A striking feature of the world economy in the 1970s and 1980s has been the frequency and magnitude of the shocks to many economies as a result of changes in the price and availability of natural resources. Resource-poor countries have suffered of course, but paradoxically, resource-rich countries have not been immune. Resource-based booms have been blamed for a tendency towards 'deindustrialisation' in many countries, while the macroeconomic performance of countries with large resource sectors has been less than satisfactory.

This paradox of affluence, which has become known as the 'Dutch Disease', has been the focus of a CEPR project on 'Natural Resources and the Macroeconomy' under the direction of Program Director J Peter Neary and Research Fellow Sweder van Wijnbergen (CEPR Bulletin No. Four). The project culminated in a conference held in London on 10-11 June 1985, funded by the Ford Foundation. It was attended by over 30 economists from a wide range of countries, both theorists who have analyzed models of the Dutch Disease and empirical researchers who have studied the effects of natural resource-based booms in particular countries.

Peter Neary (University College Dublin and CEPR) and Sweder van Wijnbergen (World Bank and CEPR) in their opening paper, 'Natural Resources and the Macroeconomy: A Theoretical Framework', set the scene for the applied studies presented at the conference, with a synthesis of recent theoretical writings on the Dutch Disease. They examined the effects of natural resource discoveries using a series of models, each of which was designed to focus on a specific issue. In the simplest static model, a resource boom affects the economy in two ways. The first is the 'spending effect': higher domestic incomes as a result of the boom lead to extra expenditure on both traded and non-traded ('home') goods. The price of traded goods is determined by international market conditions and does not rise as a result of the extra domestic spending; the price of non-traded goods is set in the domestic market and does rise. The higher relative price of non-traded goods makes domestic production of traded goods less attractive, and output declines. If in addition the booming sector shares domestic factors of production with other sectors, its expansion tends to bid up the prices of these factors. The resulting 'resource movement effect' reinforces the tendencies towards appreciation of the real exchange rate and a squeeze on the tradable goods sector. These predictions are common to most models of the Dutch Disease. However, care should be taken in applying these models to particular countries. For example, the tradable goods sector is often identified with the industrial or manufacturing sector. Nevertheless, protection through import quotas may make prices in some industries whose outputs would normally be thought of as tradable depend on domestic rather than international demand: these industries should then be considered as effectively non-traded, so there is no presumption that their output should decline as a result of the boom.
A booming resource sector has often been identified as one of the causes of unsatisfactory macroeconomic performance. Neary and van Wijnbergen identified two cases in which a resource discovery could lead to unemployment. Unemployment may emerge if the rise in prices of non-traded goods requires a fall in workers' real wages but because of pre-existing contracts or trade union power real wages are downwardly sticky. The second case involved a purely nominal rigidity. The resource boom raises incomes and therefore the demand for money in real terms. In order to restore equilibrium in the money market and avoid unemployment, the real money supply must increase, which will not happen if the price level cannot fall and the nominal money supply is held constant. A moderately expansionary monetary policy which provides enough liquidity to accommodate the boom can avoid this type of unemployment.

Neary and van Wijnbergen then turned to a model with two periods, in order to examine questions about the desirable rate of exploitation of the natural resource. Policy-makers are often concerned that the industrial sector, if allowed to decline during a resource boom, will be unable to benefit from technological progress arising from 'learning by doing'. One can represent this learning effect by assuming that second-period productivity in manufacturing depends on first-period output. There is then a case for a subsidy to manufacturing whether or not new resources are discovered. However, a resource boom may increase the optimal level of subsidy in certain cases. This is especially likely if there are capital market constraints which prevent the economy from smoothing consumption over time by acquiring foreign assets in the period when the natural resources are being exploited. A subsidised manufacturing sector thus acts as a surrogate for foreign assets and helps avoid the need for a sharp drop in consumption when the natural resource runs out.

Commenting on the theoretical issues raised in the Neary and van Winjbergen paper, Ronald Jones (Rochester) noted that the squeeze on other tradables caused by a booming resource sector also occurred in models where all goods were tradable; the introduction of non-traded goods into the model merely intensified this squeeze. On the positive side, a boom in an industry producing non-traded inputs would have favourable effects on the tradables sector by reducing the costs of these inputs. William Branson (Princeton and CEPR) suggested that the learning-by-doing argument presented by Neary and van Wijnbergen needs to be qualified when allowance is made for the diversity of manufacturing industries. During a temporary appreciation of the real exchange rate, competitors expand their output of certain tradables, so there is more learning by doing abroad. Therefore a different structure within domestic manufacturing may be required when the appreciation is reversed. Partha Dasgupta (St John's College, Cambridge and CEPR) analysed the choice between investing resource revenues abroad at a fixed rate of return, or investing them in the domestic physical capital stock. He emphasized the importance of technological progress in extending the scope for worthwhile investment in the domestic capital stock. Some participants doubted whether this result would hold outside the simple optimal growth model used by Dasgupta.
How do governments react to the increase in their revenues brought about by oil-price increases? This recurring theme was addressed specifically by Alan Gelb (World Bank) in his paper, 'Adjustment to Windfall Gains: A Comparative Analysis of Oil Exporting Countries.' He began by estimating the average size of the windfall gain from increased oil prices to Algeria, Ecuador, Indonesia, Iran, Nigeria, Trinidad and Tobago, and Venezuela. The gain represented about one quarter of non-oil GDP; about four-fifths of this windfall accrued to producer governments, whose reactions therefore primarily determined the ultimate effects of the price increase. The countries studied were very diverse, but their responses had much in common. They were dominated by the speedy use of about half the windfall to fund domestic, and overwhelmingly public, capital formation. This share was reduced by greater public and private consumption in the late 1970s, while the proportion invested abroad remained stable at one-quarter.

What have been the effects of these investment programmes? Have they boosted growth in the non-oil economy? Very little, according to Gelb. Most infrastructure investments were subject to long gestation lags. The oil price shock itself rendered some of the initial capital stock obsolete. Much public investment was in large-scale, complex projects, which were prone to substantial cost overruns and disappointing operating performance. Accelerated public investment, entailing large recurrent expenditures, acquired a momentum which proved hard to restrain after oil revenues fell. Such downward inflexibility left these economies particularly vulnerable to the global recession of the 1980s, especially since their access to international capital markets tended to vary with the level of oil prices. Gelb argued that with hindsight these countries would have obtained greater benefits from their windfalls if they had limited domestic investment by applying market criteria more rigorously and had invested a higher proportion of the increased revenues abroad. This would have increased the volume of funds to be recycled by world capital markets, however, at a time when the markets' ability to cope with this process was in doubt. That may have created pressure to absorb the windfall domestically.

Sebastian Edwards (University of California, Los Angeles) pointed out that the temporary or cyclical nature of some resource booms could justify a policy of restraining real exchange rate appreciation, in order to reduce adjustments needed on both the upswing and downswing of the boom. One way that governments might achieve this was by limiting their own spending on non- traded goods.

The importance of the government's use of windfall gains was also emphasised by Jeroen Kremers (Nuffield College, Oxford) in his paper, 'The Dutch Disease in the Netherlands'. Some of the predictions of the Dutch Disease literature were in fact observed in the Netherlands during its natural gas boom, but the complex structure of the economy and the small share of gas output in GDP suggested caution in attributing these developments to the gas boom alone. Resource movement effects were slight, because the gas reserves were easy to exploit. Spending effects operated mainly through the government, which absorbed more than three quarters of the revenue. The increased government revenues from gas had however merely reinforced an existing trend towards increased public consumption and transfer payments. There was little evidence of increased public or private investment to smooth the time path of consumption; Kremers suggested that the use of gas revenues to finance investment rather than current government spending might have helped to achieve this end.

During the 1970s there was a squeeze on profitability in the tradables sector, due to a strong guilder, upward wage pressure and increasing energy prices. Profit margins absorbed the pressure at first, but significant effects on output and employment followed. Government support, sometimes in the form of subsidies, was provided to industry. Some of these industries now appear to have a healthy future; in this sense the government support was successful. The decline in the tradable goods sector also contributed to a sharp rise in unemployment. Policies to deal with this had been less effective. Unemployment benefits have been raised and public sector employment increased; at the same time increases in social security contributions, minimum wages and labour market regulations have exacerbated the unemployment problem.

According to Jean Frijns (Central Planning Bureau, The Hague) Dutch experience underlined the similarity between the spending effects of the gas boom and those of an increase in the government deficit. Each contributed to a squeeze on the tradable-goods sector, which was reinforced by the supply-side consequences of the oil-price shock and labour market policy. Profitability in the traded-goods sector remained low for an extended period. This was only partly explained by firms' use of profit margins as a buffer against exchange rate appreciation; subsidies to ailing industries had also delayed adjustment and the recovery of profitability.

An economy with large reserves of an exhaustible resource is faced with important choices concerning the rate at which the resource is depleted and the size, composition and financing of its investment effort. This can only be analyzed within an explicitly intertemporal framework, which was provided by Kemal Dervis (World Bank), Ricardo Martin (World Bank) and Sweder van Wijnbergen (World Bank and CEPR) in their paper, 'Policy Analysis of Shadow Pricing, Foreign Borrowing, and Resource Extraction in Egypt'. Their framework was a long-run optimal growth model of the Egyptian economy. The economy was divided into three sectors: oil and gas, other traded goods and non-traded goods. Domestically produced traded goods were imperfect substitutes for imported goods. Both oil and traded goods could be exported, but the latter faced a downward sloping demand curve. The government maximized the (discounted) stream of domestic consumption over time, subject to the economy's production possibilities and finite reserves of the natural resource. There were also limitations on the speed at which extraction capacity and domestic investment could grow. Foreign borrowing was constrained by an interest rate which increased with the amount borrowed.
Van Wijnbergen and his co-authors found that key features of the optimal development path for Egypt were relatively insensitive to changes in assumptions concerning future oil prices and oil discoveries. Much of current income arises from decumulation of exhaustible resources and from transfers such as workers' remittances and foreign aid. These may be only temporary, so it is desirable to achieve rapid structural adjustment via high rates of investment in the traded-goods sectors. The analysis also generates time paths for shadow prices of goods and factors, which have more general implications for project analysis and foreign borrowing policies in other countries. For example, the existence of absorptive capacity constraints which are gradually relaxed over time implies that the accounting rate of interest (ARI) should be high while the constraints bind tightly and should decline over time. A constant ARI is commonly used in project evaluation, and this can lead to very misleading results. The optimal path of foreign borrowing depends on the trend in the relative price of consumption goods to foreign goods, since such borrowing must be repaid in terms of the latter. The optimal plan involves a gradual depreciation of the real exchange rate, which makes foreign borrowing more expensive in terms of consumption foregone to repay the loan.

Alasdair Smith (Sussex and CEPR) noted that constraints on the rate of oil extraction were required to prevent implausible 'boundary' solutions in the model; in practice, uncertainty about future oil prices might be a more important stabilising factor. Smith also argued that the shadow prices implied by the fully optimal plan in the model were not identical with those which could be used to identify small welfare-improving projects. The authors of the papers accepted this argument, but noted that current decisions had to be taken about large projects with long life-spans; the shadow prices produced by their long-run model were more appropriate for this purpose than those based on static models, even if the latter had the advantage of incorporating more detail on short-run distortions in the economy.

The discovery of substantial reserves of oil has greatly increased the total wealth of Norway. At the same time, it has shifted the composition of Norway's national wealth towards oil, a riskier asset, thereby increasing uncertainty about the future path of national income. Methods of dealing with this increase in long-term uncertainty were examined by Iulie Aslaksen and Olav Bjerkholt (Central Bureau of Statistics, Oslo) in their paper, 'Certainty Equivalence Methods in the Macroeconomic Management of Petroleum Resources'. The planner's problem was formally modelled as one of maximizing the present value of consumption over a finite planning horizon, together with the present value of wealth at the end of the period. National income is viewed as the return on a set of assets, one of which is oil; the planner's decision resembles that of a private agent, who must allocate his portfolio among different assets. In each period the planners decide how much to consume and how to distribute total investment over a set of assets with different mixtures of risks and returns. Aslaksen and Bjerkholt assumed that the rate of return on petroleum reserves was equal to the rate of growth of oil prices, net of marginal extraction costs. If national wealth was to be allocated optimally among the different assets, the rate of return on oil reserves, adjusted for risk aversion, should be equal to the common 'certainty-equivalent' rate of return on all other assets. Optimal consumption is then simply a linear function of total wealth, a result which depends on the assumption of an exponential utility function.

Aslaksen and Bjerkholt applied their theoretical framework to the Norwegian case, by considering a simplified planner's preference function. They based their estimate of the degree of risk aversion on a government advisory report and assumed a rate of time preference of one per cent. The interrelated risk/return characteristics of a simplified asset menu were estimated from historical data. While the results were sensitive to the degree of risk aversion and depended on several simplifying assumptions, analysis suggested that there were clear benefits to the substitution of oil resources for assets with a more certain return.

David Newbery (Churchill College, Cambridge and CEPR) commented on the nature of uncertainty concerning oil prices in the model. The movement of the oil price around its trend value was assumed to follow a random walk. As a result, the model in effect interpreted any movement in the price away from its trend as a shift in the trend itself, which would imply a large change in wealth over the longer term. One consequence was a large change in optimal consumption, which was linearly related to total wealth; a 10% fall in the net oil price called for a 6% fall in consumption. Another participant noted that the costs of adjusting consumption were neglected in the model, another reason why consumption changed so much in response to a change in the oil price. The nature of oil price uncertainty was also an issue: a high price now might lead to lower prices in future because of recession in the oil-importing countries, conservation, or the development of new oil fields and alternative energy resources. Newbery also argued that the depletion rates suggested by the model were highly sensitive to small changes in expected future oil prices. The authors had also assumed that foreign assets yielded a risk-free rate of return, and their results were very sensitive to the assumed rate of return on this asset. Finally, Newbery pointed out that the estimated degree of risk aversion was very large, in that it implied little scope for intertemporal substitution of consumption.

Could a simple computable general equilibrium model be used to analyse policy issues related to resource booms and slumps in a developing country? Lance Taylor (MIT), Kadir Yurukoglu and Shahid Chaudhry (World Bank) attempted this in their paper, 'A Macro Model of an Oil Exporter: Nigeria'. The framework was Keynesian; it allowed scope for the relative price effects which the resource-boom literature suggested were important, as well as for aggregate demand, which was thought to play an important role in the Nigerian economy. Data were taken from the national income accounts, budgetary statistics and the banking system, while the parameters were chosen by various methods, including international evidence and expert opinion. The model was simulated over the late 1970s and early 1980s, to ensure that it replicated observed data accurately. It was then used to explore a variety of policy options, including devaluation of the exchange rate. These policy options were compared to a 'base' policy of modest fiscal contraction and increased foreign borrowing.

A policy of devaluation causes the federal government's oil revenues to rise in local currency terms, while the Nigerian constitution stipulates that 45 per cent of federal revenues be passed directly to state and local governments. Since the spending propensity of these bodies is close to unity, the authors argued, the expansionary fiscal consequences tend to offset the contractionary effects of devaluation arising elsewhere in the model.

Hashem Pesaran (Trinity College, Cambridge) criticised the lack of sensitivity analysis in the paper. There was no indication how the effects of policies depended on the particular values of the parameters used to construct the model. Nor was enough information provided to allow assessment of the reliability of the model.

The short-run consequences of a commodity export boom can differ greatly between developing and developed countries, because of differences in the structure of financial markets and the operation of monetary policy. A model with assumptions appropriate to developing countries was outlined by Sebastian Edwards (UCLA and the World Bank) in his paper 'Commodity Export Prices and the Real Exchange Rate: The Money Inflation Link'. He applied it to the Colombian experience of a coffee-price boom. On the real side, the theoretical structure incorporated the spending effect of a price-based commodity export boom, which tended to raise demand for home goods. On the monetary side, the boom leads to a balance of payments surplus and an increase in international reserves. The increase in reserves was not sterilised; the monetary authorities did not attempt to control the supply of money, but rather based monetary policy on the exchange rate. As a result the increase in foreign currency receipts automatically raised the domestic money supply. If this money-creation effect of the export boom is large enough, the real exchange rate could depreciate rather than appreciate; it might also overshoot its long-run equilibrium value.

Coffee accounts for about half of Colombia's measured exports, and receipts accrue mainly to private producers. Edwards argued that his model could be used to interpret the temporary coffee- price boom of 1975-77, and the concern expressed by Colombian policy-makers over its monetary consequences. The Columbian experience was broadly in line with theoretical predictions. The analysis suggested that a key consideration was the extent to which the authorities allowed the domestic money supply to respond to a boom believed to be temporary. It was essential to develop effective capital markets, since the limits of export taxation had already been reached in Colombia.
Patrick Honohan (Department of the Taoiseach, Dublin) argued that since the increase in the money supply depended on receipts from export sales, it was revenue from coffee exports and not the coffee price alone which was the key variable; over the period studied by Edwards sharp changes in coffee prices had been offset by export volume changes. He was also concerned by the exclusively short-run orientation of the model, which left long- run equilibrium values undefined. Deletion of variables with insignificant coefficients from the estimated equations might permit telling a simpler story of how an export boom affects the money supply and inflation. This boom moved the economy to a new equilibrium in which the authorities' reactions determined the nominal exchange rate, and the price of non-traded goods adjusted to ensure the required long-run real appreciation of the exchange rate.

Cross-country comparisons often highlight the processes at work in different countries and help to avoid invalid generalisations. This motivated Peter Forsyth (Australian National University) in his paper, 'Booming Sectors and Structural Change in Australia and Britain: A Comparison'. He ignored macroeconomic disequilibrium in order to concentrate on comparisons of resource allocation effects. A large proportion of North Sea oil revenues represented pure economic rents, and this rental component made spending effects of primary importance in the UK case; in Australia, the movement of resources into the expanding minerals sector was more important, since the rental component of revenues was lower. North Sea oil was seen as likely to have a larger, more immediate impact on the United Kingdom than the longer- lasting mining boom on Australia. The Australian case demonstrated that such a mineral boom could give rise to considerable structural changes within the tradable goods sector, with a much smaller effect on overall sectoral output patterns. A substantial part of the structural adjustment required by the boom will last only for the investment phase; costs of adjustment make it worthwhile to extend this phase, rather than attempt to complete it in a few years. Trade and capital flows also helped to smooth the adjustment.

The effects of the boom on government revenue were important for the United Kingdom, but much less so for Australia. A theoretical analysis suggested that the optimal response to such a temporary revenue windfall included some additional public investment, a lowering of other taxes, and perhaps a rise in public consumption. Macroeconomic factors made it difficult to judge whether the UK government's response to its increased revenues approximated the optimal one, or if not, whether private agents had taken compensating actions. It did seem likely, however, that not enough of the windfall was being invested.

The sensitivity of the sterling exchange rate to oil price developments was also considered. Although the United Kingdom is no more than self-sufficient in oil in the long term, a rise in the price of oil has tended to strengthen sterling against most major currencies. This could be explained by the fact that most of Britain's trading partners are long-term oil importers, or by the use of higher discount rates in the foreign exchange markets than were used in the long-term calculations of the oil balance.

John Flemming (Bank of England) emphasised that observed developments in the United Kingdom were heavily influenced by macroeconomic disequilibrium. He doubted that the revenue boom induced by the 1973/4 and 1978/9 oil price increases justified increases in public consumption: estimated oil reserves had not changed since 1973, and price increases had little effect on wealth since the UK was only just self-sufficient. Actual public expenditure would also have been influenced by a downward shift in government expenditure targets, and a correction for the overshooting of these targets in the early 1980s. Other participants stressed that the time profile of a resource boom depended upon decisions made by government or firms concerning the rate of depletion of the resource. Technical factors and strategic considerations in the world oil market were relevant to the UK government's depletion policy.

The final paper of the conference, 'Indonesia's Other Dutch Disease: Economic Effects of the Petroleum Boom', compared the qualitative predictions of Dutch Disease literature with the experience of Indonesia. Peter Warr (Australian National University) examined in addition the effect of government policy responses in this paper. Since employment in the oil sector was less than 1% of the total work force, spending effects were thought to have dominated resource-movement effects. Government expenditure grew rapidly, with about two-thirds of revenue coming from oil taxes. The predicted real appreciation (a rise in the price of non-tradable relative to tradable goods) did occur. The Indonesian exchange rate was tied to the US dollar so domestic prices of traded goods could not adjust; therefore the real appreciation of the exchange rate caused by the oil price increase was brought about largely by increases in the prices of non-traded goods. By late 1978 the apparent decline in the profitability of the tradable goods sector (excluding oil) was the subject of widespread public discussion in Indonesia. The large devaluation of November 1978 seemed to have been primarily motivated by a desire to avoid an excessive contraction of these industries, a contraction which would need to be reversed when the oil ran out. This devaluation of the nominal exchange rate initially had the desired effect on the real exchange rate, but the effect had been gradually eroded. The effects of these relative price changes on the structure of the economy were difficult to identify, but comparisons with neighbouring countries suggested that the growth of Indonesian manufacturing was slower than might otherwise have been expected.

The overall policy response in Indonesia seems to have been more successful than in some other oil-exporting developing countries. Indonesia spent a considerable amount of its windfall gain on investment and did not engage in excessive borrowing. Nevertheless, greater returns might have been achieved if more had been invested in overseas assets.

Doug Purvis (Queen's University) congratulated the author on a well-researched paper. Great care had been taken in constructing the index of the relative price of tradable and non-tradable goods. Warr's data showed convincingly that the initial price effects of the 1978 devaluation were smaller but its medium-term effecrts more prolonged, than official data suggested. Nevertheless, while the relative price movements were undoubtedly reduced by Warr's recalculations, Purvis noted that such movements as remained were still quite significant.

Peter Neary chaired the final round-table discussion. Max Corden (Australia National University) distinguished three main issues: the cyclical nature of resource booms, the structural adjustment aspects, and resource extraction issues. Concentrating (as had the conference) on the first two areas, Corden emphasised the links with earlier analyses of stabilization policy and structural change. The cyclical nature of booms raised normative questions concerning the appropriate policy response to a temporary boom, given certain asymmetric rigidities. On the analytic side, Corden argued that it was preferable to model government policy as endogenous rather than to specify arbitrary policy rules. Analyses of structural adjustment abstracted from these cyclical problems in order to focus on the long-run effects of a permanent resource boom on other sectors. The Dutch Disease literature differed from earlier analyses of the interaction between growth and trade first by taking a more disaggregated approach and secondly by focussing on the implications for the returns to factors of production employed by the affected industries.

John Kay (Institute of Fiscal Studies) felt that the empirical studies had reinforced the importance of the distinction between the spending and resource movement effects. This was illustrated by the differences which arose when the booming sector was an 'enclave' which did not compete for productive factors with the rest of the economy. He reviewed some of the microeconomic issues which arose in an enclave model, where most of the returns from the boom accrue to the government, raising its share of national income and so redistributing resources from the private sector to the government.

If the government captures rents from the resource through taxation, there is a reduction in the 'shadow price' of government revenue, given that other taxes are distortionary. This means that the optimal level of provision of public goods rises; pressures to increase the actual scale of provision may also rise, since the political price of raising government revenue may fall even more. Subsequent discussion emphasised that there were many other factors influencing the optimal and actual levels of government spending, which also tended to change during a resource boom. For example, if public goods relied on non-traded inputs, the rise in the relative price of non-traded to traded goods would tend to reduce the socially optimal level of government spending. Reductions in distortionary taxes would also be an important part of the optimal policy package, so there could be no general presumption that an increase in the size of the public sector during a resource boom was desirable.

If the costs of adjustment and uncertainty about oil prices are large, one might expect markets or governments to develop means of dealing with them. Kay noted that this has not happened to any great extent. There is no long and active futures market in oil, nor have western governments negotiated long-term supply contracts with each other.

Doug Purvis (Queens University) noted that with a variety of approaches, the empirical papers had each provided useful evidence on the issues and raised new questions for theorists. He singled out the problems of intertemporal optimization and uncertainty as particularly interesting. Greater attention would also have to be paid to the nature of the boom itself; price- induced booms had worldwide effects, while resource discoveries tended to be country or region-specific. In order to successfully isolate the effects of the Dutch Disease itself, empirical studies also needed to take account of the short-run effects of macroeconomic policy.

It had been suggested that de-industrialization might not result from resource booms in LDCs, because the booms would allow greater foreign borrowing to finance investment in domestic industry. Purvis drew attention to Gelb's evidence, which showed that the constraints on foreign borrowing had indeed been relaxed, but that this had encouraged rather inefficient investments. These investments often involved attempts to incorporate the oil or gas enclave into the wider economy by building large 'downstream' projects such as petrochemical complexes or fertiliser plants, even when it was more beneficial to the economy as a whole to export the resource.

David Newbery (Churchill College, Cambridge) pointed out that oil could now be regarded as a commodity with a rather volatile price. Estimates of the gains from the stabilization of other commodity prices had been small, even in models where macroeconomic consequences had been considered. The missing element in these models seemed to be a satisfactory theory of government behaviour. Many participants had stressed the importance of government responses to large increases in commodity prices. This consideration greatly strengthened the case for commodity price stabilisation.

Newbery argued that the most important question concerning depletion policy for gas and oil was the timing of the required investment, which was large, risky and had a long life-span. Once the investment had been made, the marginal costs of extraction were low, and it usually made sense to extract the resource as fast as was technically possible.

The papers presented at the conference and described in this article will be available in Natural Resources and the Macroeconomy to be published by Basil Blackwell in March 1986.