Policy, Research, and External Affairs WORKING PAPERS Macroeconomic Adjustment and Growth Country Econorrlic- PDp-rtmnnt The World Bank October 1991 WPS 788 How Expectations Affect Reform Dynamics in Developing Countries I I CLAI .AJ Li(av ,I Bold moves were effective in trade reform in Chile, Turkey, and Venezuela. But discredited governments, .n countries with a history of policy failure, are probably better off sending no signals of policy reform and approaching it in small, cautious steps. ThePolicy.Resarch,and Exiemal Affaus CompIe, disrnbutesPRIE Working Papers todisccrr.inatethc 'indings of sok in pro)gress and to encourage the exchange of ideas among Banik staff arid all othcrs interested in development Lssucs These papers caris thc names o! the authors, reflect only then views, and should be used and cred accordinglh The findings, interprctations. and conclusions are thc authors' own They should not be attnbuted to the World Banik, its lBnard of Drtors, its managcmnen:, or any of iLs member countnies Poicy, Research, and External Affairs I Macroeoonomic Adjustment and Growth WPS 788 This paper a product of the Macroeconomic Adjustment and Growth Division, Country Economics Department-is parl of a largereffort in PRE to understand the transition in reforming economies. It was funded by a research project on "Managing the Transition in Adjustment Programs" (RPO 675-70). Copies are available free from the World Bank, 1818 H Sti3et NW, Washington DC 20433. Please contact Susheela Jonnakuty, room NI 1-039, extension 39074 (46 pages). October 1991. Reform is often flawed by delayed implementa- In the literature on second-best reform tion and, after initial acceptance, sudden rever- strategies, misperceptions about the prospective sals in public reaction. In recent years, many costs and benefits of reform add another attribute these delays and reversals to reform intertemporal distortion to the maniy already dynamics because reform, panicularly when identified. comprehensive, rarely takes place all at once. In designing reform, it is important to determine Bold moves - such as preannounced what the best sequence of reforrn would be, institutional changes - are often suggested as a under what conditions that sequence is feasible, way to signal the beginning of a new policy and how expectations will affect the success of regime. Daveri advocates a more cautious altemative reform strategies. approach when a government has been discred- ited by a history of policy failures. T'he PR E Working~ xcra The PRE Woing Paper Soyries eisnn!ae uha f-d nclin^f 'A^'l ul rd^ a n 1,- V-aW!: D..!.;. ecrr.:n xv AffairsComplex. Anobjective ofthc scries is to get these findmgs out quicklv, e-en ifpresentaions ae less than fullh polisied. The findings, interpretations. and conclusions in these papers do not necessanh represent officia! Bank policy. Produced by the PRE Dissemination Center Table of Contents 1. INTRODUCTION 2. ON THE CONSISTENCI OF REFORMS 2.1 Costs and benefits of reforms 2.2 The debate on reform dynamics 2.2.1 Stabilization and liberalization 2.2.2 The order of liberalization: two examples 3. HOW EXPECTATIONS AFFECT COSTS AND BENEFITS OF REFORMS 3.1 The role of expec.ations durinb the transition 3.2 Lack of credibility: causes, consequences, remedies 4. CONCLUSIONS ON EXPECTATIONAL ISSUES AND REFORM DYNAMICS References Figures (*) This paper is part of CECMG's research on Reform Transition. It was written while the author was a consultant at the Macroeconomic Adjustment and Growth Division of the World Bank. I gratefully acknowledge the helpful comments of Bela Balassa, Fabrizio Coricelli, Miguel Kiguel, Nissan Liviatan, Klaus Schmidt-Hebbel and Deborah Wetzel on a previous version of this paper. 1. INTRODUCTION Reforms proposed for Less Developed Countries throughout the Eighties (and for East-European Countries in the last two years) have been principally aimed at achieving an economic restructuring, supposedly conducive to higher and sustained growth, while maintaining (or helping restore) a framework of macroeconomic and political stability. Such policy measures encompass stabilization policies and changes in institutional arrangements and rules in the direction of expanding the role and scope of the private sector and, in parallel, reducing the direct and centralized control of the public sector over resource allocation by altering the incentive structure in both domestic and foreign markets/sectors and strenghtening the legal and regulatory framework where private initiatives take place. Despite their overall advisability, reforms frequently encounter flaws such as delayed implementatior. and, when initially accepted, sudden reversals. Delayed implementation of measures generating winners ard losers as a result of their application is neither new nor specific to Developing Countries. In the first half of the Twenties as well as in the wake of the Great Depression, France was a highly polarized country where a political stalemate came about consequently to sharp disagreement over the distribution of the burden of contractionary policies'. Moreover, ex-ante uncertainty on which group will be the winner and which the I With reference to stabilization policies, Alesina and Drazen (1989) make two further poinLb. FilzL, VvulL;y co Ltections often coincide with a political consolidation, that is when either political side takes the lead and forces the weaker political group sustain a disproportionate fraction of the total adjustment costs. Second, successful programmes are usually preceded by several failures and sometimes do not markedly differ from the unsucceseful ones: the 1926 French stabilization plan was very similar to the 1924 one and the same holds for the 1923 German programme vs. the former ones. 1 loser in the aftermath of the reform may add to distributional conflicts in determining lags of implementation. Not only outward-oriented reforms favour workers and entrepreneurs already active in exportable sectors, but also attract some of them previously employed in protected sectors, whose firms were producing for the domestic markets before the opening up. Besides that. the new relative price structure will enable completely new exportable sectors to grow and develop. Trade liberalization brought about drastic changes in the volume and compasition of trade in South Korea (mid-1960s and early 1980s), Chile (second half of Seventies) and Turkey (early 1980s). World Bank (1989) also quotes the relatively unknown, but nonetheless relevant, Colombia's experience with exports of cut flowers, which rapidly grew from nothing, in the late Sixties, to more than $200 million ten years later. All in all, it seems unreasonable to expect such dramatic changes whose beneficiaries are unidentified in advance to be both perfectly anticipated and supported by all individuals. The most vigorous opposition to reform proposals sometimes arises from inside the Government bodies: e.g. Taiwanese Government officials distrusted the outward-oriented reforms proposed by local economists2, while a proposal to reduce protection for an automobile assembly firm was successfully resisted by the Government agency holding shares in the firm and by the head of a trade regulatory agency tied to the firm3. Finally, implementation lags may be attributed to the short time horizon of the electorate or the Government, which causes short-term adjustment costs to be valued more than the forthcoming benefits of reforms. Once started, reforms do not always carry through the end. The Philippines 2 Fernandez and Rodrik (1990). 3 World Bank (1990, p.49). 2 (1970), Turkey (1958, 1970), Peru (1979) and Pakistan (1959, 1972) cases have been clasaified as failures for their implementation was initiated, yet subsequently reversed. In some cases, their short duration is partially due to external shocks, which make overall well-designed reforms difficult to accomplish for some unexpected events happen. Fluctuations in international prices of exported goods, downswings in aid inflows, climatical variations and civil wars are major setbacks whose frequent occurrence severely constrained the effectiveness of structural adjustment in many Sub-Saharan Countries. Wheeler (1984) attributes Kenya's good performance in the '70s to favourable environmental conditions, while Gulhati (1990) reports that reform efforts in Zambia, Zaire and Madagascar were seriously hampered by sharp terms of trade deterioration, and in Uganda by serious domestic political instability and social unrest. Another possible explanation for reversals has some bearing on one of the causes for lags of implementation mentioned above: after the reform, uncertainty partially vanishes and new information becomes available as to the identity of those who will fare better. Hence some of the losers, if not compensated, may find themselves uncomfortable with the ultimate outcome of the policy change and withdraw their support to the reforming Government, by joining those who have been actively lobbying against the reform from the beginning4. The mentioned shortcomings have to do with reforms as such, without much consideration for the fact that their success or failure strongly depends on ' The relevance of political economy considerations (strength of exporter groups, leadership commitment) in determining a backlash in trade reforms received some empirical validation in the tests carried out by Nabli (1988). 3 reform dynamics, that is characteristics such as: - intensity (how large is the change in the value of the policy instruments required to enact thp r-fm"'! - speed (how short is the time horizon over which the reform attempt takes place), - comprehensiveness (how many areas/sectore/markets are simultaneously involved, that is how broad the field of reforms is), - sequencing (the order in which reforms are undertaken), whose overall influence is related to the kind of reform(s) at issue and to country-specific initial conditions. Take the commonly required fiscal correction for stabilization purposes, which formed part of nearly any adjustment package after 1982. For some reasons (short- run rigidity of fiscal systems, political sensitivity of current Government outlays), the burden of adjustment frequently insisted on public inv-stment by the central Government, especially in social spending. Though budget cuts and improved access to social services by the poor are, in theory, simultaneously possible, toc large a decrease in health and education, or too short the period o0- -:-r . -I. t ^_ _ rz, 'z ...,_ra . X=; sause dramatica±Ly aaverse consequences in Developing Countries than, say, in East-European Countries, where the values of Human Development Indices (such as life expectancy at birth, …_'__' '.ta'-e, ''teac-Cy .ate) a&re cossly similar to -nose ot western Countries. This is to point out that sharp discrepancies in initial conditions suggest variable reform intensity. The classical argument on the speed of implementation of economic liberalization, made by Little, Scitovsky, and Scott (1970), is that reforms should proceed gradually (one at a time or according to a plan determined in advance), so as to 4 minimize adjustment costs in the presence of significant inflexibilities and hurt the smallest possible number of people. Compensation for losers (e.g. firms and workers of previLusly protected sectors, includirg the public sector) should be provided through incomes policy and adjustment assistance. On the other hand, it has been argued5, too slow a move towards reforming could discourage resource reallocation, leave more room for lobbying against reforms and, ultimately, drive people to doubt on their persistence6. As to comprehensiveness, "packaging" reforms allows to exploit complementarities so as to maxe the follow through more automatic: for tariff reduction to produce its highest attainable gains accompanying measures in exchange rate, regulatory and public sector policies are clearly required. According to Hirschman (1963), undertaking global reforms is also advisable in order to allow the implementation of a reform, desirable by itself yet neglected in the face of more popular and visible ones. So essentially did Celso Furtado when the sugar cane problem was dealt with in the SUDENE plan for the Brazilian North East. Moreover, in a comprehensive programme, losers from specific measures can receive some compensation elsewhere. In these cases, the more comprehensive a package of reforms, the more likely its success. However, it should not be forgotten that comprehensive plans may be excessively demanding in terms of management and administration capacities, thus raising coordination problems. The implementation of actual economic reforms requires the collaboration of many individuals and agencies with distinct personal and organizational interests. For instance, though in the presence of full agreement 5 For instance, Neary (1982). 6 On this point, see more in section 3. 5 on general policy goals, the limited significance :f such general statements interplayed with the shortage of administration skills, in generating internal inconsistency and short-run orientation of structural reforms in many African Countries7. However desirable and particularly when comprehensive, reforms rarely take place all at once. Determining whe'ther an optimal sequencing exists and under what conditions it is feasible is an important element in the design of policy reforms, given that not all sequences end up producing ultimately identical results. Yut recommending a specific sequencing has proved a difficult task, nainly because the optimal path of sequencing depends on the kind of distortions Initially affecting the economy and on how Governments evaluate each of them. When simultaneous distortions affect the adjustment process, eliminating one of them does not guarantee t'e economy to approach the first-best. What is worse, no general presumption can be inferred on the direction of the misperception, that is if private expectations overshoot or undershoot the "true" social costs and benefits; hence, very little guidance is provided to ascertain whether ruiorms are inappropriately sequenced and either too rapidly or too slowly implementedO. In practice, reforming Governments in Developing Countries and international organizations proved more eager to address budgetary and trade distortions, taking the rigidities in labour markets and the apparent biases in the composition of Government expenditure as given. This most likely occurred because I See Gulhati (1990), chs.2,3. 9 The bulk of the economic knowledge on sequencing is effectively summarized in Dixit (1986). 6 labour markets and Government expenditure are considered politically contentious areas9 rather than as a consequence of welfare maximization criteria. In the absence of general guidelines, the economists' perspective on reform implementation has taken quite a narrow view and the discussion, especially stimulated by the Southern Cone policy experiments, dealt with few issues such as the precedence of stabilization over liberalization and the order of liberalization (current versus capital account, domestic versus external financial liberalization). Some statements on the principal steps to undertake have been formulated and are sometimes considered 'common knowledge". They can be summarized as follows. First, stabilization is a precondition to any form of liberalization, because not only macroeconomic instability lessens the efficacy of liberalization , but also trade and financial liberalization have adverse macroeconomic implications. Second, domestic capital market liberalization is a precondition to external capital market liberalization, because otherwise foreign capital would flow out, thus destabilizing an economy with poorly working financial markets. Third, external goods market liberalization should precede external capital market liberalization, because otherwise major foreign capital inflows would cause the real exchange rate to appreciate, at variance with the requirementa for a successful trade liberalization. Nevertheless, a broader consideration of the implications of stabilization and liberalization packages raises some question marks even on the few consolidated points reached so far. Special emphasis in the paper is devoted to analyze the role of people's misperception regarding the reforms. 9 Labour markets are not even mentioned among the possible areas of reform in the otherwise comprehensive survey of the main current issues in Latin America provided by Williamson (1990, ch.3). 7 Misperceptions add to objective Costs by lessening people's willingneas to carry out the necessary behavioural adjustment or making their efforts less effective and, in some cases, harmful to the final rvnl rof +-hrt '-Pf9m1 The question that arises is whether there exists scope for the Government to increase the net gains from reforms by avoiding expectational mistakes. Properly announcing the future course of the reform or signalling policy-makers' determination to go through with it, regardless of contingencies, might significantly decrease the uncertainty faced by individuals and open the way to the new regime. Institutional reforms&0 aimed at enabling the Government to carry out binding precommitments on its future course of conduct, as well as establishing reputation, say by overshooting the target, have been proposed as remedies to credibility problems. let often overlooked are the costs entailed by such "investments" in reputation: in the final part of the paper, a careful assessment of their benefits and costs is recommended as a necessary precondition before credibility-enhancing measures can be successfully implemented. An organizing framework to discuss costs and benefits of policy reforms is set up in section 2.1, while section 2.2 surveys some of the issues concerning stabilization and liberalization (section 2.2.11 and t.he oL:dwL u liberalization (section 2.2.2). Section 3 summarizes how expectational issues affect costs and benefits of reforms (section 3.1), with special reference to lack of credibility - ,…"-.eione o II~LUn ex.ectat.c.LW a..d e,L.uu dy,cuuL& liow in section 4. 10 Such as the divorce between Central Bank and Treasury as well as the voluntary participation in binding international agreements. 8 2. ON THE CONSISTENCY OF REFORMS 2.1 Costs and benefits of reforms Reforms are geared to generate the economy's approptiate response in terms of resource reallocation, which was impossible under the former regime. Such reallocation brings about the creation of new fiLme and industries attracting workers and entrepreneurs from previously protected sectors (including the public and parastatal ones), together with a flat reduction in wasteful directly unproductive, rent--seking activities, which usually develop in parallel with controls and regulations. The bulk of the measurable gains from reforms thus consists of higher efficiency. The extent of benefits from the reforms are also influenced (sometimes significantly hampered) by administrative efficiency in reform design and execution. Yet moving resources from one allocation to another also involve transitional costs, deriving from all sorts of preexisting market imperfections and externalities, which injure the success of an overall well conceived and socially advisable rAform&'. Domestic market disruption and temporary unemployment of people laid off in the previously protected sectors; labour retraining and loss of installed and non-shiftable capital, together with macroeconomic external and domestic imbalances, are examples of such temporary costs. They are the price paid for a change of regime and are often considered unavoidable, at least in the short run. 11 Stern (1989, p.616) provides an accurate list of reasons for both market and Government-induced failures, based on the scrutiny of the assumptions underlying the two fundamental theorems of welfare. It is striking how little consideration causes of distortions and externalities are given in the literature onn seqienminga speed and timina of reforms. 9 As Fischer (1986) effectively pointed out, the costs of reforms are usually immediate and highly visible, while their most relevant gains are delayed, dynamic and hard to measure. In a world ridden with imperfections, it comes as no surprise that reforms encounter some flaws. Some of them can be represented through the organizing framework that follows. Consider an economy living infinitely after carrying out a refo- at time 0, and going through two different phases, the transition and the new regime. A reform is undertaken when its social let benefits are strictly positive (that is whenever the total benefits accruing to the economy over time more than offset total costs). Call "transition" the period during which costs per unit of time are greater than benefits per unit of time, while the "new regime" comes about as a result of the reform, when benefits per unit of time are greater than costs per unit of time. The transition ends when gross benefits and costs per unit of time net out. Now define more precisely how costs and benefits evolve over time. Costs stem from existing distortions, such as monopoly or monopsony power, limited factor mobility, price rigidities'2, which result in unemployment, balance of payments and Government budget deficits at the aggregate level, and, sectorally, in temporary unemployment, foregone training costs, bankruptcies for firms and banks (especially in import-competing industries). Assume that coste per unit of time, C, are smoothly decreasing over time, say at an exponential rate a, and starting from an initial level of co, which seems to be in accordance with their hypothesized transitional nature: 12 All of them can be endogenously determined within the private sector (as a result of staggering of contracts or collusion between firms and trade unions within protected sectors so as to prevent entry of new comers) or superimposed by the Government on the private sector through regulation and controls. 10 (1) q - co-exp{-atj O0 ; Obo>0 ; ObO embodies the presumption that the economy has to undergo some "sacrifice" in order to move to the new regime; this implicitly raises the issues of whether people will accept to sustain this trade-off and how much the Government should commit to future benefits through "promises", which will be both taken up in section 3. Equating (1) and (2) (benefits and costs per unit of time) allows to determine the end of the transition phase t': 'n; co /bo) (4) t2 B+a which is positively related to the initial level c. costs and negatively related to the initial level of benefits and to the rate of change of costs and benefits over time. A question to raise at this stage is why a far-sighted reforming Government 11 should care about the length of the transition: if a reform is socially beneficial, the length of the transition should not be influential on welfare grounds, because rnly the present value of net discounted benefits should matter. However, the social discount rate may be too high and substantially lessen the scope for intertemporal substitution, due to politieal instability or fear of external shocks. If the Government faces strong and competitive political opposition, its time horizon may not be long enough to await until the net benefits of reforms turn out positive. The possible occurrence of external shocks might also suggest the society as a whole to prefer a shorter transition period rather than an uniform distribution of the burden over time. Figure 1 highlights how the end of the transition is determined as well as the trade-off between gradualism and "cold turkey". A shock reform cost-equivalent to a gradual reform is drawn for higher values of the parameters co (c*) and a (a.), which leads to a shorter but tougher transition period (t,t'). An intense and rapidly accomplished reform indeed increases temporary costs at the outset, but may still be advantageous, if there are strong motives for the Government to fear the future or if policy-makers enjoy reasonably high support at the beginning. Moreover, the more generous external assistance to adjustment, the less costly to compensate losers for the country; thus the more likely for benefits from a "cold turkey" strategy to outweigh its costs. Vice versa, the net effect of increasing plan comprehensiveness on the length of the transition appears a priori indeterminate because two conflicting elements may move the benefit curve upwards and downwards: a better exploitation of reform complementarities raises its benefits, while congestion effects may contribute 12 to reduce them. 2.2 The debate an reform dynamics 2.2.1 Stabilization and liberalization The advisability of reforms is not only influenced by their prospective costs and benefits, but also by the costs of maintaining the current policy regime. High and unstable inflation and persistent balance of payments deficits as well as microeconomic distortions are undoubtly costly. However, they often coexist in many developing and socialist countries. Iigh inflation by itself is often said not to be too damaging'3, while the welfare costs of price instabilit-' are to a greater extent emphasized. First, in an inflationary environment, the Lse of money as a unit of account is undercut, for contracts become riskier". Second, the increased uncertainty due to instability reduces the informational content of relative prices, potentially worsening centralized as well as decentralized resource allocation. Productive factors are devoted to exploiting opportunities for arbitrage, while the planning horizon of firms shrinks; a defensive and waiting attitude develops in the 13 As long as it is fully anticipated and institutions are inflation-proof. See Fischer and Modigliani (1978) on this point. " The phenomenon called "dollarization" can be taken as an extreme case of the progressive loss of confidence in local currency as the relevant unit of account. 13 private sector as a whole and capital flight may take place either officially (when capital movements are allowed) or unofficially (through over-invoicing and under-invoicJncv of trmlp +- oˇr c tri;h.t smuggl Iig). In turn, persistently negative external balances cause, in the absence of exchange rate corrections, excessive foreign indebtness and/or drawing down of hard currency reserves, which is often exacerbated by the speculative attacks described in Krugman (1979). Microeconomic distortions due to both "market failures" and "Government failures" involve static and dynamic resource misallocation, which decreases the overall efficiency of the economy. Take trade restrictions as an example"5: protection results in distorting the pattern of production and the use of inputs into production and in distorting the pattern of consumption, because price signal3 do not reflect the actual trading opportunities open to the country. Moreover, these costs are probably outweighed by the additional costs of protection due to administration and compliance tasks and individuals' rent-seeking behaviour'6. Hence, costs from macroeconomic instability as well as microeoconomic distortions seem to be very high; both fiscal, monetary and exchange rate measures, on the Anf ,o"A… .- an',.a. iiujralization, on tne other, would be very often necessary to restore the potential for growth as well as to enhance welfare and equity. Yet, economists'7 addressing this question in the ':st t_n y_a_- yrae that ..e.aicALzaL1vn and stao±i.zation packages should not 5 The argument follows Corden (1987). 16 Administration and compliance costs are often very high in the presence of a licensing system, which is also likely to generate rent-seeking costs as a result of wasteful competition in order to obtain the licenses. 17 Among others, McKinnon (1973, 1982), Edwards (1985), Corbo and de Melo (1987), Dornbusch (1983). 14 be undertaken together. Their main argument refers to the conceivable difficulty of establishing an efficient relative price structure in a perturbed macroeconomic environment. In what follows, additional macroeconomic inconveniences, possibly brought about by trade and financial liberalization, are summarized. Macroeconomic implications of trade liberalization As to external liberalization, wrong sequencing (e.g. an abrupt move from quotas to free trade) as well as excessive speed may contribute to the adverse effects on the external balance of liberalizations intended to be welfare-improving'8, thereby triggering off stabilization problems, especially in economies with no access to international capital markets. Moreover, historical episodes, such as the much discussed ones in the Southern Cone, have shown that even liberalization programmes which appeared to be compatible with balance of payments equilibrium from an ex-ante point of view turned out to be harmful if implemented simultaneously with exchange-rate based stabilization policies'9. The use of the exchange rate as a nominal anchor narrows the feasible path for obtaining the real devaluation necessary when liberalizing trade so as to materially support the desired expenditure switching, 18 Temporary budget and balance of payments deficits, though costly in the short-run, may well stem from optimal intertemporal resource allocation as long as they are sustainable, that is if they respect the intertemporal budget constraints and the so called "no-Ponzi-game" condition. '9 Kiguel and Liviatan (1988, 1990) distinguish between exchange rate-based and other types of stabilization policies and investigate their empirical features. 15 while maintaining the balance of payments in equilibriums'. To put it another way, a liberalization that is compatible with the maintenance of external equilibrium may not be compatible with an exchanye-rate based stabilization, for too many tasks are simultaneously assigned to the exchange rate. Alternative ways of staging trade liberalization are also likely to have effects in terms of tax revenues, whose positive or negative sign depends on the price elasticity of imports and exports and the intensity and length of the liberalization process. While a switch from quotas to equivalent tariffs is often supposed relatively soft on the fiscal side, moving from a tariff-based regime towards free trade would probably involve the need to find alternative sources of revenues21. Among others, seigniorage has often been used, especially in Latin America, in the aftermath of the debt crisis, which implied sharp increases in inflation rates and, whenever the exogenous seigniorage needs were not compatible with any steady state inflation, the emergence of hyperinflationary episodes (Bolivia 1985, Peru 1989-90). The alternative route of promoting fiscal reforms to enlarge the tax base while reducing tax rates has rarely been attempted. Macroeconomic implications of financial liberalization As a first approximation, financial liberalization appears to be harmful to stabilization purposes if undertaken early, no matter whether it is domestic or X When the nominal exchange rate is fixed, a real devaluation can obtain only as a result of a domestic deflation. 21 See Edwards (1989) for an elegant formalization of the problem, using duality theory. 16 external. On the one hand, liberalizing domestic financial markets, before stabilization Lakes place, nuuntua to rwduLisLn Guverument revenues C,wu aeigniorage for any level of the budget deficit (through the reduced real demand for money consequent to the increased availability of positive interest-bearing instruments for saving), thereafter potentially fuelling the rate of inflation2. Moreover, the underlying macroeconomic instability is likely to originate high domestic real interest rates, which would increase the burden of Government debt repayment and the likelihood of default. On the other hand, opening up the capital account when stabilization reforms are still under way would permit capital flight and further worsen the external accounts, quite at the wrong moment, that is when additional resources would be especially needed to finance compensatory measures in a non-inflationary way. Further aualifications The seemingly unfavourable macroeconomic consequences of trade and financial liberalization seem to confirm the case for stabilizing first, in order not to worsen domestic and external accounts. Yet: - Though liberalizing first appears not to be advisable, it ought to be borne in mind that stabilizing a distorted economy is presumably costlier than stabilizing an undistorted, perfectly competitive economy. Members of organized coalitions, such as farmers, landlords, or unionized workers, may collude with the goal of appropriating a higher share of national income during the disinflation, thereby 2 McKinnon and Mathieson (1981), Dornbusch and Reynoso (1989). For an application to Argentine in the late Seventies, see Sjastad (1983). 17 increasing price stickiness and slowing down the speed of disinflation when agents' expectations are pessimistic'. The way in which stabilizations are carried out is not at all neutral on the possibility of successfully carrying out liberalization afterwards. For instance, as tariffs are employed to rapidly and effectively raise Government revenues in countries heavily relying on trade taxes>, it may be the case that today's stabilization (without permanent fiscal correction) undermines the success of trade liberalization in the future. - Stabilization policies based on temporary price, wage and exchange rate freezes appear to significantly reverse the timing between costs and benefits of reforms>. Benefits accrue quickly and at low costs, while costs in terms of real exchange rate overvaluation and balance of payments emerge more slowly. After an initial honeymoon period, when benefits net of costs are positive (rapid disinflation and no output contraction), the wedge between the sharply rising costs of real exchange rate appreciation and freeze-induced rigidities in resource allocation and the thin benefits of temporary disinflation widens, so that, at some point in time, the reform is abandoned: this was the case in Argentina, Peru and Brazil in 1985-86. Two countervailing elements w,re at work in the Israeli successful story: the substantial inflows of US foreign assistance and the switch to a nonaccomodating attitude of the Central Bank. 2 A nice formalization of such policy dilemma can be found in Blanchard and Fischer (1989, p.611-612). I Collier and Joshi (1989) point out that the use of trade policy as a support to the macro management Is commonplace, especially in African countries. Raising tariffs to obtain a rapid budget deficit reduction has also played a major role in recent stabilization plans in Argentina and Peru. 2 Kiguel and Liviatan (1991) evaluate pros and cons of so called heterodox stabilization policies. 18 - When stabilizations appear to be systematically delayed>6, structural reforms would run the risk of waiting forever. In such a case, the claim "stabilization first" seems to conflict with common sense, which would suggest that some liberalization steps could be cautiously initiated. However, the "common sense" solution raises an additional problem: the most effective steps on the way to perfect competition do not necessarily ease the way to stabilization. Dismantling the oligopolistic structure of labour and goods markets would surely be aimed at reducing price and wage stickiness, but would also presumably contribute to deteriorate the social climate, thereby laying the foundations for either a rekindling of inflation or a loss of consensus for the manoeuvre. 2.2.2 The order of liberalization While the precedence of stabilization over liberalization reforms as a whole appears to be reasonably robust (though with the caveats expressed in section 2.2.1), nothing comparable can be set out concerning the order of liberalization between domestic and external goocs and capital markets. Th. view that domestic liberalization (in the form of privatization and decentralization) should come before external liberalization, so as to dismantle local monopolistic positions before opening up trade maintains that foreign 2M Alesina and Drazen (1989) provide a conflict-based view of stabilizations as war of attrition games, where the "stabilizer" group, giving up the fight, bears a disprortionate share of the burden of adju3tment. 19 competition should wait outside until a "proper" domestic environment has been set up. However, this fails to address an important question: the development of competition and markets has dynamic effects in the provision of incentives for learning, discovery and innovation. If Governments are not strong enough to bear the unpopularity brought about by the transitional costs of external liberalization, it is not easy to figure out how they could succeed in breaking off local monopolies. Perhaps, foreign competition could provide a less direct way of reducing the degree of monopoly in the economy" and take advantage of incentive provision. Another rationale for delaying the opening up of the economy, provided by Rodrik (1989b), is the "evd for maintaining a certain and predictable structure of incentives for the private sector as a feasible goal for policy reforms (as opposed to liberalization at all coats). It can be refuted that pointing to predictability of the after-reform regime as a major goal may implicitly lend support to the maintenance of the status quo; moreover, as Edwards (1989, p.24) remarks, maintaining a predictable structure of incentives may be too costly when 'nitial Uen aje vezy hiiyh *n.i interrwlnea. Although drawing answers on broad issues such as privatization, deregulation, and management-property relationship remains difficult, recent advances in the theory . ucdulai.i oryanization would prooably provide some clue to a more thorough understanding of the consequences of rigidities and market failures and of the role played by the ownership structure and the type of goods produced than 27 Improving the overall competitiveness of the system does not necessarily amount to establishing a perfect competition world, with many small price-taking firms. A higher market contestability (i.e. decreasing sunk costs to entry and exit) would be enough to deliver competitive-like results (price equal to marginal cost) regardless of the number of incumbent firms. 20 standard welfare models, based on the perfect competition-full information- complete markets paradigm, do. In the absence of such an imperfect competition approacn&, any commitment tu more Pe[;i beJuW1sLe.1Y VdL!1b 4A1 t.. LiA UIll= dJQve mentioned ie likely to be founded on preconceptions. Leaving these general remarks aside, I come to the two aspects of liberalization sequencing which have received explicit attention in the literature: whether domestic financial markets ought to be liberalized before external capital is permitted to flow in and out and whether trade liberalization should precede external capital market liberalization. Domestic financial liberalization as a precondition for external financial liberalization one of the few propositions enjoying generalized consensus is that the capital account should be opened last, because early capital flows can be harmful in two ways, depending on whether they flow out or in after the opening up. In the first place, capital flight takes place when, as mentioned above, opening the capital account precedes stabilization reforms as well as domestic financial reforms. Moreover, apart from stabilization reasons, domestic financial liberalization should come before external opening up in order to prevent too sharp a disruption of existing local financial institutions. The case for low controlled interest rates (resulting in "repression" of domestic 2 Matsuyama (1990) provides a formal discussion of the advisability of temporary protection as a credible threat to persuade local firms to invest and become competitive along the way to complete liberalization. 2 In the 1970s, Argentina and Uruguay started with capital market liberalization, while Chile did the opposite in the same period. Indonesia simultaneously opened both current and capital accounts. 21 financial markets) has often been advocated from Keynesian standpoints either in order to offset people's preference for liquidity or to permit Governments to finance their budget deficits at low costs. Moreover, old and nec-etvucturalists mainly emphasized the stagflationary effects of tight monetary policy and high interest rates. Leaving auide theoretical justifications, loan rate ceilings have been extensively used in many LDCs in conjuction with import restrictions to encourage industrialization through import-substitution. McKinnon (1973) and Shaw (1973) challenged this orientation, pointing out that interest rate ceilings associated with administrative allocation of credit were likely to have adverse effects on sa7ings and growth. The main drawbacks attributed to financial repression were that it discourages risk-taking, prevents lending institutions from charging risk premia, promotes low-yielding direct investment as a substitute for depositing money in a bank and encourages the adoption of capital-intensive techniques. Those advocating domestic liberalization rely on McKinnon-Shaw-type arguments: they expect interest rates in the official credit market to be pushed up so as to make domestic capital markets more appealing to household savings and to eauate thB price of capital with its true marginal cost. However, concentrating on the level of interest rates as the ultimate objective of domestic financial markets liberalization3 can be partially misleading. UndArtnkInn a complete financial reformu requires, am.ong other things, establishing reliable deposit-taking institutions, reducing loan market segmentation (which produces market outcomes unsatisficing on efficiency as well as on equity grounds), allowing for a more indirect Government intervention in 3 This is more often done by those who oppose it. The World Development Report 1989 contains a well developed survey of the requirements for financial development. 22 credit allocation and, at the same time, strengthening the necessary role of supervision and regulation played by the monetary authority. If too s.Luw a seL wZ sdsujwb .L&. puL it pJ.i-w, uc.s -L1 ia sy .oi&ie about. Apart from the mentioned effects on the real demand for money (which increasee the inflationary Impact of any given4 budget deficit), interest rates may well not go up aa a result of financial liberalization. Financial repression, often Government-induced, can also endogenously arise consequently to the functioning of a unregulated competitive credit market whenever collecting information about the risk characteristics of different borrowers is costly. If asymmetric information between borrowers and lenders cannot be overcome through additional contractual instruments (such as the provision of suitable collateral), the equilibrium interest rate may not increase and liberalization would thus prove insufficient to persuade domestic capital not to fly away3,. Alternatively32, especially whenever macroeconomic instability is still around, freeing interest rates may result in deposit-taking institutions paying intereat rates on deposits lower than the lowest rate before interest deregulation, while cnarg2ng a nigner iending rate Lhiaii bc.ure. For sure, domestic financial liberalization is an important factor in encouraging a balanced development in a mdrket economy, because, in allowing a more efficient coilection ano allocation of savings, it mitigates tne effecrs of snarp fluctuations in current income on current consumption, thereby smoothing the consumption path over time. Yet, it is only a necessary, but not sufficient 31 Dornbusch and Reynoso (1989) argue that, in any case, very little empirical evidence exists on the alleged link between interest rates and growth. 32 An example of this phenomenon has been Ghana in the 1980s. 23 condition for enabling the opening up of the capital account. Current versus capital account liberalization Even once macroeconomic adjustment (fiscal and monetary correction) and domestic financial liberalization have been carried out in advance, a simultaneous opening up of current and capital accounts is considered to be harmful to economic welfare. Based on the second-best proposition that, in a multiple distortion world, the removal of one distortion could have negative rather than positive welfare implications, as long as the others stand still, Edwards and van Wijnbergen (1986) show two main results: - negative welfare effects stem from capital market liberalization under trade restrictions; - gradual trade liberalization ameliorates welfare and is superior to an abrupt one under capital market imperfections. The costs of liberalizing capital first in an economy where the goods market is distorted are due to the fact that, the importable good being capital-intensive, an increase in capital endowments would result (by the Rybczynski theorem) in a higher production of importables, which would worsen welfare33. On the contrary, a gradual trade liberalization strategy enhances welfare in the presence of a capital market distortion, because it modifies the consumption discount factor, thereafter increasing saving and investment with respect to the 33 Edwards and van Wijnbergen consider the case where the additional borrowing is employed to increase investment. If all of the new foreign borrowingi were used for current consumption, there would be no welfare deterioration even in the presence of tariffs. This partially weakens the argument. 24 rationed initial position9. Edwards and Edwards (1987, p.182-183) also emphasize that early opening of the capital account unnecesearily raises adjustment costs when resour.e movements are costly. The argument is as follows: given that the speed of adjustment in financial markets is higher than in goAIs markets, an immediate inflow of foreign capital is first observed, while nothing happens in the goods tnarket. This will result in increased money supply, inflation and real exchange rate appreciation, unless the Central Bank sterilizes the inflow. Alternatively, under a floating exchange rate, nominal and real appreciation ot the exchange rate will take place in order to maintain the external balance. In both cases, resources will move into the nontradable sector. However, after enough time has elapsed, the response of the goods markets to the trade liberalization will show up, triggering off a movement of resources in the opposite direction, that is out of the nontradable sector. If resource mobility is not costless, the double movement of resources causes unnecessary costs, which clearly conflicts with the pursuit of successful trade reforms. 3 Once again, the result of superiority of gradualism hinges upon investment bearing a disproportionate part of the borrowing constraint: were there no distortions in the capital market the static welfare loss due to tariffs could prevail and invalidate the result. " Going a step further, some authors, such as Leamer (1980), Neary (1982), Edwards (1988, pp.177-182) and Mussa (1986), have developed models with sector- specific capital and labour market distortions il. uLdi iU ivtaluaLi &t implications of intersectoral transfer costs, wage rigidity and minimum wage legislation in terms of optimal staging of tariff reductions. In these models, the first-best policy still consists of a complete elimination of the tariff (unless income distribution issues with no lump sum transfers are taken into account). However, the presence of adjustment costs (in the form of frictional unemployment) suggests, as a second-best solution, a more gradual approach to tariff reduction. The argument runs as follows: in a two-period two-sector economy, workers are informed at the beginning that protection will last for some time in the future before being eliminated. This makes being employed in the protected industry rational in the first period, while transferring into the exportable sector may not be rational in the next period, as long as there are 25 In general, it remains unclear how the conclusion in favour of gradualism drain by Edwards and van Wijnbergen can be generalized to encompass the many additional distortions characterizing the real world. Further gualifications - Either destabilizing or anti-trade effects of foreign capital do not take place "naturally". Under fixed or managed exchange rate regimes, the damage brought about by capital outflows and inflows can be sensibly reduced, provided that an efficient domestic market for public bonds is created as a part of the financial reform and the monetary authority sterilizes the external inflow of money through open market operations. This appears to strengthen t-he case for domestic financial reform as a precondition to external financial opening up. Once public bonds can be issued and a market for them develops, since then policy mismanagement partially shares the blame for the anti-trade liberalization effects commonly attributed to early capital inflows. - Do the results obtained by Edwards and van Wijnbergen also apply to the timabt e apeencing of liberalization? In principie yes, because the welfare loss in their model is supposed to turn up whenever wrong relative prices drive resource allocation; this would establish an overall presumption for liberalizing goedef f-"t arnd capital later. YeL the recent reform attempts in Socialist Countries suggest some caution with respect to such unfettered generalizations. Liberalization of goods markets implies that firms are allowed to set prices so as to accommodate wage pressures, while,in the absence of formal credit markets, positive adjusttent costs. The same applies when the model allows for on-the-job training: leaving the protected sector, after gaining some sector-specific experience in the first period, is costly. 26 an active interfirm credit market develops. In turn, either the Central Bank accommodates inflation demand by creating money or refrains from playing the role of the lender cf lact reccrt and does not accommodate price increases, causing a credit crunch which propagates more rapidly the more widespread is the mechanism of interfirm credit before the stabilization36. Therefore, liberalizing the domestic goods market in a financially repressed economy results in either inflationary pressures or indiscriminate bankruptcies. In the latter case, firms are unable to face the switch to hard budgeting promoted by the Central Bank. In such a case stabilization reforms are incompatible with goods market liberalization in the absence of an intermediate domestic financial liberalization. - - It must be pointed out that grouping public and private capital flows together, as it is often done when arguing against an early opening of capital markets, can be somewhat misleading. Unlike export credits and private investment and lending, which are usually heavily correlated to business expectations about the evolution of the economy, foreign aid and concessional lending are meant to obey different goals, one of which is to reduce the mentioned transitional costs ru- adjusL.tienL a&id external shocks. Therefore, auuh flows are unlikely to play the same destabilizing role as private capital37. Making this distinction between official and private capital flows, however, does not weaken the argument for postponing the opening of the (private) capital account to the latest possible stage. 3' See Calvo and Coricelli (1990). 31 However, excessive amounts of unconditioned foreign official inflows cause shortcomings, too: they give rise to currency overvaluation and often lend support to the pursuit of wrong domestic policies. This was probably the case with the generous international assistance provided to Somalia and Ethiopia in the second half of the Eighties. 27 3. HOW EXPZCTATIONS AFFECT COSTS AND BENBFXTS OF POLICY REFORMS Section 3.1 describes how costm anil 1-efite of rcf=--m.z a_c affccted by people's expectations about the the new regime, while section 3.2 summarizes the main causes, consequences and remedies to Government lack of credibility, drawing on the current literature on the topic. 3.1 The role of expectations during the transition Expectational problems are usually considered transitional while "transition" is defined as the period in between the long-run and the short-run39; yet what is meant by "transition" is not always clear. An operational definition has been given above and is as follows: in the presumption that the overall benefits of policy reforms are greater than the costs of change, then the transition to adjustment can be seen as the period for which the current costs of the reform are greater than its current benefits per unit of time. Flawed expectations regarding future costs or benefits result in too little or too much adjustment (or adjustment in the wrong direction) with respect to the cpt-..al d_g.c_ -' .B_curc allocatlor, oun U.lm V& oi the private sector wnicn tne reform is intended to produce. In fact, a change of regime involves more than a mere switch from an initial "bad" state to a final "good" state: noninsurable uncertainty on its costs and M Since expectations are deemed to be fulfilled in the long-run. 3 See, for instance, Edwards (1988, p.175). 28 benefits and on what category, group or class has to bear the burden of adjustment drives people to form either excessively optimistic or pessimistic expectations. On the contrary, engendering the appropriate response from agents may be crucial to the success of reforms, at least in market-oriented economies, where resource allocation is to a large extent decentralized. Governments may wonder whether there is some room for increasing the net benefits from the reforms by reducing misperceptions. On the one hand, the absence of any commitment on the part of the policy-makers would imply to leave the private sector observing costs higher than benefits during the transition, which would weaken the consensus for the manoeuvre. On the other hand, unreasonable promises may well succeed in deceiving the private sector today and capturing an easy consensus, but at the price of paying the price of reduced Government credibility in the future. This leads to consider private sector's errors of perception, which may be either Government-induced or "autonomous". Mussa (1986) gives examples regarding trade liberalization. In order to evaluate benefits from moving out of the protected sector into the exportable sector, firms have to evaluate the Vy ue of installed capital in the two industries: if expectations on the future path of earnings are wrong, so will values assigned to the single units of capital. When entrepreneurs have static expectations (so that they expect the present differential in sectoral rate of returns to stay constant over time), the private evaluation of benefits from the reform will overstate the "true" social value. If, on the contrary, they correctly expect that, as accumulation takes place in the new sector, the wedge between the rates of return will narrow down, but they are too pessimistic in evaluating the pace at which this will occur, private 29 expectations of benefits will undershoot their social value. In both cases, people will improperly adjust, respectively, too much or too little. Among the Government-induced sources of externalities, the presence of taxes on factor incomes, reducing the privately perceived benefits of the adjustment with respect to the social one, causes a slowdown in the adjustment process, at least when individuals only take into account the after-tax present discounted value of their "investment"'. As to private market failures in capital markets, individual discount rates may differ from the social one (in both directions); moreover, the country as a whole may face a rising supply curve of credit, due to country-risk premium, while individuals do believe they face a horizontal curve4". In such a case, a tax on foreign borrowing would enhance welfare, avoiding the over-adjustment consequent to euphoric expectations. As an illustration, consider the consequences of external shocks using the same graph as in section 2. Permanent shocks (in the form of discrete shifts upwards and downwards in the benefit curve) result in symmetrically shortening or prolonging the transition period, depending on whether they are positive or negative. Positive (negative) temporary shocks misperceived as permanent may induce people to overadjust (underadjust) and expect an earlier (later) end of the transition with respect to the actual date (see figure 2). For instance, a positive external terms of trade shock feeds into higher productivity and real wages in the exportable sector, yet, when the day of reckoning comes, individuals observe net actual benefits lower than expected, which may result in a complete ^ It is here implicitly assumed that Government expenditure does not affect the private agents, utility function, which may not be the case. 41 This seems to have been the case in Chile after the opening of the capital account. 30 reversal of their attitude towards the reform42. This is to emphasize that people's present suspicious attitude finds its roots in past unjustified euphoria. ±ne rise ana rail or many popu L1sr poiicLes in i..aiLn America are examples of the permanent costs that excessive optimism at the outset, fed by unlikely promises on future policy results, leaves to be paid in terms of the ineffectiveness of future economic policies. Among the possible causes of misperceptions, much work has been devoted to analyze causes, consequences and remedies to "lack of credibility", the label attached to the expectational difficulties encountered by economic policies when forward looking and optimizing private sector and Government confront in a distorted environment. Some of the related issues are dealt with in section 3.2. By now, figure 3 shows how lack of credibility implies, as a result of "rational pessimism", that the curve of the expected benefits (EB) lies beneath the actual benefit curve, so that agents are led to expect a longer and tougher transition than the actual. This would complicate the consensus problem for the Government from the beginning. 42 This is not the only conceivable case. If significant pecuniary externalities exist which, in a "big push" fashion, raise investment profitability as more people invest, then it could be worthwhile for the Government to be committed to optimistic promises, which, if believed, would become self-fulfilling. However, relying on self-fulfilling expectations for the effectiveness of reforms might be dangerous especially if expectations are unstable. 31 3.2 Lack of credibility: causes, consequences, remedies Since credibility has proved a catch-all word in the literature, a precise specification of causes, effects and remedies to lack of credibility seems to be in order. As argued above, not all policy failures derive from lack of credibility; moreover, even in the case in which credibility issues matter, they may not be as relevant as other more tangible factors. A number of determinants of lack of credibility can be listed. All of them appear to ultimately descend from three sources: objective inconsistency, time or dynamic inconsistency, and uncertainty. First, if reforms are blatantly inconsistent with other policies, the violation of budget constraints may be recognized by a rational, forward-looking and fully informed private sector, which would tend not to believe in policy-makers' announcements. Second, in the dynamic programming terminology, a policy course of action is said to be time inconsistent when it does not respect Bellman's principle of optimality, i.e. it forms part of an optimal plan formulated at an initial date, while being no longer optimal at a later date when it has to be implemented. Because of distortions (e.g. irreversible factors, staggered contracts, unavallability of lump-sum taxation; also, conflicting goals between Government and private sector4), ex-ante and ex-post optimal policies available to the Government may be different, thus tempting it to renege on its commitments. 4 Chari, Kehoe, and Prescott (1989) show that no time inconsistency problems arise in team environments, where agents form a grcup sharing a common objective, so that each private agent's preferences coincides with the ones of the society as a whole. 32 However, the private sector, when considering whether the policy package it is faced with is credible, correctly perceives Government incentives, drawing the conclusion that policy-makers will not stick to their declared policy. In such a case, the public neither distrusts nor shows any other prior negative attitude towards Government behaviour, but only provides a rational evaluation of the policy-makers' promise. Another important source of lack of credibility is uncertainty, either exogenous or endogenous to the characteristics of Government and private sector. In an uncertain environment, a well-intentioned Government may be simply unable to respect its commitments if unforeseen contingencies occur. As Schelling (1982) points out, exogenous uncertainty springs up because the most a Government can commit itself to is an input (the programme) rather than an output (the result of the programme). The suggested advice in order to lessen "bad luck" costs is for the Government to distribute the information it is provided with, so as to enable individuals to link "input" (say, the growth rate of the money supply) and "output" (say, the inflation rate) in the same fashion as the Government does, without incurring in misunderstandings. Coming to "endogenous" uncertainty (regarding agents' characteristics), the private sector may be unsure about the real intentions of the Government, because it does not know whether the Government is "strong" or "weak" in its attitude towards the declared goals. Depending on whether this suspicious attitude is justified or not, the strong Government, determined to carry out the reform, has to undertake different actions. If being strong (S) or weak (W) implies different goals of economic policy, S will find signalling its true attitude convenient in 33 order to make W's mimicking costliert hence an incentive to overadjusting arises44. If, vice versa, being S and W refers to policy-makers, ability to respect their promises` under coincident goals, it would be irrational for S not to compromise46. The most direct implication of lack of credibility is that reforms are expected to be short-lived. As a result, either resources are not reallocated at all and the private sector adopts a "wait-and-see" behaviour, generating a coordination problem47, or people react to the intertemporal distortion induced in relative prices by the transient reform. Ir. some of his papers, Calvo (1987, 1988a)4 investigates how lack of credibility results in people overconsuming imports or overstoring durable goods during the free trade regime, perceived as temporary. People's anticipations cause unambiguous welfare losses: the shorter the period of free trade, the sharper the intertemporal distortion and, consequently, the welfare loss. In addition to this, were policy uncertainty about the survival of the reform associated with investment irreversibility (e.g. in the presence of positive entry and exit costs), the rise in investment necessary for the reform to be '4 As long as the conditions for a separating equilibrium to show up hold. See Vickers (1986), Rodrik (1989a) and Persson & van Wijnbergen (1989). 4s Barro (1986). 46 Cukierman and Liviatan (1991) propose a solution, which is still time inconsistent in a world without precommitment: a half-way announcement does not eliminate the incentive for the Government to deceive the private sector as long as it has the last "move". 47 Dornbusch (1988), Simonsen (1988), Rodrik (1989d), provide rationales for these coordination failures. 4 The papers contain further generalizations of the main argument reported in the text, when allowing for nontradable goods, money and ruling out intertemporal substit-ution in consumption. 34 sustainable' may not take place. In turn, an insufficient level of investment ends up increasing the probability of reversals: hence policy uncertainty generates the posaibility of self-fulfilling expectations and multiple equilibria and the economy may be stuck in the "wrong" equilibrium, though nothing is wrong with the fundamentals. Remedies differ depending on the source of lack of credibility; in any case, establishing (or restoring) credibility is indeed costly. As figure 4 shows, the curve of expected benefit shifts up as a result of successful building up of reputation from time t' onwards, but the higher resource costs incurred to provide the signal would result in a similar upward shift of the cost curve. The net outcome in terms of the end of the transition period would thus be indeterminate: B', could in fact lie on the right as well as on the left of B. If public's distrust derives from objective inconsistency, the remedy consists of taking budget constraints into account-. if time inconsistency is the cause of lack of credibility of economic policies, institutional reforms directed at making precommitments enforceable ("tying the Government hands in advance") are usually advocated. Yet, this solution still leaves the question of credibility of institutional reforms by themselves open to discussion. We are thus left with two more opportunities: eliminating or 49 Rodrik (1989d) discusses the role of private sector's expectations in determining the reversal of policy reforms. Among the numerous arguments advanced to explain the consensus enjoyed by protection (Governments unwilling to renounce to easy revenue collection, uneven distribution of gains from trade, short-term electoral concern), Fernandez and Rodrik (1990) emphasize the role of the uncertainty surrounding individuals' status after the reform, which determines a bias in favour of the status quo. In this way, ex-post advisable reforms are not agreed on ex-ante, nor carried out. s Once again simple in theory, though often overlooked in practice. 35 lessening the distortions which cause time ilconsistency to arise or, given the existing distortions, establishing reputation (at least, as long as the time span of the reform is not too short and the "game" is played without much change in environmental conditions). When uncertainty is at the origin of imp,rfect credibility. anntiincements and aignalling may be carried out to bridge the "credibility gap". However, expectations are very unlikely to be influenced by mere announcements or isolated attempts to reverse an ongoing tendency (at least, as long as people are not only forward-looking, but also able to see through Government real incentives). Too simple announcements would produce inconsistent predictions, while detailed or contingent announcements would require all the alternatives be specified in advance, which has two shortcomings. First, an announcement respecting requirements in terms of objective consistency (coherence of means and ends), timing and speed, if strictly applied, would severely restrict the freedom of choice of the Government in cases in which it enjoys informational advantages not easy to spread around; second, contingent announcements would complicate the interpretation of the reform to the public. In fact, the question that arises is whether Governments should rather build up favourable expectations along the way, without bothering very much about the state of expectations at the outset. In addition to being tied up in its announcements, the Government is also subject to constraints on the way in which it signals its true attitude as well, especially when signals are formulated in the form of over-adjustment (which is usually the case advocated by the related literature). 36 Signals, likewise announcements, are costly", because they use up resources, which could be destined elsewhere. If a Government wants to demonstrate its serious intention to Libera.Lze by suiDeLctzing imports (rather than simply reducing or eliminating tariffs), this will probably bring about more severe strain on the balance of payments side as well as on the Government budget constraint. Other items will have to give way so as to maintain external and domestic balance; otherwise, reduced expectational mistakes may be fully offset by increased transitional costs of other nature. 5 In principle, this is taken into account in the literature (for example, Rodrik (1989a)), yet, when policy implications are drawn, this requirement is often left implicit. 37 4. CONCLUSIONS ON EXPECTATIONAL ISSUES AND REFORM DYNAMICS Hisperceptions of reform costs and benefits add another uistortion (of an intertemporal nature) to the many already identified and investigated in the literature, which prevent the economy from reaching a first best solution. Some results from the literature on credibility, partially summarized in section 3.2, point to the necessity of addressing credibility problems directly through preannounced institutional changes acting as signals that a new policy orientation is being followed. In fact, bold moves in trade reforms such as Turkey's elimination of quantitative restrictions to imports, Venezuela's abolition of the agency in charge of foreign exchange allocation, Chile's elimination of quotas associated with a drastic reduction in the highest tariff rates are reported as examples of credibility- enhancing measures in World Bank (1990, p.50). However, even taking for granted that the mentioned cases are success stories, some doubts can be cast on such seemingly widely held view. This is not to deny the relevance of the concept of credibility in the analysis of policy reforms, but rather to emphasize that credibility considerations do not necessarily justify reforming Governments to take on too risky attitudes towards reform implementation, especially in the presence of only partial and not yet consolidated policy prescriptions. On the stabilization side, the quoted example of heterodox shock policies reminds that "signals" not respecting the budget constraint or too long restraining the functioning of a market economy, such as price freezes, rapidly undermine, rather than reinforce, the credibility of a reform, witnessing the difficulty of directly targeting credibility by itself. They may succeed in making 38 stabilization not as painful as traditionally believed, but end up postsponing the solution of problems to the future in the conventional populist fashion, while narrowing the path of future liberalization reforms. Honeymoon periods only bring about temporary benefits, unless they are backed by adjustment in the "fundamentals". At the same time, countries like Bolivia and Mexico, which obtained a surprisingly rapid disinflatiQr. in the recent past, while failing an equally rapid return to sustained growth, can be taken to indicate that adjusting the "fundamentals" may not be enough. If investors retain their option to "wait and see" (Dornbusch, Rodrik), a coordination problem arises and the desired supply response lags behind. In such a case, expectations might be in theory positively influenced by appropriate Government intervention or, better, conditional loans at concessional terms from international organizations, which would help lessen uncertainty about the future. As to liberalization and welfare-improving resource reallocations, the theoretical case for caution seems to be overall strengthened. Froot (1988) points out that, whenever a country has limited access to foreign credit, and is thus forced to deplete its foreign currency reserves to finance external deficits, deviations from the first-best policy, such as a slower pace of liberalization, would lessen the distortion caused by lack of credibility. Moreover, the Edwards-van Wijnbergen proposition on liberalizing trade first and capital later is also strengthened by the arguments brought forward by Calvo (1988b): given that imperfect credibility results in amplifying the existing distortions, simultaneously opening up markets with different speeds of reaction (such as goods and capital) would trigger off a more rapid deterioration in the external accounts. 39 rn general, it remains unclear whether a strong precommitment to a specific path of reforms is ever effective on credibility grounds in those cases when currently undertaken reforms come after a number of past policy failures. As Calvo and Rodrik have shown, when reforms are perceived as temporary, their anticipation may be very costly. This points to the inappropriateness of long-standing precommitments, while suggesting that discredited Governments would better not make any announcements, nor send "signals" at all. A much more advisable strategy would thue consist of starting from the easiest and safest reforms and narrowing the "field" of the game at the beginning, while delaying more ambitious moves to a later stage. 40 REFERENCES ALESINA, A. and A. DRAZEN (1989), Why are stabilizations delayed?, NBER Working Paper 3053, August ANDERSEN, T.M. (1989,, Credibility of policy announcements, European Economic Review, 33, 13-30 BARRO, R.J. (1986), ReF"tation in a model of monetary policy with incomplete information, Journal of Monetary Economics, 17, January, 3-20 BLANCHARD, O.J. and S.FISCHER (1989), Lectures on Macroeoconomics, Boston, MIT Press, ch.11 CALVO, G. (1987), On the costs of temporary policy, Journal of Development Economics, 27, 245-262 CALVO, G. (1988a), Costly trade liberalizations: durable goods and capital mobility, IMF Staff Papers, 35, October, 461-473 CALVO, G. (1988b), Notes on credibility and economic policies, mimeo, April CALVO, G. (1989), Incredible reforms, in CALVO G., R. FINDLAY, P. KOURI, and J. BRAGA de MACEDO, Debt, stabilization and development, Oxford, Blackwell CALVO, G. and F. CORICELLI (1990), Stagflationary effects of stabilization programs in Reforming Socialist Countries: supply-side vs. demand side factors, mimeo CHARI, V.V., P.J. KEHOE and E.J. PRESCOTT (1989), Time consistency and policy, in R.J. BARRO (ed.), Modern business cycle theory, Oxford, Basil Blackwell CHOKSI, A. and D.PAPAGEORGIOU (1986), Economic liberalization in Developing *urU flvfnf-rd, Blackwell COLLIER, P. and V. JOSHI (1989), Exchange rates in Developing Countries, Oxford Review of Economic Policy, 5, 3, September, 94-113 CORBO, V. and J. de MELO (1987), Lessons from the Southern Cone policy _xp_r:-.ente, World Bank Research OhbtrvAr.; JTilv. 111-143 CORDEN, M. (1987), Protection and liberalization: a review of analytical issues, IMF, Occasional Paper 54, August CORDEN, M. (1990), Exchange rate policy in Developing Countries, The World Bank, PRE-WPS Working Paper 412, April CUKIERMAN, A. and N.LIVIATAN (1991), Optimal accomodation by strong policy-makers under incomplete information, Journal of Monetary Economics, forthcoming DIXIT, A.K. (1986), Tax policy in open economies, in AUERBACH, A. and M.FELDSTEIN (eds.), Handbook of Public Economics, Amsterdam, Elsevier Publishing DORNBUSCH, R. (1983), Real interest rates, home goods and optimal external borrowing, Journal of Political Economy, 91, February, 141-153 DORNBUSCH, R. (1988), Notes on credibility and stabilization, NBER Working Paper 2790, November DORNBUSCH, R. and A. REYNOSO (1989), Financial factors in economic development, The American Economic Review, 79, 2, 204-209 EDWARDS, S. (1985), Stabilization with liberalization: an evaluation of ten years of Chile's experience with free market policies, Economic Development and Cultural Change, 33, January, 223-254 EDWARDS, S. (1988), Terms of trade, tariffs and labor market adjustment in Developing Countries, World Bank Economic Review, 2, 2, 165-185 EDWARDS, S. (19e9), On the sequencing of structural reforms, NBER Workin Paper 3138, October EDWARDS, S. and A. COX-EDWARDS (1987), Monetarism and liberalization: the Chilean experiment, Cambridge MA, Ballinger Publishing Company EDWARDS, S. and VAN WIJNBERGEN (1986), The welfare effects of trade and capital market liberalization, International Economic Review, 27, 1, February, 141-148 FERNANDEZ, R. and D. RODRIK (1990), Why is trade reform so unpopular? On the status quo bias of policy reforms, NBER Working Paper 3269, February FISCHER, S. (1986), Issues in medium-term macroeconomic adjustment, World Bank Research Observer, 1, 2, 163-182 FISCHER, S. AND A. GELB (1990), Issues in socialist economy reform, mimeo, November r`SCMER, S. and F. MODIGLIANI (1978), Toward an understanding of the real effects and costs of inflation, Weltwirtschaftliches Archiv, 810-832 FROOT, R.A. (1988), Credibility, real interest rates and the optimal speed of trade liberalization, Journal of International Economics, 25, 71-93 HIRSCijAg, a.O. (1963), Journeys towards progress: studies in economic policy- making in Latin America, New York, Twentieth Century Fund KIGUEL, M. and N. LIVIATAN (1988), Inflationary rigidities and orthodox stabilization policies: lessons from Latin America, World Bank Economic Review, September KIGUEL, M. and N. LIVIATAN (1990), The business-cycle associated with exchange- rate-based stabilization policies, The World Bank, PRE-WPS Working Paper t513, October KIGUEL, H. and N. LIVIATAN (1991), Lessons from the heterodox stabilization programs, The World Bank, PRE-WPS Working Paper #671, May KIGUEL, M. and K.SCHMIDT-HEBBEL (1990), Managing reform transition: issues and research questions, mimeo, November K'IUEGER, A. (1986), Problems of liberalization, in CHOKSI, A. and D.PAPAGEORGIOU (eds.), Economic liberalization in Developing Countries, Oxford, Blackwell KRUGMAN, P. (1979), A model of balance-of-payments crisis, Journal of Money, Credit and Banking, 11, 3, August, 311-325 LEAMER, E.E. (1980), Welfare computations and the optimal staging of tariff reductions in models with adjustment costs, Journal of International Economics, 10, 21-36 LITTLE, I., T. SCITOVSKY, and M.SCOTT (1970), Industry and trade in some Developing Countries, London, Oxford University Press MATSUYAMA, K. (1990), Perfect equilibria in a trade liberalization game, The American Economic Review, 80, 3, June, 480-492 Mc KINNON, R.I (1973), Money and capital in economic development, Washington, Brookings Institution Mc KINNON, R.I (1982), The order of economic liberalization: lessons from Chile and Argentina, in BRUNNER, K. and A.H. MELTZER (eds.), Economic policy in a world of change, Amsterdam, North Holland Mc KINNON, R.I and D.J. MATHIESON (1981), How to manage a repressed economy, Princeton Essays in International Finance 145, Princeton University, December MICHAELY, M. (1986), The timing and sequencing of a trade liberalization policy, in CHORSI, A. and D.PAPAGEORGIOU (eds.), Economic liberalization in Developing Countries, Oxford, Blackwell MUSSA, M. (1982), Government policy and the adjustment process, in BHAGWATI, J.N. (ed.), Import competition and response, Chicago, The University of Chicago Press MUSSA, M. (1986), The adjustment process and the timing of trade liberalization, in CHOKSI, A. and D.PAPAGEORGIOU (eds.), Economic liberalization in Developing Countries, Oxford, Blackwell NABLI, M.K. (1988), The political economy of trade liberalization in Developing Countries, mimeo, November NEARY, J.P. (1982), Intersectoral capital mobility, wage stickiness and the case for adjustment assistance, in BHAGWATI, J.N. (ed.), Import competition and response, Chicago, The University of Chicago Press PERSSON, T. and S. van WIJNBERGEN (1989), Signalling, wage controls and monetary disinflation policy, NBER Working Paper 2939, April RODRIGUEZ, C.A. (1991), Inflation in Peru, mimeo, May RODRIK, D. (1989a), Promises, promises: credible policy reform via signalling, The Economic Journal, 99, September, 756-772 RODRIK, D. (1989b), Liberalization, sustainability and the design of structural adjustment programs, J.F.Kennedy School of Government (Harvard University), Discussion Paper 177D, March RODRIK, D. (1989c), Credibility of trade reform: a policy-maker's guide, The World Economy, 12, 1, 1-16 RODRIK, D. (1989d), Policy uncertainty and private investment in Developing Countries, NBER Working Paper 2999, June ROEMER, M. (1986), Simple analytics of market segmantation: what case for liberalization?, World Development, 14, 3, 429-439 SCHELLING, T.C. (1982), Establishing credibility: strategic considerations, The American Economic Review, 72, 2, May, 77-80 SHAW, E.S. (1973), Financial deepening in economic development, New York, Oxford University Press SIMONSEN, M.H. (1988), Price stabilization and incomes policy: theories and the Brazilian case-study, in BRUNO, M., G. Di TELLA, R. DORNBUSCH and S. FISCHER (edo.), Inflation stabilization, Boston, MIT Press SJASTAD, L. (1983), Failure of economic liberalism in the Cone of Latin America, The World Economy, 6, March, 5-27 STERN, N.H. (1989), The economics of development: a survey, The Economic Journal, 99, September, 597-685 VICKERS, J. (1986), Signalling in a model of monetary policy with incomplete information, Oxford Economic Papers, 38, 443-455 WHEELER, D. (1984), Sources of stagnation in Sub-Saharan Aftica. World Development, 12, 1 WILLIAMSON, J. (1990), The progress of policy reform in Latin America, Institute of International Economics, January, ch. 3,4,5 WORLD BANK (1989a), World Development Report, New York. 0,ford t1ni-fvarity Press WORLD BANK (1989b), Strengthening trade policy reform, mimeo, Trade Policy Division, November WORLD BANK (1990), Report on adjustment lending II: policies for the recovery of growth, Macroeconomic Adjustment and Growth Division, March Costs, benefits e~~ L time eA t*f Figure 1 - Gradualism and cold turkey C'S~~~~~~~~~~C t U Figure . - Effects of misperception of temporary positive shock C,B~~- , C~~~~~~~~~t - - - - ~~~~t,r Figure - Effects of pessimistic expectations (lack of credibility) C,S~~~~~~~~~~S t ime Figure 4 - Effects of investment in reputation PRE Working Paper Series Contact JJ1Q A bhog for zasr WPS771 Macroeconomic Structure and Ibrahim A. Elbadawi September 1991 S. Jonnakuty Policy in Zimbabwe: Analysis and Klaus Schmidt-Hebbel 39074 Empirical Model (1965-88) WPS772 Macroeconomic Adjustment to Oil Ibrahim A. Elbadawi September 1991 S. Jonnakuty Shocks and Fiscal Reform: Klaus Schmidt-Hebbel 39074 Simulations for Zimbabwe, 1988-95 WPS773 Are Ghana's Roads Paying Their Reuben Gronau September 1991 J. Francis Way? Assessing Road Use Cost and 35205 User Charges in Ghana WPS774 Agricultural Pricing Systems and Mark Gersovitz October 1991 B. Gregory Transportation Policy in Africa 33744 WPS775 The Macroeconomics of Public William Easterly October 1991 R. Martin Sector Deficits: A Synthesis Klaus Schmidt-Hebbel 39065 WPS776 Enforcement of Canadian "Unfair" Mark A. Dutz October 1991 N. Artis Trade Laws: The Case for Competition 37947 Policies as an Antidote tor Protection WPS777 Do the Benefits of Fixed Exchange Shantayanan Devarajan October 1991 A. Bhalla Rates Outweigh Their Costs? The Dani Rodrik 37699 Franc Zone in Africa WPS778 A Dynamic Bargaining Model of Eduardo Fernandez-Arias October 1991 S. King-Watson Sovereign Debt 31047 WPS779 Special Programme of Research, Janet Nassim October 1991 0. Nadora Dlqvelonment and Research Training 31019 in Human Reproduction WPS780 Optimal User Charges and Cost Ian G. Heggie October 1991 P. Cook Recovery for Roads in Developing Vincy Fon 33462 Countries WPS781 The Korean Consumer Electronics Taeho Bark October 1991 N. Artis Industry: Reaction to Antidumping 37947 Actions WPS782 The Economic Effects of Widespread Patrick Conway October 1991 N. Artis Application of Antidumping Duties Sumana Dhar 37947 to Import Pricing WPS783 The Origins and Evolution of J. Michael Finger October 1991 N. Artis Antidumping Regulation 37947 WPS784 Chemicals from Poland: A Tempest Andrzej Olechowski October 1991 N. Artis in a Teacup 37947 PRE WorkinQ Paper Series Contact BIe AhQ for paper WPS785 How niid th0 Aqi;n (nr. t;n.. AY- ;- !: ';- 2Obu I D Y I b. King-Vatson the Debt Crisis? 31047 WPS786 Fiscal Policy for Managing Sadiq Ahmed October 1991 B. Prasertwaree Indonesia's Environment 82477 WPS787 Private Investment Under Macroeco- Klaus Schmidt-Hebbel October 1991 S. Jonnakuty nomic Adjustment in Morocco Tobias Muller 39074 WPS788 How Expectations Affect Reform Francesco Daveri October 1991 S. Jonnakuty Dynamics in Developing Countries 39074