Interactions between macroeconomic policy and the agricultural sector are particularly important to developing economies, whose agricultural sectors account for major shares of economic activity and income. Moreover, the effects of liberalizing agricultural trade on income distribution have potentially significant implications for the political economy of reform. Papers presented at a joint conference organized by CEPR and the OECD Development Centre in Paris on 22/23 April analysed interactions among trade, macroeconomic and sectoral policies, with particular emphases on the sequencing of reforms, their impact on agriculture and their implications for world trade patterns. The conference the first joint meeting between the OECD and CEPR was organized by Ian Goldin, Programme Head at the OECD Development Centre, and L Alan Winters, Co- Director of CEPR's International Trade programme. Financial support for the conference was provided by the OECD Development Centre.

Sectoral Policies

Kym Anderson (GATT) opened the conference with a paper on `International Dimensions of the Political Economy of Agricultural Policy'. According to the neoclassical political economy approach, `poor' countries switch from taxing to subsidizing agriculture as their economies develop, which suggests that protectionism will increase over time. Anderson argued that international factors, such as US and East European pressures on the European Community to open up its agricultural market and the increased budgetary costs of the Common Agricultural Policy entailed by any expansion of its membership, will modify this bleak outlook. Growing understanding of the financial and environmental costs of agricultural protectionism and the potential benefits from liberalization may also turn policy-makers against protectionism. On the other hand, food safety and other environmental concerns may lead to an increase in non-tariff barriers to trade.

David Blandford (OECD) criticized Anderson for over-simplifying the political economy approach. Many poor countries encourage the import-substituting parts of their agricultural sectors, and certain industrialized countries such as Australia and New Zealand discriminate against agriculture relative to industry. Increased pressure on governments to reduce the budgetary costs of protection may not induce measures to increase transparency and efficiency.

In a paper on `Infrastructure, Relative Prices and Agricultural Adjustment', Riccardo Faini (Università di Brescia and CEPR) developed a model in which countries' participation in World Bank/IMF programmes is endogenous in order to assess the overall impact of adjustment lending. His results confirmed earlier findings that these programmes appear to raise agricultural growth and reduce investment in agriculture but exert no significant impact on economic growth or the current account. Faini then used an indirectly estimated revenue function distinguishing agricultural, industrial and services sectors and labour, private and public capital to assess the contributions of price and non-price determinants (such as infrastructure) to the observed growth of agricultural production. Both price factors and the availability of infrastructure influenced the agricultural sector's performance significantly, but industry was more responsive than agriculture to public investment.

Paul Seabright (Churchill College, Cambridge, and CEPR) suggested that taking account of adjustment programmes' duration, their policy content and the extent of their implementation might improve the model's performance. He also noted that public investment appeared to be more effective than private investment in the agricultural sector, while Christian Morrisson (OECD) noted that the OECD Development Centre's case-studies on agriculture in Malaysia and India had found substantial time-lags between infrastructural investment and supply responses.

Santiago Levy (Boston University) then presented his joint paper with Sweder van Wijnbergen on `Agricultural Adjustment and the Mexico-USA Free Trade Agreement'. This assessed the impact of liberalization on Mexico's relatively protected agricultural sector: its employment and wages, income transfers among six household types, and aggregate efficiency and welfare, under various assumptions concerning government intervention and migration. Liberalizing the maize market with full government compensation for welfare losses realized considerable efficiency gains, which more than doubled once allowance was made for migration. Without compensation, owners of irrigated land gain, while other rural households lose, but migration mitigates these effects; urban households also suffer but to a lesser extent. Introducing rural employment and urban transfer programmes greatly reduces and in some cases reverses these effects; moreover the government's revenue gains from liberalizing maize trade more than offset its costs to the public budget. Including other grains and the abolition of US tariffs on fruit and vegetables presented a more varied picture: government revenue generally increased, allowing compensation programmes to be targeted on the net losers.

Peter Lloyd (Melbourne University) questioned whether the small country assumption applied to Mexico and suggested incorporating the US-Canada FTA, since the current negotiations are trilateral. Gerard Adams (University of Pennsylvania) suggested considering the impact of US foreign direct investment in Mexico and extending the analysis of the industrial sector, while several participants criticized the assumption that migration costs are exogenous and fixed.

Exchange Rate Reform

In a paper on `Exchange Reforms, Supply Response, and Inflation in Africa', Ajay Chhibber (World Bank) estimated the responses of export supply and total agricultural supply (controlling for manufactured goods' availability in rural areas) to real devaluations, which were positive in both cases. He developed a model of inflationary responses to nominal devaluations to determine for African economies the contributions of imported and `domestic' inflation and to divide the latter into cost-push and wage/demand-pull effects and those stemming from the deregulation of prices.

Chhibber found that devaluation initially leads to cost-push inflation, while its second-round effects depend on the financing of the budget deficit, the availability of external financing and the response of exports. The concerns of the UN Economic Commission for Africa about structural adjustment programmes are legitimate but do not reverse the conventional wisdom: the extent of exchange rate misalignment, loss of markets to Asian competitors and the development of parallel markets leave African economies no choice but to pursue exchange rate reform programmes.

Jean-Paul Azam (Université de Clermont-Ferrand) argued that for countries with dual or multiple exchange rates, a devaluation of the official rate can affect the parallel rate if exporters switch to the official market or the expectation of future devaluations leads to capital outflows. He attributed Ghana's low inflation during its devaluation to the inflows of refugees into the agricultural sector and the availability of external finance to subsidize imports. In Madagascar, in contrast, with no significant external finance, successive devaluations destroyed the purchasing power of monetary assets with severe consequences for certain social groups.

In their paper, `Do the Benefits of Fixed Exchange Rates Outweigh their Costs? The Franc Zone in Africa', Shantayanan Devarajan (Harvard University) and Dani Rodrik (Harvard University and CEPR) stressed that the choice of fixed versus flexible exchange rate regimes is particularly important for countries that rely on a few primary exports with volatile world prices. They used a model with government-set output and inflation targets to assess the wisdom of the CFA Zone countries' irrevocable commitment to fixed parities. Their preliminary results indicated that fixed exchange rates maintain low inflation while output fluctuates with the terms of trade; while flexible exchange rates pose no check on inflation but reduce output fluctuations. They then conducted a cost-benefit analysis to show that the CFA Zone countries' fixed rate regimes have impaired their adjustment to external shocks which the benefits of lower inflation have not fully offset.

Emil-Maria Claassen (Université de Paris Dauphine) argued that a simple comparison between CFA Zone countries and others took no account of the latter's variety of exchange rate regimes. The decline in many CFA Zone countries' output stemmed not from the exchange rate regime, but rather from the policies they had adopted after favourable but usually short-lived external shocks. The harmful effects of a terms-of-trade deterioration could therefore be offset in part through restrictive monetary policies, even within the CFA Zone.

Implications for Welfare

In a paper on `Sequencing and Welfare: Labor Markets and Agriculture', Sebastian Edwards (University of California at Los Angeles) developed a two-period general equilibrium model of a small developing economy to compare the welfare effects of alternative sequencing of capital and current account liberalization. For a three-sector economy with net agricultural exports and manufacturing imports subject to import tariffs, with minimum wages in the manufacturing and services sectors and rural-urban migration, Edwards showed that current account-first (or tariff-first) sequencing is not usually optimal, while with a minimum wage across the whole economy in the first period only, the capital market should be liberalized first. Further, the market distortions associated with a given sequencing of reforms significantly affect agriculture, employment and output. Migration and intersectoral reallocation of capital and natural resources depend on real exchange rate movements, which constitute the main transmission channel for dynamic effects.

Helmut Reisen (OECD) disputed the assumption that the real exchange rate appreciation associated with financial liberalization increases labour demand in the services sector and thereby reduces agricultural employment. In fact it leads to search unemployment in the urban informal services sector, disguised low-productivity rural unemployment, and a tendency for employment in industry to contract while agricultural employment expands. Further, Edwards's production function considers only the demand for output and labour input and so says nothing about pre- versus post-reform profitability; hence it cannot pick up the shift from low- to high-productivity uses of resources that liberalization and structural reform are intended to achieve.
François Bourguignon (Département et Laboratoire d'Economie Théorique at Appliquée, Paris) and Christian Morrisson (OECD) then presented their joint paper with Akiko Suwa, `Adjustment and the Rural Sector: A Counterfactual Analysis of Morocco'. Morrisson stressed that the sustainability of adjustment programmes in the face of urban opposition is critical to their success. Nevertheless,careful monitoring is required to ensure that they do not permanently damage human capital by either increasing absolute poverty or worsening the health and education conditions of the politically weak rural majority. Following the adjustment loans of 1984-5, Morocco experienced satisfactory GDP growth, a significant reduction in the external deficit, little change to budget deficit and an increase in farmers' mean real income. Bourguignon then reported simulations of foreign exchange and financial liberalization, tariff reform and a reduction in export duties, which they compared with base-run estimates using alternative closure rules: the first picked out short-run price and wage rigidities in the modern sector with full product and labour market competition in agriculture; while the second focused on the long run, with full price and wage flexibility throughout the economy.

With rigidities in the modern sector, the costs of a crisis and the benefits of a successful adjustment both concentrate in the rural sector. Relative prices determine the intersectoral bias of adjustment, so either reduced agricultural tariffs or reductions in the export tax on phosphates will harm Morocco's rural sector. An initial rise in exports leads to a real currency appreciation, reduced exports and increased imports, as the agricultural terms of trade and the share of agriculture in GDP both fall. If prices and wages in the modern sector are flexible, however, the effects of tax changes are spread more evenly, so their impact is less pronounced. Lastly, some quite unexpected indirect effects of macroeconomic policy reforms on particular sectors may dominate their expected overall effects, at least in the short-to-medium run.

Shantayanan Devarajan suggested extending the model to differentiate rural and manufacturing capital markets, and to examine the effects of infrastructure on income distribution within sectors.

Trade Policies

In their paper, `Trade Reform and the Small Country Assumption', David Evans (Institute of Development Studies, University of Sussex), Ian Goldin (OECD) and Dominique van der Mensbrugghe (OECD) explored the implications of relaxing the small country assumption. This is particularly important for producers of tropical beverages, which face highly inelastic world demand and are among the poorest of developing countries. They used the applied global general equilibrium Rural-Urban North-South model to assess possible changes to existing patterns of trade and export taxes on traditional primary commodity exports, paying particular attention to coffee and cocoa. Their preliminary results suggested that a 25% cut in producing countries' export taxes on these tropical beverages would impose substantial costs and GDP losses. This suggests that coordinated policy-conditional assistance should take explicit account of the removal of the small country assumption. Goldin stressed in conclusion that developing countries' legitimate concerns regarding a fallacy of composition in policy design must be met without eroding the wider commitment to freer trade, including reductions in OECD protectionism.

Christopher Bliss (Nuffield College, Oxford, and CEPR) noted that cost-benefit analysis requires accurate predictions of future prices and hence recognition of the `fallacy of composition' argument. He questioned whether export tax agreements would work in practice. If these exporters' terms of trade are in long-term decline, their comparative advantage must lie in other products, and distorted pricing will suppress diversification. Goldin agreed with this interpretation and replied that the simulation results highlighted both the significance of the sequencing of reforms and also the case for a slower rate of reform of selective trade interventions. This would be associated with higher levels of GDP and investment in agriculture, which could be directed towards diversification. In contrast, current adjustment packages are often more associated with reductions in such investment.

In his paper, `Taxes versus Quotas: The Case of Cocoa Exports', written with Maurice Schiff, Arvind Panagariya (University of Maryland) used a simple model of demand and supply to assess the options available to a country facing a `fallacy of composition' for one major primary export. They compared equilibria under Nash-Bertrand behaviour, where countries fix export taxes taking other countries' taxes as fixed and given; Nash-Cournot behaviour, where they fix national quotas taking other countries' quotas as given; and Stackelberg behaviour by one or more countries under both tax and quota games. Applying their model to the world cocoa market to estimate ten countries' optimal taxes and quotas, they found that Nash taxes yield lower profits and Nash quotas higher profits than most countries' actual taxes. With Côte d'Ivoire as the Stackelberg quantity leader, its profits rise but other countries' profits and aggregate profits fall. Quota cartels have a poor record of success, however, and Panagariya recommended that these countries diversify their production. He noted that the World Bank has consistently refused concessionary loans to expand tea, coffee and cocoa production since 1968, whenever an alternative has been available.

Peter Lloyd questioned whether the model's behavioural forms adequately captured the rules of the real world and the characterization of an oligopoly by ten players. Game theory ignores the net loss arising when supplying countries' gains from restricting supplies are outweighed by importing countries' losses.

Prices and Taxes

In a paper on `Domestic Price Stabilisation Schemes: Forward Markets or Marketing Boards?', Christopher Gilbert (Queen Mary and Westfield College, London, and CEPR) developed a model in which farmers set production levels relative to expected prices and which explicitly linked aggregate market behaviour to that of individual farmers. A farmer may choose to sell his annual crop on the cash or the forward market or to a monopsonistic marketing board, while all other farmers remain in one of the three regimes: hence there are nine possible outcomes overall. For simulations over a 200-year period, Gilbert found that the mean price level is highest for the (general) cash market, where production is lower than under forward sales and marketing boards, while standard deviation is lowest for forward sales. The higher production levels under marketing boards and forward markets reflect these regimes' lower risk, but average income levels are highest under cash sales. An individual farmer moving from cash to forward markets under a cash sales regime faces a trade-off between reduced average levels and greater stability of gross revenue and income. The same move under a forward sales regime will increase the individual's average revenue and net income, and this will always increase average income and reduce its variability under marketing boards. Gilbert concluded that forward markets have weaker income-stabilization properties than marketing boards, although the reduction of risk stimulates extra production in both cases, which may cause crop prices to fall. If widely adopted, this may leave farmers worse off as a group.

Walter Labys (University of Western Virginia) observed that Gilbert's results were heavily dependent on the precise specification of the product supply functions. Jean-Marc Boussard (Institut National de la Recherche Agronomique, Ivry) noted that marketing boards in much of Africa are a legacy of the colonial period and are intended on the whole to benefit consumers rather than farmers. Moreover, the differences between marketing boards and the other two regimes should be much larger than Gilbert's results suggest, since in theory the supply to a marketing board should be infinite.

Ron Anderson (Université Catholique de Louvain) presented his joint paper with Andrew Powell on `Markets, Stabilization and Structural Adjustment in Eastern European Agriculture', in which they reported preliminary work that drew on case-studies of Hungary and Poland. He noted that permanent shocks, the move to hard currencies, the liberalization of domestic markets and privatization are causes of uncertainty in Eastern Europe. Achieving the transformation from a planned to a market economy while ensuring stability involves not just stabilization around a given historical mean but choosing a paradigm for the future course of the East European economies' development. Anderson argued that the required adjustment forms a disequilibrium process with a starting-point far off the final equilibrium, which may be subject to continual shocks and will result in numerous entries and exits by individual enterprises. He then assessed the relative effectiveness of trade tools (variable levies and export tariffs), domestic market tools (public buffer stocks) and market-based risk contracts (futures markets) in achieving stability in the agricultural sector. The criteria were their effectiveness as stabilizers, their ability to facilitate adjustment, and their contribution to the creation of new private enterprise institutions. He found that trade tools would create distortions that would be difficult to remove in future, had weak stabilization properties, and required significant administrative capacity; buffer stock schemes on the other hand involved serious credit access problems. He therefore concluded by favouring futures markets.

Gerard Adams agreed with Powell and Anderson's notion of a radical disequilibrium and noted that rigorous analysis and planning were needed to determine where true comparative advantage lies, which institutions need to be established and what the probable long-run market prices will be.

Vito Tanzi (IMF) presented a paper on `Structural Factors and Tax Revenue in Developing Countries: A Decade of Evidence', which assessed the impact of structural changes on the level and structure of developing countries' tax systems. He analysed changes in the level and structure of taxation for developing countries as a group, for different regions, and for groups of countries with given common economic characteristics, using data on tax levels (as percentages of GDP) and tax structures in 88 countries during 1978-88. He then regressed tax/GDP ratios on per capita income and shares of agriculture, imports and public debt in GDP. His results suggested that average tax levels for the whole group changed very little (at 17-18% of GDP), despite the major structural changes over the decade. Africa and Latin America had the highest average tax levels and experienced the greatest increases; average taxation in the 17 highly indebted countries fell after 1982 but regained its former level by 1988. Per capita income became progressively less significant in explaining differences in tax levels over the decade, while the share of agriculture in GDP increased in significance. The shares of agriculture, imports and external public debt in GDP explained nearly half the cross-country variance in tax/GDP ratios; and the importance of Value Added Tax in these countries' tax structures increased, while export duties gradually disappeared.

Nicholas Stern (LSE) questioned the inclusion of countries as diverse as China and India in the same population distribution. He noted that the LDCs had done well to maintain their tax levels during the adjustment period but had performed poorly on the expenditure side. Sebastian Edwards suggested analysing the impact of inflation on tax structure, while Helmut Reisen argued that total public debt, not just its external component, should be considered. Tanzi welcomed these remarks but he noted that collecting data on total debt was very difficult in practice.


The papers presented at this conference will be published in early 1992 in a joint CEPR/OECD volume, to be edited by Ian Goldin and L Alan Winters.