, atour stage of development, is not yet the centralproblem that it is often made out to be. 17

There is, in fact, a critical and much neglec-ted role for public policy at the microeconomiclevel. At the region's present stage of develop-ment, rapid and continuous productivity gainsdepend far more on the rapid acquisition anddi�usion of technology than in raising ownR&D. There are, however, important costsassociated with the search and identi®cation of

that modern technology most suited to localconditions. Pro®table as such search be, theretends to be considerable underinvestment insame, because the ideal strategy for each indi-vidual ®rm is to let others incur the searchcosts, while one quickly imitates and soproduces but at even lower costs. Since thefruits of systematic search are not themselvesinnovations, patent protection is both unavail-able and undesirable as an instrument. Hence, astrong case can be made for public policy tospeed the massive and systematic search for,transfer and rapid di�usion of those technolo-gies and management practices best suited todomestic conditions; and not just in one orseveral sectors but, given the huge, economy-wide productivity gap, massively and across theboard, irrespective of sector.

Technology can be acquired by research andinvestigation, by reverse engineering and byimitation, by visiting and seeing. It can comeembodied in machinery, blueprints and paten-ted processes, or be acquired tacitly by ``learn-ing by doing'' or by interaction with others,both those interested in our acquiring theirgoods (suppliers of machinery and intermediateinputs) as well as those interested in oursupplying them with inputs to their precisespeci®cations. Moreover since learning iscumulative, it is subject to important economiesof scale and agglomeration. Finally, even whereits sale or licensing is feasible, it is bogged downby the information breakdown intrinsic to suchmarkets: to wit, the purchaser does not knowhow much the knowledge is really worth to himuntil he is told it, but once told, it is necessarilyworth less. In short, knowledge is a nonrivalgood, whose bene®ts are not solely limited tothose who undertake the costs of developing it.It is characterized by uncontrollable spilloversand externalities, for which the price system,however attractive it otherwise might be for theallocation of resources, is particularly poorlysuited.

Indeed, even more than our limited relianceon sectoral policies, if one group of policies setsLatin America apart from the Southeast Asiantigers, it is our neglect of the whole host ofpolicies pursued by them to accelerate thetransfer of technology. Kim (1997) has addeddetail to just what much of this entailed inSoutheast AsiaÕs quest to achieve creative imi-tation. In e�ect, informal, nonmarket mediatedmechanisms proved decisive in acquiring tech-nology. Most transfer of technology is rela-tively inexpensive if there is local reverse


engineering capacity. This depends amongother factors on the capacity to be a fast``strategic follower'': to learn by watching andby visiting. This kind of learning entails, amongothers, technological surveillance, the hiring offoreign specialists; the use of technical assis-tance from foreign ®rms; attendance at inter-national trade shows; the invitation ofinternational suppliers to visit the country; thesending of engineers to foreign universities; themanning of technological missions abroad;attendance at international professional meet-ings; links to foreign research institutes; tech-nological assistance centers for SMEs; andsectoral research centers for the absorptionand adaptation of key technologies, generic andotherwise (this latter de®nitely being a casewhere a solid scienti®c and technologicalinfrastructure is an important prerequisite).

There are three policies, that are especiallylacking in the region in this regard and could beexpected to have a very signi®cant impact atthis ``microeconomic'' level. 18

First, it is insu�ciently appreciated in theregion just how important the systematicpromotion of technological licensing can be atits current stage of development, as it was inJapan and Korea at this stage of their devel-opment. Recent studies (Agosin & Saavedra,1998) have established, for example, that foryears Japan paid out 0.25% of GNP in royal-ties, whereas Brazil and Chile pay out but0.05% to 0.08%. This is a virtual order ofmagnitude di�erence in an area in which LatinAmerica still seems tied to its past, more apt toconsidering that the proper policy is to limitsuch payments rather than to accelerate andexploit fully the potential in licensing. Clearlythis is a throwback to the old ISI days, whenFDI and technology licensing were looked atwith suspicion, as instruments of exploitationby foreigners. Latin America has changed withrespect to FDI and transnationals, but seems tobe still lagging insofar as licensing goes. Muchremains to be done here.

Moreover, licenses need not be taken out justwith the leading international ®rms. There areany number of less well known, but strong®rms, that possess important technologies, butwho simply do not have experience in interna-tional markets. Niosi (1997) makes the point inarguing for technology tracking institutes, thatwhereas Asian ®rms come to Canada knowingwhat they want, and where alternative tech-nology suppliers exist, what they o�er andunder what conditions, Latin American ®rms

are laid back and for the most part areapproached by and receive visits from Cana-dian sellers of technology o�ering their wares.The ®rst obviously know what they want andwhat alternatives exist; whereas the LatinAmericans not only are at the mercy of thegood will of the technology providers whocome their way, but really are not activelyengaged in taking it apart and knowing why itis worthwhile; they just take it passively. This ishardly an active policy of technology acquisi-tion and di�usion.

Second, for the most part Latin American®rms lack systematic e�orts to acquire andtransfer technology through the cheapestmeans available, learning by visiting andobserving. One of the most successful compo-nents of the Marshall Plan for the reconstruc-tion of Europe is its less well-known technicalassistance program. The idea was that sinceEurope had to reconstruct, rather than repro-duce the plants in use in 1939, why not recon-struct with best practice techniques? The ideawas to do ``industrial extension in reverse'':rather than bring experts to Europe and tellthem how to do it, take European entrepre-neurs, managers, engineers, technicians to seebest practice plants of similar scale and condi-tions in what was then the sole center of bestpractice, the United States. For six weeks suchgroups (typically 20 per subsector) visited andstudied six best practice plants (one per week)and scrutinized them in detailÐthe equipment,the process, the organization of production,layouts, inventory and quality control, laborrelations, marketing techniques, etc.Ðand thenwrote a report (today this would be accompa-nied by videos) on what seemed the bestcombination of practices most suited forconditions in their countries. Upon their returneach member was required to spread the newsto 10 other ®rms. It is estimated that the resultsof such missions raised productivity by 35±50%without any signi®cant increases in invest-ment. 19 It was simply the fruit of a systematicstudy of the range of best practices, and thediscussion and interaction among the inter-ested parties with their developed countriesÕcounterparts.

It is, of course, well known that before that,Japan ®rst and later, Korea, did much of thesame on their own with stupendous results.Similar results could be expected in LatinAmerica. It is thus proposed that Latin Amer-ican countries organize technology missionsabroad (15±20 members subsector) to visit best


practice plants in the world (now not just theUnited States, but Europe, Japan or all three asbe most appropriate to the industry in ques-tion), again for 6±8 weeks, and videoing as wellas studying the practices, so as better be able todi�use the ®ndings at home. The governmentwould help make the contacts, organize andco®nance such missionsÐincluding possiblyattendance at trade fairs.

If the government were to cover travel andper diem and member ®rms cover salaries, thegovernment component of such missions wouldbe of the order of $15,000 per visitor or$300,000 per mission of 20. Fifty such missionsa year could be ®elded at a mere $15 million ayear, signi®cantly raising the total factorproductivity of the investments of the 1000participating ®rms as well as that of the possi-bly 5±10,000 ®rms indirectly bene®ted throughthe postvisit di�usion in the order of 35±50%,were it to replicate the experience of theMarshall Plan. 20 For a medium-sized coun-tryÐsuch as ChileÐa program on this scalewould bene®t directly or indirectly all of itsexporting ®rms in a single year, and wouldextend similar bene®ts to the bulk of potentialexporters in ®ve years or so. This is a programthat is massive, as required, is relatively inex-pensive, and has an extraordinarily high rate ofreturn, not just in terms of the resourcesdirectly granted it (which are, in fact few), butvery especially in strongly raising the produc-tivity of new investment (and this latter is large,of the order of 20% of GNP). Obviously, largercountries, such as Mexico and Brazil wouldneed multiply their e�orts, but these wouldrepresent a trivial percentage of resources (ofthe order of 0.2% of GNP per year). 21

Third, it would seem that the region is ripefor the introduction of ``model'' plants for eachsector and prevalent scale of production. Forthen the costs of systematic search could beshared, while di�usion could be inexpensiveand rapid. The drawback in such a proposal isthat the costs of setting up model plantsÐtherebeing so many sectors and di�erent scales ofproduction, each with its own idiosyncrasiesÐcould be quite high. Thus it is suggested thatthe government create ``virtual model plants,''that is to say, rather than building them, itshould consider sponsoring (and ®nancing)competitions for the design of sameÐeachdesign to include not only the technology andprocess considered best suited for that scale ofproduction, but the market niche and quality itshould aim for, the layout and organization of

production considered ideal etc. The selectionof the winners would be made by a panel ofsectoral leaders and specialists, and the winners(or top three) would earn important prizes, saysumming $0.5 million per virtual model plant.Awards could be made to 30 subsectors peryear at a relatively low cost and the details ofeach ``virtual plant'' would be available to allinterested parties at no or minimal costs, so asto accelerate di�usion.

(c) The resource e�ort

Before closing, a word must be saidconcerning the order of magnitude of therequired e�ort in such competitiveness policies.Otherwise, we risk generating a ``shopping list''of programs with no sense of impact or ofpriorities. Most countries have policies in manyof these directions, especially at the mesoeco-nomic level. Yet the resources so dedicated tendto be meager, given the needs and prospects. Ine�ect, if the annual social rate of return of suchprograms were to average a very solid 30% (asis estimated by Evenson & Westphal, 1995), forevery 3% of GNP invested in such programs(including those private resources mobilised bypublic funds), total factor productivity (TFP)would grow an additional 1% per year. To raiseTFP growth to 2.5% per year would thusrequire mobilizing 7.5% of GNP. In point offact, however, it can be estimated that currentexpenditures in such meso and microeconomicpolicies are of the order of 2±2.5% of GNPÐless than a third of needs. 22

Second, we must be aware of the targetuniverse which microeconomic policies aim toreach in order to raise national productivity.Just for the sake of establishing orders ofmagnitude, there are close to 100,000 ®rms inthe formal sector alone in Chile, not to mentionsome 500,000 microenterprises. If we use thecapacity to export as an approximation oftechnological modernization and internationalcompetitiveness, of these 100,000 ®rms, that arethe ultimate object of competitiveness policies,only 6,000 export. Thus the e�ectiveness ofcompetitiveness policies must be gauged in howmany years it takes to reach and signi®cantlyupgrade and modernize the remaining 94% offormal sector ®rms. It is thus obvious that anye�ort that reaches and signi®cantly impacts lessthan, say, 5% of formal sector ®rms per year,will be well below needs and potential. More-over, given the relatively modest resourcesassigned to such meso and microeconomic


policies, it is small wonder that the number of®rms signi®cantly a�ected by most of ourcurrent programs is but a fraction of this.

5. CONCLUDING REMARKS

Though any number of instruments of thepast associated with ``industrial policies''Ðsuchas high and varying tari�s, preferential access tocredit, imported inputs and public sector utili-ties, etc.Ðhave been rejected as misplaced orhad their scope reduced by the emergence of the``new economic model'' in Latin America, it isthe thesis of this paper that a new set of policiesis required, rather than no policies, if the highgrowth rates characteristic of all successful late-developing countries is to be attained in theregion. For at the heart of successful ``catch-up''is skipping stages of development and movingrapidly toward the technological productionfrontier. But this is not an automatic process.Rather, it requires the implementation ofproductivity enhancing policies to close the``technological gap.'' Such ``competitiveness''policies shift the emphasis from what best toproduce (the sectoral policies of the past) to howbest and most e�ciently produce it.

While the reforms of the new economicmodel have succeeded in dramatically lowering

in¯ation and markedly increasing exports, thegrowth of output as well of productivity at theeconomy-wide level has been quite disappoint-ing, even below that under ISI. It is thus thethesis of this paper that such acceleratedgrowth requires a ``second generation ofreforms,'' focusing on sharply raising produc-tivity through policies at the much neglectedmeso and microeconomic levels.

Such neglect can in part be justi®ed by thedominant role played by macroeconomicdisequilibria after the debt crisis. But now thatthese disequilibria have been brought largelyunder control in most countries, the time is ripeto move down to the meso and microeconomiclevel. I suspect that the neglect of meso andmicroeconomic policies has also been due to thebiases and assumptions of the ``WashingtonConsensus.'' Thus this paper has made a speciale�ort to indicate, through concrete examples, anew policy agenda that might be formulated atthe meso and microeconomic levels, beyond thelimits set by that ``Consensus.''

Finally, there is a problem of scale as wellas of approach and adequate instrumentdesign. Anything less is to give lip service to themeso and microeconomic policies needed fordevelopment, but to go on relying, in fact, onglobal reforms and no more, whatever therhetoric.

NOTES

1. For extensive treatment of the old and new policy

regimes, see Peres (1993, 1997).

2. By contrast, the productivity of countries on the

technological frontier grew three times faster, so that in

the last 50 years Latin America has lost ground vis �a vis

the countries of the OECD. Conversely, the productivity

of the Asian tigers grew ®ve times faster than Latin

America's (double even that of our ``golden era,'' 1950±

73), and almost one percentage point per year faster than

that of the OECD, thus strongly reducing the produc-

tivity gap with countries on the frontier.

3. To be sure, there are important exceptions, among

the more notable being the preferential treatment

accorded the automobile sector in most economies.

Then too, Costa Rica successfully landed INTEL's

fourth chip plant in the world with a $500 million

investment through direct negotiations on sector speci®c

bene®ts. Less successful, however, was Brazil's attempt

to develop its own computer industry, a program which

ultimately had to be abandoned because of its price and

quality costs for Brazilian ®rms forced to purchase

locally.

4. Only one important addition has been made to

existing horizontal policies. This is the case of export

promotion policies in increasingly open and globalized

economies. Thus public sector resources have been

provided to help new exporters enter foreign markets

as well as help traditional exports penetrate new and

di�erent international markets.

5. While this assures much greater linkage between

research and production, such ®nancing has the draw-

back that much of what is now called research is

operational and ®rm-speci®c, thus minimizing work of a

``public good'' nature in favor of a private, or at most a

limited group, good.

6. Such results are far from peculiar to Latin America.

Though productivity is far more homogeneous by ®rm


size in the United States than in Latin America,

nevertheless the adoption of new technologies in the

United States is greatly a�ected by size. Small ®rms (less

than 100 workers) adopt at a rate 1/3 to 2/3 that of

medium ®rms (100±499 workers) and these in turn adopt

at rates 1/3 to 1/2 those of large ®rms (above 500

workers). Moreover, these adoption rates, in turn are 1/3

those of Japan, where, presumably, because of closer

supplier relations, the investment in the transfer of

technology is greater, and so faster. The point is that

adoption rates and the speed of response are di�eren-

tiated among ®rms and highly sensitive to policy

inducements. See Shapira (1990).

7. In addition to appropriate reserve requirements for

di�ering degrees of risk and transparency in the deter-

mination of nonperforming loans, this also entails,

among other things, strict lending limits to any one

sector, ®rm or conglomerate, with limits, if not prohibi-

tion, of the ``triangulation'' of funds to circumvent such

regulations; as well as limits on loans to ®rms belonging

to the same owner or conglomerate as the banks.

8. I am expressly leaving out the issue of selective

sectoral policies, because of space limitations. But my

own thoughtsÐespecially in favor of ``soft'' sectoral

policiesÐare in print (Ramos, 1998), and strongly

espouse some form of consensual and strategic planning

in this direction.

9. It is worth noting that by measures of this sort in the

mid-1970s Chile, among others, was able to increase tax

intakes by a full four percentage points of GNP, without

increasing tax rates!

10. This problem is probably even more acute with our

SMEs. Yet, my point is that this problem a�ects all but

the handful of ®rms traded in the stock exchangeÐthose

responsible for the bulk of value-added.

11. Much capital has moved into such sectors, but not

at the pace required. By and large it has gone in as direct

investment, where property and ®nance go together, but

there continues to be a dearth of ®nance available for on-

going ®rms, which constitute the bulk of ®rms in these

activities, and a large part of value-added.

12. This proposal does not deal with the issue of the

quality and relevance of education, a truly central issue,

but one that would require another paper, if not a book,

to deal with properly.

13. Who can dispute that some union power is exces-

siveÐpossibly copper workers in Chile, petroleum

workers in Mexico and Venezuela, some utility workers

elsewhere, and stevedores generally? Yet the extent of

unionization in Latin America is, outside of Argentina,

relatively low, below 20%, and the bulk of unions are

weak, not strong as in the few above cited cases. As for

severance pay, the typical amount (one month per year of

service) is not unreasonably costly for economies gener-

ally lacking in unemployment insurance. Moreover, just

as one would expect, to the extent that such severance

pay is perceived as a form of unemployment insurance,

albeit imperfect, its costs are largely absorbed not by the

®rm but by the worker in the form of lower wages, so

that it does not signi®cantly a�ect the ®nal costs of labor

to the ®rm! Finally, the level of minimum wages in real

terms is for the most part quite low, cited by ideologues,

but rarely by ®rms, as a limitation to hiring.

14. Weitzman (1985) worked out the theoretical attri-

butes of this model. The empirical evidence in favor of the

practice can be found in Blinder (1990). Evidence (favor-

able) of the occasional introduction of such schemes in

Latin America is to be found in ECLAC (1992).

15. Chile put in place just such an instrument in the

aftermath of the debt crisis and it was eminently

successful in raising nontraditional exports at a very

rapid rate, the share of manufactures (including second,

but not ®rst processing, of its natural resources) today

equalling 25% of its exports.

16. As with patents and other inducements to innova-

tion, this should be necessarily transitory, for only the

initial innovator or innovators warrant the bene®t, and

not those who follow suit thereafter.

17. It is true that investment in R&D in Latin America

(at most 0.8% of GNP) is well below that of developed

countries (2.5% of GNP). Yet, what is truly pertinent is

how Latin American e�orts compare to those of

successful late-starters, such as the ``Asian tigers,'' at

similar stages of development. Latin AmericaÕs e�orts in

R&D are in fact not signi®cantly below those of the

``tigers'' when they were at our current, relatively low

level of per capita income (as late as the mid-1980s,

Korea spent but 0.9% of GNP in R&D). What is notable

is that their e�orts, as be®ts countries then in the early

stages of catch-up, were heavily focused on accelerating

the transfer of technology through instruments designed

to promote creative imitation, instruments which for the

most part have but token presence in Latin America

today: technological missions abroad, reverse engineer-

ing, the heavy promotion of licensing (at a rate 3±5 times

current use), technological surveillance programs, etc.

18. It goes without saying that the three that follow are

my sense of those most needed. Yet the literature is rich


in further suggestions, including, among others, bench-

marking, industrial extension services, subcontracting,

export processing zones, networking, utility patents,

incentives, as in Singapore, to the ®rst ®rms to introduce

processes and technologies new to the country, etc. See

in this regard, for example, Dahlman (1994), Nadvi and

Schmitz (1994), Lall (1990) and Shapira (1990).

19. For an exposition and evaluation of this program

by two of its original organizers, see Silberman and

Weiss (1992). I might add that the intellectual author of

this proposal for modern times is Carl Dahlman of the

World Bank, at whose behest the above study was

commissioned. Dahlman proposed this program for the

ex-Soviet Union. The idea is splendid, but Latin

America is far better suited for its implementation for,

unlike Russia, the region has in place and in working

condition the principal institutional prerequisites of a

market economy.

20. This, of course, is but a ballpark ®gure. The e�ect

could be lower to the extent that Latin America's

absorptive capacity as of today is much less than that of

postwar Europe. On the other hand, since the impact

also depends on how great a productivity gap exists, the

greater the current gap, the greater the e�ect of

modernization. This explains why ``backward'' countries

such as China today and Korea in the mid-1950s, with a

far lower absorptive capacity than postwar Europe,

exhibited far greater productivity growth upon modern-

izing than did postwar Europe. This is, of course, but a

re¯ection of ``catch-up'' and the single great advantage

of late-developing countries.

21. Preliminary results in Chile are quite positive. I am

also grati®ed to see that Williamson (1997) has included

this recommendation in his latest updating of the

``Washington Consensus.''

22. A rough estimate can be made on the basis of the

following: R&D (public and private) averages no more

than 0.8% of GNP in the region; investment in training is

likely to be less than 1% of GNP, for the goal of most

training programs is to invest 1% of the wage bill (or

1/2% of GNP in direct expenditures, up to 1% if we

include as well, as we should, the opportunity cost of

labor); ®nancing programs for SMEs, even in Mexico,

which has among the best programs in the region

(Kagami, 1995), account for only 5% of SMEs ®xed

capital investment (less than 1% of GNP), so that the

subsidized component of such programs are not likely to

amount to more than 0.33% of GNP; and public

expenditure on export promotion programs, for example,

in Chile, which has one of the region's most active

programs in this regard, amounts to less than 0.25%

of GNP (D�õaz, 1998). Altogether, then, it can be

estimated that such programs currently account for less

than 2.5% of GNP.

REFERENCES

Agosin, M. (1993). Pol�õtica comercial en los pa�õsesdin�amicos de Asia: aplicaciones a Am�erica Latina.Desarrollo Econ�omico, (October±December).

Agosin, M., & Saavedra, N. (1998). Sistemas nacionalesde innovaci�on: qu�e puede Am�erica Latina aprender deJap�on. Santiago: Dolmen.

Blinder, A. (1990). Paying for productivity:a look at theevidence. Washington, DC: The Brookings Institu-tion.

Dahlman, C. (1994). New elements in industrial compet-itiveness: implications for developing economies. InC. Bradford, The new paradigm of systemic compet-itiveness: toward more integrated policies in LatinAmerica. Paris: OECD.

D�õaz, A. (1998). Apuntes para una estrategia m�asavanzada de desarrollo. Mimeo. Santiago: Subsecre-tar�õa de Econom�õa de Chile, October.

ECLAC, Economic Commission for Latin America andthe Caribbean. (1992). Social equity and changingproduction patterns: an integrated approach. Santi-ago, Chile: ECLAC.

ECLAC. (1995). Latin America and the Caribbean:policies to improve linkages with the global economy.Santiago, Chile: ECLAC.

ECLAC. (1998). The ®scal covenant: strengths, weak-nesses and challenges. Santiago, Chile: ECLAC.

Evenson, R., & Westphal, L. (1995). Technologicalchange and technology strategy. In J. Behrman, & T.Srinivasan, Handbook of development economics, vol.3A. Amsterdam: Elsevier.

Kagami, M. (1995). The voice of East Asia: developmentimplications for Latin America. Tokyo: Institute ofDeveloping Economies.

Kim, L. (1997). Imitation to innovation. Cambridge, MA:Harvard University Press.

Lall, S. (1990). Building industrial competitiveness indeveloping countries. Paris: OECD.

Macario, C. (1995). Am�erica Latina: Competitividad ypol�õticas de formaci�on de exportaciones. ComercioExterior (March).

Macario, C. (2000). The behaviour of manufacturing®rms under the new economic model. World Devel-opment, this issue.

McKinsey Global Institute. (1998). ProductivityÐthekey to an accelerated development path for Brazil.Washington, DC: Sao Paulo.

Nadvi, K., & Schmitz, H. (1994). Industrial clusters inless developed countries: review of experiences.Discussion Paper no. 139. Institute of DevelopingStudies, University of Sussex, Brighton, UK.

Niosi, J. (1997). Industrial policies in Canada. Paperpresented at the International Seminar on Successful


Industrial Competitiveness Policy Experiences: Les-sons for Latin America and the Caribbean inECLAC, Santiago, 9±10 December.

Peres, W. (1993). Industrial policy: where do we stand?CEPAL REVIEW (December).

Peres, W. (1997). Pol�õticas de competitividad industrial.Am�erica Latina y el Caribe en los a~nos noventa.Mexico: Siglo XXI.

Peres, W., & Stumpo, G. (2000). The behaviour ofSMEs and large industry groups under the neweconomic model. World Development, 28 (9), thisissue.

Ramos, J. (1998). On the development of clustersaround our rich natural resource base. CEPALREVIEW (December).

Shapira, P. (1990). Modernizing manufacturing. Wash-ington, DC: Economic Policy institute.

Silberman, J., & Weiss, C. (1992). Restructuringfor productivity: the technical assistance programof the Marshall Plan as a precedent for the formerSoviet Union. Mimeo. Washington, DC: WorldBank.

Weitzman, M. (1985). The simple macroeconomics ofpro®t sharing. American Economic Review (Decem-ber).

Williamson, J. (1997). Washington consensus revis-ited. In L. Emmerij, Economic and social developmentinto the XXI century. Interamerican DevelopmentBank. Washington, DC: Johns Hopkins UniversityPress.


Policy Directions for the New Economic Model in Latin America

Documents

Transcript of Policy Directions for the New Economic Model in Latin America