Chapter II. The world economy.

Barrell, R.J. ; Gurney, Andrew ; Dulake, Stephen 等

Uncertainties over the prospects for the oil price mean that the short-term forecast is subject to a wider margin of error than usual. The Iraqi invasion of Kuwait could lead to a reduction of world oil production by at least eight per cent immediately and has already produced a significant rise in oil prices. Our central forecast assumes that in the short turmoil prices will stay firmer, but that in the longer run there will be no major change in oil market conditions. We have assumed that oil prices (or more precisely the arithmetic average of Brent and Dubai spot prices) will be around $25 per barrel in the second half of 1990, but that they will fall thereafter. This should leave crude oil prices in the range 20-22 per barrel by the end of 1991. Annex I to the chapter investigates the effects of an oil price rise on the world economy, and looks in particular at the distribution of the effects across the major economies. The conclusion of the Annex is that a 25 per cent rise in oil prices is likely to raise inflation in the major seven economies by only a quarter to a half a per cent in the short run, and to lower output by up to a half per cent.

We do not expect oil prices to stay high for long because the reduction in supply from Iraq and Kuwait only partly offsets the current situation of oversupply in the market. Unused oil production capacity in the world is probably only just less than iraqi and Kuwaiti production, although some of this capacity is located in the Gulf. Even if output fell by the full amount of the two countries' production, stock levels are very high. The combination of strategic reserves, commerical stocks and stocks in producer countries combine to produce stocks that would cover almost a full year's consumption. After some initial turbulence the existence of these stocks, which relative to output are double the level available in either 1974 or 1979, will put downward pressure on prices. it is also the case that the response to the 1974 and 1979 oil price shocks increased the flexibility of the capital stock, and it is possible to shift between primary energy sources more quickly now than was possible in 1974. Chart 1 plots the past level of the real oil price and also our forecast up to 1999. Although prices fall in the short term we expect them to gradually strengthen over the next decade. There are many differences between the scale and impact of the last two oil crises and the current one. The oil intensity of output in the advanced world has declined by 25 per cent over the last decade, and oil production is only around its 1979 level. Any oil price change will hence have less effect on prices and output. We are also not anticipating that there is any prospect of oil price rises on the scale observed in 1974 and 1979 partly because demand has become more elastic, but also because OPEC is both less important, with less then 45 per cent of output, and also much less cohesive than it was in the 1970s.

It is not only the oil market that is now different. The 1974 oil crisis was preceded by rapidly rising commodity prices and a very sharp acceleration in world demand and inflation. Commodity prices have now been falling sharply for two years, and world output has already begun to slow down. The 1979 oil crisis was preceded by a period of rising inflation accompanied by relatively stagnant output, and although the oil price rise contributed to inflation it was not the major cause of the recession in 1980/1. The change of administration in the US (and the UK) was accompanied by a shift in policies. Real interest rates in the US, which had been negative, rose very rapidly, and the US real exchange rate also rose. Real of the early 1980s had left most of the major economies operating well below capacity. Chart 2 plots capacity utilisation rates in the major economies since the mid-1980s. Utilisation at the turn of the decade was higher than for at least ten years, and as a consequence inflationary pressures began to develop in all the major economies. Chart 3 plots recent and prospective inflationary developments in the major four economies.

The reasons for the surge in growth during the late 1980s are still unclear, but a number of factors contributed. The defence of the Louvre accord exchange rates in 1987 led to large scale increases in foreign exchange reserves especially in Germany and Japan. Although these flows could have been fully sterilised neither country wanted to take on the full interest rate consequences that this would have required and as a result monetary conditions were loosened and monetary targets were overrun in both countries. German central bank money targets were overrun in 1986 and 1987, and even the change to M3 targeting in 1988 could not prevent a further overrun. Although the defence of the Louvre accord exchange rates failed it did allow the US to export some of its inflationary pressures to the rest of the world. The effect of this monetary loosening was compounded by the effects of financial deregulation which has led to a temporary rise in consumer spending in a number of economies. The process of deregulation probably started earlier and has proceeded further in the UK and consumption has risen strongly as a result. Deregulation has also been under way in Japan, albeit at a slower pace. Both France and then Italy have, partly as a consequence of the decision to remove capital controls in 1990, undertaken some degree of deregulation. The process is likely to extend further, especially as the success of the European Commission's 1992 programme depends heavily on increased competition in financial services. This is likely to have a particularly strong influence on the availability of consumer credit in West Germany, where the banking system has until recently been dominated by rather cumbersome regional banks owned by the Lander.

The basically monetary cause of the increase in growth has been followed by a basically monetary response from the authorities. Chart 4 plots recent changes in short-term interest rates over the last year in the major four economies. German and Japanese short rates have risen sharply whilst those in the US have been falling from peaks that were achieved earlier. Chart 5 plots recent and prospective long rates. These rose sharply in late 1989 and early 1990, especially in Germany and Japan. Higher long rates indicate either that the market is anticipating higher short rates over the period to maturity or that the risk premium on long assets has risen. Anticipated short rates may have risen either because inflation expectations have been revised up or because real interest rates are expected to be higher. We feel that the rise in long rates, especially in Germany, reflects all three factors. Unification with the East is likely to put downward pressure on wages and increase profit opportunities, raising rates of return. However, perceived risks on the average future investment may have risen and as a result the liquidity premium is likely to be higher. Finally, and we feel most significantly, inflation expectations do appear to have been revised upward throughout the major

The monetary response to rising inflation colours our whole forecast, which is set out in table 1. US output growth appears to have been slowing over the last year, and recent data revisions have reduced estimates of growth in 1989 to 21/2 per cent, which is below our estimate of capacity growth. The prospects for slower growth and hence declining inflation have allowed the US authorities some space for reducing interest rates, and the looser stance of monetary policy, along with continuing Congressional delays to producing a tighter fiscal stance, indicate that although growth may be as low as 1.5 per cent this year it is likely to rise into 1991. Inflation in the US should reach a peak in 1990, and we are forecasting that it will fall to 41/4 per cent in 1991. Output growth in the US is expected to stay below capacity until 1993, puffing downward pressure on price inflation. We expect that inflation will have fallen to around 2.75 per cent from 1 994 onward.

The prospect for a soft, but successful deceleration in US inflation has consequences for the rest of the world. In particular we expect that world trade growth will slow down from an average of around 7 per cent over the last four years to around 4/4 per cent in 1990 and 5% per cent in 1991. Slower trade growth is a factor behind slower Japanese growth in 1991. The weakness of the Yen in 1989 and 1990 is already boosting Japanese exports, and this, in combination with strong domestic demand and a strong contribution to GNP from net property income from abroad, is expected to raise Japanese growth to around 5.5 to 6 per cent in 1990. Japan is still a relatively closed economy, and the fall in the Yen is likely to have only a small impact on domestic inflation, which we expect to settle just above 21/2 per cent a year.

Recent developments in Europe have insulated the major economies there from the effects of the slowdown in the US. The prospect of developments in Eastern Germany has given the West Germans a boost to their growth, and hence has affected the prospects for the rest of Europe. French and Italian growth are both expected to fall in 1990 after two to three years of growth in excess of capacity. The slowdown is more marked in italy than in France. The latter country is currently expected to return to full capacity growth with low inflation relatively quickly.

Developments in Germany remain rather uncertain. The arrival of over a million migrants in the last eighteen months has raised the prospect for growth without excess inflation. We anticipate that the new migrants will be absorbed into the workforce relatively slowly. These workers, along with those who will now travel across the border to West Germany and West Berlin will raise the rate of growth of productive potential in the current Federal Republic. They will also put downward pressure on wages. The combination will, we expect, allow West German growth to average 3 per cent a year for much of the 1990s, compared to 2 per cent in the 1980s. Despite this acceleration in growth we are anticipating that inflation will not exceed the level experienced in 1989. Our forecast for German growth is relatively smooth, reflecting both the use of an estimated time series model and also a belief that market mechanisms take time to adjust. However there are clearly severe risks on both sides of our forecast. This may be the last time that we are able to produce a forecast for the Federal Republic alone. There are already plans to change the basis upon which statistics are produced. The Bundesbank intends to move the current account statistics over to an all-German basis immediately. This will be better than treating intra-German trade as domestic trade, and it will reduce the world current account discrepancy. it will make the comparison of new data with the recent past more difficult. It is to be hoped that national accounts data will continue to be produced for West Germany alone, as this will make the interpretation and hence control of events considerably easier. In the short term we do not anticipate that events in Eastern Europe outside the GDR will have any noticeable impact on developments in the West, although their longer-term implications may be significant.

(c) Longer-term prospects

Our base forecast is predicated on a belief that the current oil crisis is temporary and that the world economy resumes a process of steady growth relatively quickly. in the 1980s discussions of international economic problems were dominated by the issue of balance of payments imbalance. A combination of developments in Europe and structural changes in Japan appear to have reduced the potential scale of such imbalances. Chart 6 plots the current account to GNP ratios for the major 3 economies. This shows that we are expecting the scale of capital flows to gradually decline. The pattern of world trade imbalances that we are forecasting depends in part on our projections for world trade shares. The forecast appreciation of the Yen along with some structural change will produce some stabilisation of the Japanese trade share at around 7 per cent. Meanwhile we expect that the US will continue to regain its share of world trade which we expect to rise from 12 per cent in 1990 to around 14 per cent by the end of the decade.

We are anticipating that US pressure in the GATT round talks to reduce food production and export subsidies will be successful, and that this will contribute to the process of balance of payments adjustment in the world economy. Chart 7 gives past paths for real commodity prices along with our forecast and table 2 contains our forecast for all commodity prices. We expect that real metals' prices will continue to fall as the rate of technical change in mining and extraction is expected to continue to exceed the rate of growth in demand. However we are projecting rising free market real world food prices. This does not necessarily mean that prices to consumers in the advanced world will be any higher than they are currently. Agricultural subsidies are estimated to cost the advanced world $250 billion a year. There are at least three different farm support systems, all of which put downward pressure on free market prices. The US subsidises and supports production of many commodities. Prices are generally left to find their own level, although up to the mid-1 980s the US authorities had considerable stockpiles which were partly sold off during a vigorous export drive which put downward pressure on prices in 1985 and 1986. The Japanese basically prevent imports of rice, and domestic rice prices are more than ten times higher than world prices. The gradual reduction of these barriers will help support grain prices in general. The European Community has a system of import levies and export subsidies that keep European prices high and depress free market prices. This system is more visibly distortionary than that in the US, but it has no more damaging effects on the world economy. We are assuming that by the end of the 1990s real world food prices will have returned to the levels experienced in the first half of the 1980s. These developments will improve the US current balance and worsen those of both Japan and Germany. Exchange rates and interest rates Over the last three months the Dollar has weakened against all the major currencies, partly in response to declining US short rates, especially relative to those in Japan and Germany. Over the year to the end of July 1990 the Yen had depreciated by 14 per cent and the D-Mark had risen by 9 per cent in effective terms.

The weakness of the Yen has been a little hard to understand given the size of the Japanese current account surplus, and as we anticipated in our last Review, the Yen has recovered (by 3 per cent in effective terms) over the last three months. Over the last decade the Japanese current account surplus has averaged over 2 per cent of GNP, and in the 4 years prior to 1989 it averaged more than 3.5 per cent. Part of this large scale capital outflow reflects the gradual abolition of exchange controls and restrictions on portfolio composition. (These restrictions, and their implications for the rest of the world, are discussed in Pain (1990), which emphasises their importance to the UK). Flows resulting from portfolio reallocation can be seen as temporary, but we have argued elsewhere that there are demographic and other reasons for a sustained capital outflow from Japan (see Barrell and Wren-Lewis 1989). We expect those outflows to continue for some time, and hence a relatively large current account surplus has to be achieved. Annex 11 to this chapter shows that there has been some structural change in the Japanese trade accounts, and that at constant competitiveness the current account would now be some $36 billion higher if income and price elasticities had stayed the same as they were in 1985. We believe that the markets and the authorities had not fully taken account of these structural changes when they allowed the Yen to appreciate strongly into 1989. The rather rapid deterioration of the current account in 1989 can be seen as news', and it caused the market to re-adjust its evaluation of the Yen.

There has been much discussion of the FEER (fundamental equilibrium exchange rate) for the Yen and its relation to PPP (purchasing power parity). There is normally a large discrepancy between these two estimates of the equilibrium exchange rate, especially if the PPP rate is in excess of 200 Yen per Dollar, which would imply a much larger surplus on the current account. We prefer the FEER, which is based on both our estimates of the trade elasticities and on our judgement of the scale of structural capital flows. it is not possible to uniquely define an equilibrium bilateral rate against the Dollar when we discuss the FEER because it is the level of competitiveness against all currencies together that determines the size of the current account. However, at current exchange rates for currencies such as the D-Mark and the Franc we would judge that the FEER for the Yen is around 140-5 Yen per Dollar. The evolution of the Dollar Yen FEER over time depends not only on the complete pattern of bilateral rates but also upon the prospective paths for prices in all the countries involved. As the Yen is at around its FEER and as Japanese inflation is expected to stay below that in the US and elsewhere we would expect the Yen to continue to appreciate in nominal terms in order to stay around its real FEER level. We also expect the scale of structural capital flows to decline over the next decade, and this will put further upward pressure on both the real FEER and, we presume, the nominal effective exchange rate. Table 3 gives our projections for all exchange rates and chart 8 the past and expected paths for effective exchange notes.

It is much more difficult to assess the prospects for the D-Mark in terms of its FEER. Until last Autumn we had been expecting large scale capital outflows from Germany to continue for some time. These reflected not only the savings behaviour of the population but also the relatively low rate of return to real investment in Germany compared to those available elsewhere. Developments since November have changed our judgement. Firstly the influx of migrants is likely to reduce savings and also, by putting downward pressure on real wages, raise returns to investment. Secondly unification with Eastern Germany will open up considerable scope for further investment both in the Federal Republic, and also in the East, which for current account purposes has always been treated as part of Germany. As a result we expect that capital outflows will fall.

It is easy to predict that capital flows will decline, and to presume that the D-Mark will be under pressure to appreciate, and the recent strong appreciation of the D-Mark reflects this pressure. However it is difficult to judge the level of the D-Mark at which the FEER will be achieved. If all other exchange rates stay round about their current levels we would judge that the FEER for the D-Mark against the Dollar would be around 1.55 to 1.65 D-Mark per Dollar. However there is considerable uncertainty as to this range. In particular it depends upon our estimate of the Sterling/D-Mark FEER. it is difficult to assess the UK FEER as it depends on our estimates of capital flows (see Wren-Lewis et al where the issue is discussed more fully using the Institute's domestic model).

The judgement of the level of equilibrium exchange rates is important when assessing the prospects for stability in the European Exchange Rate Mechanism, and also for evaluating the prospects of the successful formation of a European Monetary Union. Even inside a Monetary Union the current balance of any component member will be driven toward some sort of equilibrium. If on entry the exchange rate is such that a large current account deficit exists then there will be a tendency for wealth to decline. if the rate of change in wealth exceeds that desired by savers they will adjust their consumption plans in order to reach their target wealth path. Individuals in the economy will only reach their target wealth path when the current account deficit is consistent with that target. Our estimates of the FEER depend upon a current account that is consistent with domestic equilibrium. Although there are automatic forces causing adjustment these are likely to be slow acting, and it would seem desirable for entry to EMU (or even the ERM) to be judged in the light of the relationship between the real exchange rate and the FEER for each member of the Union.

We are assuming that these considerations affect the policy making process in Europe, and that we will see a two speed progression to union. Given current exchange rates and projected inflation paths and fiscal policies the core European economies of France, Germany and the Benelux countries are assumed in our forecast to successfully form a monetary union in 1995. Italian membership of a Monetary Union will have to wait until their fiscal stance is less expansionary. Given current policy plans we think that membership will be a serious possibility by 1997. In that year we are assuming that both italy and the UK will be able to join the Union without any major disruption to their economies. Of course, if Italian fiscal policy tightens more rapidly then we are projecting so that the equilibrium wealth trajectory is reached earlier, then they may be able to join the Union in 1995. Similar caveats concerning inflation and the exchange-rate projections have to be made about our assumption that the UK joins the Union in 1997.

Table 4 gives our forecast for interest rates in 1990 and thereafter. As we are assuming that the shock to the oil price is temporary we see no need for a tightening of monetary policy. in the short term we are expecting that Japanese rates will rise because of rising concern over inflation. Elsewhere we expect that rates will begin to decline slowly as inflation falls from its 1990 peak. We assume that in the long run real interest rates will be around 3 per cent, slightly higher in current balance deficit countries and slightly lower in surplus countries. As Monetary Union approaches interest rates in Europe have to converge. This is only partly because the authorities will be pushing rates together. The major convergence will, however, result from market forces. Although it is possible to justify country specific risk premia even in a monetary union (as there are state specific risk premia in the US) these must be limited in scale. We assume that there will be no risk premium on the French Franc (or more correctly to ECU holders in France) vis a vis the D-Mark, and that as a result of Union German rates will be a little higher, and French rates a little lower than they would otherwise have been. We are assuming that Italian rates in general stay above those in the rest of the Union. Barrell (1990) argued that such differentials could exist because of portfolio preferences. However, the 'risk premium' must be much lower than the ex-post 4 per cent observed between Germany and Italy in the 1980s. This was probably the result of both effective capital controls and a more successful anti-inflationary stance than the markets had anticipated. The United States Since our May forecast there has been a revision of the US national accounts statistics. The revised figures show US GNP growth of 2.5 per cent in 1989 compared to 3.0 per cent in the previous estimates. The most significant revision has been to consumers' expenditure, now estimated to have grown by 1.9 per cent in 1989, down from 2.7 per cent previously. The figures for real personal disposable income have also been revised sharply from 4.1 per cent to 2.4 per cent for 1989, so that despite the lower growth of expenditure, the saving ratio in 1989 is also now thought to be lower than previously estimated.

Preliminary estimates of GNP are available for the second quarter of 1990. These indicate GNP growth of 1.2 per cent compared to a year previously, with domestic demand growing by 1.1 per cent and net exports adding 0.1 percentage points to GNP growth. GNP growth in the year to the first quarter is estimated at 1.3 per cent, with domestic demand growing by 0.9 per cent. It now appears that GNP growth in 1990 will be the lowest since the recession of 1982.

The weakening of consumers' expenditure growth that appears in the revised 1989 figures has continued into the first half of 1990. in the second quarter of 1990 consumers' expenditure is estimated to have been only 1. 1 per cent higher than a year previously. The decline in consumers' expenditure growth is associated with a lower growth of real personal disposable income, which in turn appears to have been caused by a lower growth in employment in the fourth quarter of 1989 and the first half of 1990. investment expenditures have also been weak in recent months. In the second quarter of 1990 housing investment was 3.5 per cent lower than a year previously and business investment only 0.2 per cent higher. Domestic demand growth has therefore entered a cyclical downturn following seven consecutive years in which GNP growth exceeded 2.5 per cent per annum.

The decline in the rate of growth of demand has not yet had a significant effect on the rate of inflation. In June consumer price inflation stood at 4.7 per cent, down from 5.3 per cent in February, but only fractionally below the 4.8 per cent for 1989 as a whole. The appreciation of the Dollar over the first 3 quarters of 1989 will have reduced imported inflation in the early months of this year, but our forecast of a subsequent depreciation, together with higher oil prices will cause import prices to rise again in the second half of the year.

The labour market shows some signs of weakening demand. Total non-farm payrolls increased by 123,000 per month in the second quarter compared to gains of 160,000 per month in the second half of last year. Manufacturing employment fell by 27,000 in May and 31,000 in June. Unemployment, on the other hand, also fell in June to 5.2 per cent, suggesting that the labour market remains fairly tight. Weekly earnings in the second quarter were 3.7 per cent higher than a year earlier, up from 3.4 per cent in the first quarter.

Monetary policy has remained unchanged, despite the evidence of a deceleration of the rate of growth. This has led to some speculation that the Federal Reserve would lower interest rates. The Fed Funds rate was eased from 8.25 per cent to 8.0 per cent in mid-July, but the official discount rate remained unchanged at 7 per cent. Our assumption is that the Fed is likely to maintain interest rates at current levels for the remainder of the year. A reduction might occur, however, if third quarter GNP figures suggest that the downturn was developing into a recession, although the inflationary impetus of the rise in oil prices is likely to reinforce the case for maintaining current levels.

Our forecast for the United States is presented in Table 5. The data available for the first half of this year suggest that 1990 will see the lowest growth-rate since 1982, at 1.5 per cent. The main cause is the slower growth of domestic demand, with consumers' expenditure and non-housing investment growing less strongly, and housing investment again in recession. Net exports should however continue to add to GNP growth, aided by a depreciation of the Dollar since the third quarter of 1989, and by the slow growth of domestic demand while export markets remain relatively buoyant.

Some commentators have expressed a fear that the decline in the growth of economic activity may develop into a recession. This is a risk that cannot be completely discounted, especially if the disruption in the oil market further reduces business optimism. We expect, however that consumers' expenditure will pick up again in the second half of the year, providing a stimulus for investment and GNP growth. Our forecast for 2% per cent growth in 1991 is based on an oil price of $20 per barrel by the middle of the year. The simulation contained in Annex 1 indicates that if oil prices remain $25 per barrel then US growth in 1991 would be around 21/4 per cent.

Inflation this year is expected to be 4.7 per cent. On the assumption of oil prices of $20 per barrel we expect that inflation will decline to 3.9 per cent next year, but an oil price of $25 per barrel would result in inflation remaining at around 4.5 per cent. In these circumstances it is unlikely that the gradual easing of interest rates that we have assumed in our main forecast would actually take place.

The US current account deficit is a half solved problem. The depreciation of the Dollar from its peak in early 1985 has enabled a deficit of $144 billion in 1987 to be reduced to $1 1 0 billion last year. Our forecast, presented in Table 6, indicates further declines this year and next. As a proportion of GNP the deficit has reduced from 3.2 per cent in 1987 to a forecast 1.8 per cent in 1990. it is unlikely that gains will be as rapid in the years to come, primarily because we expect that the nominal value of the Dollar will remain at around its current level against a basket of other currencies. Our medium-term forecast suggests that the current account deficit has now been reduced to a sustainable level, but that the problem has not been completely eliminated since the size of the deficit will constrain the rate of US domestic demand growth. We anticipate that fiscal policy in particular will have to remain tight in order to prevent a re-escalation of both the current account and the public sector deficits.

Table 7 gives our projections for the US fiscal deficit. The Federal deficit for the first 9 months of fiscal 1990 stood at $163.1 billion, well in excess of the Budget prediction of $124 billion and the Balanced Budget net target of $1 00 billion. The target for fiscal 1991 is $64 billion, but it seems very unlikely that it will be achieved. Growth has been slower than anticipated this year and hence revenues are likely to be lower than originally forecast. In the longer term there remains the prospect of cuts in defence expenditure, but the cost of guarantees extended to the troubled savings and loan sector will also be higher than estimated a year ago. We are projecting that these could add over $10 billion to spending in the short run, although this may be offset by asset sales in the longer run. President Bush has now conceded that tax increases will be necessary to reduce the deficit, but there are still doubts about political resolve, especially since there are Congressional elections later this year. We anticipate that the Federal deficit will be gradually reduced, but at a slower rate than prescribed by the Balanced Budget Act. The deficit of the public sector as a whole is currently around $40 billion less than the Federal deficit, and is forecast to be in surplus by 1997, mainly because of an increased surplus on the Social Security Fund.

Japan

A sustained bout of uncertainty afflicted the financial side of the Japanese economy in the first half of this year. The Yen depreciated by 10.5 per cent between the end of January and the end of April, stock-market prices fell by 28 per cent between the end of December and the beginning of March, and long-term interest rates rose by 1.2 percentage points between December and March. These developments cannot be attributed directly to the performance of the real economy, which grew by 5.6 per cent in the year to the first quarter of 1990. Nor do they seem related to the prospects for inflation, which while rising from 1.7 per cent in 1989, registered only 2.6 per cent in the first quarter of 1990.

The adverse reaction of both financial and foreign-exchange markets is probably best explained in terms of a correction of their previous misalignment. Shares on the Japanese stock-market have for some years had markedly higher price-earnings ratios than similar shares in other markets, which could only be economically justified if there was a compensating expectation of substantially higher earnings growth for Japanese firms. The appreciation of the Yen between 1985 and 1988 resulted in a reduction of the current account surplus from 4.3 per cent of GNP in 1986 to 2.0 per cent in 1989. This was perhaps excessive given the extent of Japanese capital outflows, necessitating some depreciation of the Yen to bring current and capital account flows into balance. The simultaneous adjustment of stockmarket portfolios may have resulted in a temporary increase in capital account outflows, exacerbating the short-term depreciation of the Yen. We retain our assumption that the Yen will appreciate in the second half of this year to 146 Yen/Dollar in the fourth quarter, as this temporary outflow subsides.

The Japanese monetary authorities have faced a dilemma in responding to these developments. On the one hand the weakness of the Yen could lead both to an increase in imported inflation and to an increase in external demand for Japanese goods. Indicators of capacity utilisation suggest that the economy is already close to capacity output, and hence there is a high risk that extra demand will further fuel inflation. Our model equations for Japanese wholesale and consumer prices contain significant effects from capacity utilisation. On these grounds there was a case for raising interest rates. On the other hand an increase in interest rates risked throwing financial markets into greater turmoil. in the event the official discount rate was raised by 1 percentage point in March to 5.25 per cent, following rises of 0.5 percentage points in October and December of last year. Monetary policy is therefore much tighter than a year ago. This has helped to stabilise the value of the Yen, which appreciated 3.5 per cent between the end of April and the end of July. Even so the Yen was 12.5 per cent lower than in July 1989, and there are fears that interest rates will have to be raised further in order to dampen the inflationary consequences. Renewed weakness in the Yen evident in early August and the rise in the oil price following Iraq's invasion of Kuwait make a further increase more likely. We have assumed that interest rates are increased 0.5 percentage points in the third quarter of 1990, and start to come down again during 1991.

Japanese GNP grew by 5.6 per cent in the year to the first quarter of 1990, up from 4.8 per cent for 1989 as a whole. Domestic demand grew by 5.5 per cent and net exports added 0.1 percentage points to GNP growth. Domestic demand growth continued to be led by business investment, which was up by 14.5 per cent in the year to the first quarter of 1990 following rises of 15.7 per cent and 17.7 per cent in 1988 and 1989. In its May Short-term Economic Outlook the Bank of Japan reported that many manufacturing firms had revised up their capital investment plans from an increase of 7.5 per cent to one of 16.5 per cent for the year ending March 1991. Other business survey evidence suggests both continuing optimism about production prospects and a high level of capacity utilisation. It therefore seems likely that the high level of investment will continue for a third successive year. This should raise productive potential and ease existing capacity constraints.

Japanese labour market indicators reveal that the market is extremely tight. Seasonally adjusted unemployment stood at 2.1 per cent in May, up slightly from March, but still around the lowest level for 1 0 years. The ratio of job vacancies to job seekers rose to 1.37 in February, the highest for 16 years, and has since eased marginally. While employment growth has been strong, the tight labour market has been eased by the increasing participation of female workers, particularly in service industries. The effect of the tighter labour market was also evident in the settlements reached in the annual spring wage round, which averaged around 6 per cent this year, compared with 5.1 per cent in 1989 and 4.4 per cent in 1988.

Our forecast for Japan is presented in table 8. We expect GNP growth to increase to 5.9 per cent this year. Domestic demand growth is expected to fall to 4.9 per cent but net exports could add 1.0 percentage points to GNP growth. Net exports are also expected to add 1. 1 percentage points to GNP growth in 1991. This strong performance is a consequence of the depreciation of the Yen, which fell by 11.5 per cent between the fourth quarter of 1988 and the fourth quarter of 1989, and a further 1 1.0 per cent by the second quarter of 1990. The Yen is expected to appreciate in our forecast by around 2 per cent per annum, and so does not reach its 1989 fourth quarter level until the first quarter of 1996. In the interim there is a strong competitiveness gain. In this forecast we have used the new Japanese trade equations reported in Annex II. The new equations for export volumes and non-oil import volumes both have high competitiveness elasticities, and hence suggest that the depreciation of the Yen will have a significant effect.

Domestic demand growth is expected to be sustained by continued strong growth in business investment, in line with survey evidence. The growth of consumers' expenditure is likely to be around 3.0 per cent, slightly less than last year. Real personal disposable income growth has been reduced by higher than expected inflation, and consumers' expenditure has also been curbed by higher interest rates and a new sales tax introduced last year.

Inflation is expected to rise to 3.0 per cent this year. This increase can be attributed in part to the recent weakness of the Yen which may have contributed up to 1/2 of the increase. However, tight labour markets and high capacity utilisation have also put upward pressure on costs. Inflation in 1990 will however no longer be boosted by the new sales tax introduced in April 1989.

Our main forecast is based on an oil price of $20 by the middle of next year. This leads to a decline in Japanese growth to 41/4 per cent in 1991. The principal cause is a decline in domestic demand, following the increase in interest rates which we are expecting in the second half of this year. The growth in business investment declines to 3.5 per cent and the level of residential investment falls by 2.3 per cent. We anticipate a further increase in inflation to 31/2 per cent, as the effects of the Yen's depreciation and the increase in oil prices continue to be felt. However we also expect the Yen to start appreciating, so that the inflationary impulse arising from a weak currency will start to reverse.

Germany

German GNP grew by 4.4 per cent in the year to the first quarter of 1990. This was stronger than the growth-rates of 3.0 per cent and 3.7 per cent recorded in the year to the third and fourth quarters of 1989. German GNP is forecast to grow strongly this year and next, partly through the stimulus expected from German economic union, and the expected political union. However the pattern of expenditure contained in the first quarter GNP figures cannot be wholly attributable to the effects of reunification. Consumers' expenditure benefited from cuts in direct taxation at the start of the year, and investment in construction was boosted by another mild winter. Export growth was also strong in the first quarter.

Germany is now entering its eighth consecutive year of economic growth and third consecutive year of growth in excess of 3 per cent. During the latter period the rate of capacity utilisation in manufacturing industry has risen from 84.1 per cent to 89.7 per cent. However inflation has remained under control. There was a temporary rise to 3.1 per cent in 1989, following the increase in indirect taxes at the start of that year, but inflation in 1990 has eased slightly. This is due both to the absence of the indirect tax effect which increased inflation in 1989, and to the appreciation of the D-Mark in late 1989 which has resulted in lower imported inflation. In May 1990 import prices were 5.4 per cent lower than a year earlier, wholesale prices 0.2 per cent lower and consumer prices 2.3 per cent higher.

In the short term further progress in reducing inflation is likely to be limited, given rising cost pressures in the labour market. Recent wage settlements of 6.3 per cent in the construction industry and 6.8 per cent in the printing industry are well above the rate of inflation. Good company profits, a high rate of capacity utilisation, and the expectation that strong economic growth will be sustained in both the short and medium-term have all contributed to the increases. However, the influx of migrants into the labour force should be putting immediate downward pressure onto wages and prices.

After raising interest rates on four occasions in 1989, the Bundesbank has maintained them at the levels set last October. Monetary policy has therefore remained constant despite some expectations that German economic and monetary union might lead to further tightening. The decline in inflation that has occurred this year may be seen as justifying this stance. There seems to be little evidence that German economic union has resulted in a surge of consumer spending, but there is a possibility that a further interest-rate rise may be provoked by the consequences of economic union. Interest rates are unlikely to fall until the Bundesbank is convinced that inflationary pressures have been curtailed. As we expect strong economic growth to continue this year, we do not anticipate any reductions in interest rates until next year at the earliest.

Economic and monetary union has altered the likely stance of fiscal policy. Government spending will rise in order to finance unemployment benefits and pensions. In addition East Germany requires considerable investment in its infrastructure, and much of this may have to come from public spending. Government revenue grew more strongly than had been expected in 1989 due to the higher than expected rate of growth, and continued strong growth of economic activity will help to finance some of the extra spending.

Our forecast for West Germany is presented in table 10. Following the strong growth of GNP in the first quarter we now expect that growth for the year will be 3.8 per cent. Consumers' expenditure is expected to grow by 4.5 per cent, boosted by the tax cuts at the start of the year, and by the spending of the immigrants who entered Germany at the end of last year. investment expenditure is also likely to grow strongly, with investment in housing boosted by the extra demand associated with immigration, and business investment by the stimulus of monetary

We expect that GNP growth will decline to 31/4 per cent next year, as consumers' expenditure grows less strongly. The stimulus of direct tax cuts will no longer be operating, although there will be some additional impetus provided by continuing immigration and higher demand for West German goods from the East. Investment expenditure should also remain strong as the German economy is re-structured to supply its expanded market.

Our forecast for inflation is 2.9 per cent for 1990, and 3.0 per cent for 1991. it is still somewhat early to judge the effect of German monetary union, although preliminary evidence suggests that the surge in demand for West German goods that some commentators had feared did not occur. The East Germans appear to have been cautious in their spending, which is understandable given the likelihood of rising inflation and unemployment in the East.

The effect on East German unemployment appears to be in line with the more pessimistic predictions. Many of the old East German enterprises appear to be unviable within the union without substantial investment from the West. Unfortunately investment from the West has been less vigorous than many had hoped. An important factor here is the confusion that remains over ownership of East German firms' assets. The initial monetary union agreement embodied the right of repossession by some pre-1948 owners of assets. This is an area that requires urgent attention in order to speed up the integration of the economies of East and West. Failure to tackle the problem will mean that East Germany will continue to suffer from insufficient investment, and the consequent high unemployment will mean that it will act as a drain rather than a spur to the German economy.

Table 12 presents our forecast for the German public sector. Strong revenue growth in 1989 led to a surplus of D-Mark 5 billion, equivalent to 0.2 per cent of GNP. This was the first annual surplus since 1973, and represented a sharp improvement from the deficit of D-Mark 45 billion recorded in 1988. As 1990 is forecast to be another year of strong growth, revenues should again grow strongly, although the income tax cut introduced at the start of the year will cost the government D-Mark 25 billion. Expenditures will also rise strongly this year because of the costs of economic and monetary union. The Federal Budget of December 1989 has had to be supplemented by two further budgets making provisions for D-Mark 11.8 billion of extra spending. We anticipate that the general government will return to deficit this year of around D-Mark 18 billion. The costs associated with political union are likely to lead to to continuing deficits in the early 1990s despite buoyant revenue growth derived from the continuing strength of German economic activity.

BOX 1. THE TREATMENT OF GERMAN UNIFICATION IN OUR FORECAST

Our forecast is constructed using our econometric model, GEM. The equations in GEM have been estimated from data on the West German economy, and reflect its performance over the last twenty-five years. The following special adjustments have had to be made to allow for the likely effects of German unification on the behaviour of the West German economy over our forecast period. Some are discussed more fully in the World Economy chapter of the May 1990 National Institute Economic Review.

(a) Consumers' expenditure. We expect this to grow by more than our equation would predict because of the extra expenditures by immigrants into West Germany, and because of additional spending in the West by East Germans following monetary union. This latter effect will be included in the National Accounts as spending by West German citizens.

(b) Investment This too should grow by more than our equations would predict. Residential investment will be boosted by the extra demand arising from immigration. Business investment is expected to be higher as German industry equips itself for meeting both the demand of the expanded German market for both consumer and producer goods, and also increased demand from other East European countries.

(c) Employees' wages and salaries. We expect downward pressure on average earnings as the economy absorbs an expanded labour supply, both from immigrants and also from workers crossing the border from lower wage areas in East Germany where they are not used to receiving western levels of pay. This downward pressure will help to reduce some of the inflationary pressures arising from higher demand, and should help the German economy to grow more rapidly than in the 1980s without incurring inflationary bottlenecks.

(d) Public sector. We have made various adjustments to our model of the German public sector. Firstly transfers have been increased, by DM 15 billion a year, in the early years of the decade. This represents both payments to migrants and support for the East German economy. Secondly we have raised West German public debt by DM30 billion. This represents the sum that the Federal government has taken over at par to ensure that the banking sector in the East remains solvent.

(e) Balance of payments. We have made some adjustments to our trade equations. The German authorities have never treated intra German trade as external trade, and hence although exports to the East appeared in the national accounts they did not appear in the trade accounts. Our export volume data came from the trade accounts and hence should be basically unaffected by the changes in East Germany. We have however made some adjustment to our German exports forecast on the assumption that they increase their share of Central European markets over the next decade.

France

French GDP rose by 3.6 per cent in 1989. Although this was lower than the 3.9 per cent growth experienced in 1988 it was nonetheless higher than the government's forecast of 3 to 31/2 per cent. The growth of the French economy during 1989 can be attributed to the sustained expansion of business investment and buoyant exports, as well as strong growth in consumers' expenditure. Business investment grew by 7.2 per cent in 1989, reflecting both the profitability of and capacity constraints on French industry. Capacity utilisation was 86.0 per cent in the fourth quarter of 1989, its highest level since 1973. The volume of exports grew by 8.3 per cent in 1989, aided by a lower exchange rate, lower labour costs (trend unit labour costs fell for the third year in succession), and strong demand growth in the rest of the world. Net exports contributed positively to the growth of the French economy for the first time since 1984. Unemployment continued to fall in 1989, whilst inflation increased modestly from 2.7 per cent in 1988 to 3.3 per cent in 1989.

The overriding objective of French monetary policy has been to maintain the value of the Franc vis-a-vis the D-Mark. To this end, monetary policy was progressively tightened during 1989 as interest rates were raised in Germany. This tightening led the Franc to appreciate to its highest level in effective terms since mid-1986, causing the Banque de France to cut its intervention rate by 0.5 percentage points in April of this year. Our forecast assumes firstly that the exchange rate moves in line with risk adjusted interest-rate differentials and also that a monetary union between France, Germany and the Benelux countries is formed in 1995.

The growth of real GDP slowed to 0.6 per cent in the first quarter of 1990 from 0.9 per cent in the fourth quarter of 1989. Strong growth in consumer spending and also in fixed investment was more than offset by a decline in industrial output and stocks. The slowdown in the French economy has manifested itself both in terms of higher unemployment and lower inflation. Unemployment rose by 3200 to 2.512m in June, although unemployment is still 0.6 per cent lower than it was a year ago. Consumer prices rose by only 0.2 per cent in June and the inflation rate fell to 3.0 per cent, the lowest since 1988.

Our forecast for the French economy is set out in table 13. in 1990 we are forecasting that the growth of the French economy will slow further to 2.6 per cent. Consumption growth is forecast to remain strong through 1990, reflecting buoyant real personal incomes. (We have recently recast our French personal income database and re-estimated all the associated equations giving us a much firmer base for our forecast.) Although higher interest rates will result in a slowing of private investment growth to 4.2 per cent in 1990, compared to 5.7 per cent in 1989, we expect the growth of domestic demand to decline only marginally. Therefore the slowing of the French economy in 1990 may be attributed almost entirely to the deterioration in the external environment. A higher exchange rate and a slowing in the growth of world trade will result in net exports reducing GDP growth by 0.8 per cent in 1990. However, a higher exchange rate and higher interest rates are likely to exert renewed downward pressure on inflation, and we are forecasting that inflation will fall to 2.7 per cent this year. These developments, coupled with the strong growth of domestic demand in Germany, are liable to result in French inflation falling below that in Germany, albeit marginally. This is likely to give the authorities the opportunity to make further cuts in interest rates later this year, and we are assuming that short-term interest rates are reduced by 0.4 percentage points between the second and fourth quarters of 1990.

The deterioration in the external environment and the slowing of the French economy are liable to have a profound effect upon the export and import volume of goods in 1990. We are forecasting a significant reduction in the growth of trade volumes in 1990: exports of goods are forecast to grow by only 4.2 per cent in 1990 after their strong performance in 1989, whilst imports are set to grow by 4.4 per cent in 1990 as compared to 7.3 per cent in 1989. We are forecasting that the current account will move further into deficit in 1990 despite an improved visible balance. The immediate decline in the invisibles balance in response to an appreciation is forecast to more than offset the J-curve effect on the visible balance. This improvement can be attributed to an improvement in the terms of trade. In 1991 we are forecasting a significant deterioration in, the current account, reflecting both the continued movement into deficit of invisibles and a worsening in the terms of trade. Italy

Italian GDP grew by 3.2 per cent in 1989, lower than the 4.2 per cent growth experienced in 1988. The economic slowdown in 1989 may be attributed partly to the tightening of economic policy implemented during the latter half of the year. The move to a more restrictive policy stance checked the growth of domestic demand, which grew by 3.3 per cent in 1989 as compared to 4.7 per cent in 1988. The strength of demand in the rest of the world, however, meant that the contribution of net exports to GDP growth, although still negative, improved during 1989. Inflation rose during 1989, rising to 6.1 per cent from 4.8 per cent in 1988, reflecting both persistent pressures on capacity and an acceleration in unit labour costs (which rose by 7.6 per cent in 1989 as compared to 3.4 per cent growth in 1988).

The monetary authorities cut the discount rate by 1 percentage point in May as the decision to incorporate the Lira within narrow bands in the ERM resulted in large capital inflows and upward pressure on the Lira. Our assumption of a two-speed Europe implies that the Lira moves in line with risk adjusted interest-rate differentials until 1997, thereafter becoming irrevocably fixed against the D-Mark and the Franc as Italy begins to participate in the monetary union formed between Germany, France and the Benelux states in 1995.

In May 1990 the italian government introduced a fiscal programme aimed at reducing the budget deficit in line with the targets contained in the 1990 Finance Act, even though the 1990 budget projections were revised upward in March. In its Economic Outlook published in June this year the OECD has assessed the sustainability of fiscal policy in each of its member states. Significantly the OECD use a forward-looking indicator of sustainability which not only takes account of foreseeable changes in spending and revenues, but also the degree of expenditure reduction/revenue enhancement which is required to maintain a stable debt-to-income ratio. On the basis of current and expected budgetary measures the OECD concludes that italian fiscal policy is unsustainable. (This is in line with the argument presented in Barrell (1990) where Italian membership of a monetary union was only seen as feasible if fiscal policy tightened considerably.) We are assuming therefore that fiscal policy will tighten further during the 1990s reducing the deficit from 10 per cent of GNP to around 4-5 per cent of GNP by the end of the decade. A tighter fiscal stance will not only restrain the growth of the national debt and lower the cost of capital, but will in the short term also increase national savings. An increase in national savings, ceteris paribus, will tend to improve the current account. A tighter fiscal policy will also bear down on inflation. These latter two consequences of a tighter fiscal policy will not only reduce the differential between Italian inflation and inflation in Germany and France, but will also reduce the possibility of a major realignment of the Lira before Italy participates in monetary union.

Our forecast for the italian economy is set out in table 15. With a higher exchange rate and continued high interest rates, and an expected decline in world demand, the Italian economy is expected to grow more slowly this year. Domestic demand growth is forecast to slow to 2.5 per cent this year as consumption and investment are restrained by higher interest rates. Net exports are forecast to reduce the growth of GDP by 0.7 per cent in 1990. However, the current account is expected to improve this year, despite a deterioration in the invisibles balance. This improvement can be attributed to the favourable movement in the terms of trade that we are forecasting for 1990. Whilst we are not anticipating any marked change in the italian inflation rate this year, we expect inflation to fall in the long run as continued membership of the ERM disciplines the behaviour of wage and price setters.

Canada

In 1989 Canadian GDP grew by 3.0 per cent, lower than the 4.4 per cent growth experienced in 1988. This slowdown can be attributed to both a reduction in the rate of growth of domestic demand and a significantly more negative contribution from net exports. The growth of domestic demand slowed to 4.1 per cent in 1989 from 5.0 per cent in 1988, partly in response to the sustained high level of Canadian interest rates. The Canadian Dollar continued its appreciation against the US Dollar during 1989, rising by some 3.8 per cent, this in turn reflected an unexpected widening of the differential between Canadian and US interest rates. During 1989 the current account deficit widened to 2.6 per cent of GDP from 1.7 per cent of GDP in 1988. This deterioration in the current account is the result of a deterioration in the invisibles balance. This improvement can be attributed to the favourable movement in the terms of trade that we are forecasting for 1990. Whilst we are not anticipating any marked change in the italian inflation rate this year, we expect inflation to fall in the long run as continued membership of the ERM disciplines the behaviour of wage and price setters.

Canada

In 1989 Canadian GDP grew by 3.0 per cent, lower than the 4.4 per cent growth experienced in 1988. This slowdown can be attributed to both a reduction in the rate of growth of domestic demand and a significantly more negative contribution from net exports. The growth of domestic demand slowed to 4.1 per cent in 1989 from 5.0 per cent in 1988, partly in response to the sustained high level of Canadian interest rates. The Canadian Dollar continued its appreciation against the US Dollar during 1989, rising by some 3.8 per cent, this in turn reflected an unexpected widening of the differential between Canadian and US interest rates. During 1989 the current account deficit widened to 2.6 per cent of GDP from 1.7 per cent of GDP in 1988. This deterioration in the current account is the result of a worsening in the visible balance, which can be attributed to the sustained appreciation of the Canadian Dollar over the last three years.

After initially falling in early 1990, interest rates were raised to seven-year highs in March as the exchange rate came under intense pressure. We anticipate that Canadian interest rates will stay around their present high levels for the remainder of the year and that the Canadian Dollar will appreciate further, albeit modestly, against the US Dollar in 1990. in the longer run we expect the Canadian Dollar to depreciate against its US counterpart as the exchange rate moves down the risk adjusted open-arbitrage path.

Our forecast for Canada is presented in table 16. We anticipate that the growth of the Canadian economy will slow during the course of 1990 as high interest rates bear down on the growth of domestic demand. Canadian GDP is forecast to grow by only 1 per cent in 1990 despite some recovery in net exports. Industrial production is set to fall by almost 1 percentage point this year. The slowdown in 1990 will be accompanied by both higher unemployment and lower inflation. The unemployment rate is forecast to increase to about 8 per cent in 1990 as compared to 71/2 per cent in 1989. Consumer price inflation is projected to decline to 4.4 per cent this year, reflecting both the slowdown in domestic demand and the strength of the Canadian Dollar. However, inflation is set to rise in 1991 following the introduction of the Goods and Services Tax. The strength of the Canadian Dollar, together with a slowing in demand in the rest of the world is likely to result in a further deterioration in Canada's external position. The current account deficit is forecast to widen to 3.9 per cent of GDP in 1990, a position which is not reversed by the end of the decade. Such a deficit can only be sustained in the long run if Canada is able to attract the requisite capital flows form the rest of the world. We think that the resource rich Canadian economy will be able to-attract such funds. If these flows are not forthcoming then there will be a need to adopt an even more restrictive policy stance.

REFERENCES Barrell, R. (1 990), 'European currency union and the EMS', National Institute Economic Review, No. 132, pp.59-66. Barrell, R.J. and Wren-Lewis, S. 1989),'Fundamental equilibrium exchange rates for the G7', National Institute Discussion

Paper No. 155. Pain, N. (1990), Financial liberalisation and foreign portfolio investment in the United Kingdom', National Institute Discussion

Paper No. 184. Wren-Lewis, S., Westaway, P., Soteri, S. and Barrell, R. (1989), Choosing the rate: an analysis of the optimum level of entry

for sterling into the ERM', National Institute Discussion Paper No. 171. Recent events in the Gulf indicate that there is a possibility that the world economy may see a significant rise in oil prices. On 6th August 1990 the iraqis were still occupying Kuwait, and their annexation of the country gave them control of 8 per cent of the worlds oil supplies. Their intention is to use their bargaining power to lever up oil prices. These had been relatively weak for some time as Iraqi and Iranian production had both returned to full capacity after the Gulf war and some producers, such as Kuwait, were producing above quota. This annex investigates the effects of a rise in oil prices and compares the estimates to those produced by the OECD in 1986. We concentrate on a $5 pb rise to $25 pb, but we also include some results for a $10 pb rise to $30 pb.

(a) A temporary rise in oil prices

We have assumed for our main simulations that the rise in oil prices is only temporary. The oil price is assumed to rise by $5 pb and stay 25 per cent above base for 3 years. Table Al given the effects on output and inflation in the G7, along with the effects on world trade. Table A2 gives the associated effects on exchange rates. We have assumed that: - (1) Exchange rates are forward looking.

(2) Monetary policy is not accommodating. and (3) Fiscal policy is unchanged from base

assumptions.

These assumptions are more fully spelled

out below.

The rise in world oil prices shifts resources from the advanced world to the oil producers in OPEC, in Latin America, Africa and to the CPES. It initially raises inflation by half a per cent in the first year and then by a quarter of a per cent in each of the subsequent two years. Output initially declines in the major economies as higher prices and higher interest rates reduce both consumption and investment. We project that output would be 0.3 per cent below base by the second year, with some recovery thereafter. Industrial production would also be below base. However world trade displays a different pattern. Lower incomes in the advanced world would initially reduce trade relative to base levels, but as oil producers receive extra revenue they increase their imports. World trade rises above base in the second year of the simulation and continues increasing in the third year. This rise in the ratio of trade to output is the real concomitant of the change in the terms of trade against the advanced economies.

These differences from base are slightly smaller (for a similar shock) than those given by the OECD in the May 1986. Economic Outlook May 1986 pp164-165). The OECD ready reckoner has OECD output in the major 7 0.2 per cent below base in the first year and 0.4 per cent in the second. This is in part the result of declining energy intensity of output. We have assumed that the monetary stance will not be accommodating, and hence that real interest rates will be unchanged in response to the shock. Table A2 gives the effects of the oil price shock on interest rates and exchange rates amongst the major seven. Table A3 gives the associated changes in inflation and output. It is clear that we should expect some diversity of outturn. US inflation rises more than elsewhere. The US is both more energy intensive (although less oil intensive) and has freer energy markets than any of the other major economies. Unlike in Europe and Japan, gas, coal and electricity prices can be expected to quickly follow oil prices. Our model reflects this sensitivity and oil prices directly affect producer and consumer prices. As a result we would expect the US Dollar to fall relative to other major currencies and US interest rates to rise more than elsewhere.

The effects on exchange rates elsewhere depends upon a number of diverse factors. Countries such as Germany and Japan that have rather slow rates of pass through of external shocks to domestic prices are likely to gain in terms of competitiveness relative to those with a more rapid pass through to prices. However the Japanese are more dependent on free market imported energy sources, and the effect on their balance of payments is worse than that for Germany. We expect the Yen to depreciate initially along with the Dollar.

Not all currencies can depreciate, and as the Yen and Dollar tall, the D-Mark strengthens. This is not only because of the small effect of the oil price rise on inflation but also because they trade relatively heavily with oil producers. This link is particularly strong with the CPEs whose exports are over 20 per cent made up of oil and related products. The increase in world trade associated with higher oil prices will have differential effects depending on the pattern of trade links. Those countries that benefit will be those that trade with the high absorbers such as the CPES, Africa and Latin America.

Tables A2 and A3 also indicate that higher oil prices could produce some strain in the EMS. Initially we would expect France and Italy to appreciate, but the delayed inflationary consequences of the oil price rise are likely to induce higher interest rates and subsequent depreciations. As the model is forward looking these higher interest rates in years two and three are already taken into account when the initial exchange rate is calculated. (These results are essentially the same as those in Barrell 1990).

Both the UK and Canada are oil producers and net energy exporters, and hence the rise in the oil price improves their balance of payments and both currencies appreciate. This offsets the initial inflationary impact of the oil price rise and hence there is much less need for an increase in interest rates in order to maintain the monetary stance. UK output is above base because the UK gains from the increase in world trade without having to suffer a tighter monetary policy. UK prices are only 0.6 per cent higher after 3 years, the exchange rate initially appreciates by .5 per cent." b) Long term effects Table A4 presents some summary statistics covering the effects of a permanent rise of 25 per cent in real oil prices. The simulation is run over 15 years with the same assumptions as for the temporary shock. Output in the major seven settles slightly below base after five years and inflation is a little above base. The change in output results in part from a change in relative prices which reduces the consumption wage and hence reduces the supply of labour. The trade/output ratio, as indicated by the ratio of col 2 to col 7, rises slightly.

The model stabilises with inflation about one quarter of a per cent above base. As a result domestic demand falls relative to output to ensure that more resources flow into net exports. This rise in inflation is a result of our assumptions as well as of the model. The reduction of absorption in the advanced world could take place through one of four channels. Wealth effects are one possible channel. A sequence of current account deficits reduces wealth and hence demand. However the estimated wealth effects on the model are both rather weak and take a long while to have an effect. Fiscal policy could be tightened to induce the transfer or real interest rates could be raised. Both are assumed unchanged in these simulations. The only remaining alternative is to induce the transfer of resources by allowing inflation to be higher. The policy responses that we assume to hold produce this result.

The authorities in each of the countries in the advanced world have a number of policy options open to them, and if the effects of wealth on adjustment are too slow then either tighter monetary policy raising real rates or a tighter fiscal policy than we have assumed (which would reduce wealth more quickly by reducing the stock of government debt) might well be chosen. Indeed a tighter fiscal policy may be necessary for equilibrium because the shock will change the debt income ratio, and adjustments would have to be made to return it to a stable path. (c) A 50 per cent rise in the oil price Our model is such that a $1 0 rise in the oil price has around twice the effects of a $5 rise. Table A5 reports a simulation with oil prices 50 per cent above base levels for 3 years, and thereafter they return to our base path. Output in the major seven would be over half a per cent below base levels after two years, and US output in particular would be almost a full per cent below base by 1991. This however is not enough to push the US into a recession, as we are forecasting over two and a half per cent growth in that year. inflation in the major seven would be one per cent higher in the first year, and US inflation would be around one and a quarter per cent higher. We would expect the US authorities to raise interest rates substantially and also that the effective exchange rate for the Dollar would fall. The assumptions of constant real interest rates and constant fiscal policy are commonly used in large model simulations. Our exchange rate assumptions are more innovative. When we solve the model in forward looking mode we assume that the real exchange rate changes along the uncovered real interest parity path. The exchange rate in the first period has to be such that the real exchange rate stabilises toward the end of the run and that the current account as a per cent of GNP also returns to its base level. We are assuming that all countries including the US have to return to the base current balance to GNP ratio. The current balance trajectory on our base reflects the growth path of the world economy and the pattern of savings and investment in the world. The equilibrium trajectories for these variables, and hence for the current account depend upon demography, preferences for consumption now and on relative rates of return. Hence we would expect the world economy to return to the equilibrium set of current accounts after a shock. However we expect that this process will take some time and we allow up to 10 years after the shock has been removed for adjustment to take place.

GEM THE NATIONAL INSTITUTE'S GLOBAL ECONOMETRIC MODEL

GEM is a 640 equation macroeconomic model covering the whole of the world economy, but focusing particularly on the seven major industrial countries. The National Institute publishes the current version of GEM together with a user-friendly, menu-driven operating package designed to enable economists to produce their own forecasts and simulations.

Background

GEM is used by the National Institute to produce forecasts and analyse events and policy options in the world economy. it was developed from a model maintained by HM Treasury to produce internal world forecasts. In the last two years it has established itself as one of the leading half dozen world econometric models, participating (as the only UK based model) in comparative exercises organised by the US Federal Reserve and the Brookings Institution in Washington. it has been used in academic studies of international policy co-operation, and research and development of it is financed by the Economic and Social Research Council. The Institute has already provided the model to the Bank of England, HM Treasury and the World Bank for use in their published forecasts. Scope and coverage GEM is divided into sixteen sectors. Each of the G7 country sectors contains around 60 variables covering individual components of demand, price indices, exchange rates and interest rates, trade and the current account. Much smaller sectors exist for Belgium, Netherlands and the rest of OECD. The remaining six sectors cover OPEC, Asia, Latin America, Africa, the Centrally Planned Economies and Miscellaneous developing countries. These sectors contain equations for trade volumes and prices, which depend on five commodity price indices.

The model package

The published model package comes with a 'front-end' specifically designed for GEM (by Bahram Pesaran who co-wrote the estimation package MICROFIT) to enable economists with little or no previous experience to produce forecasts and policy analysis. A user's guide is provided, but this menu-driven programme is largely self explanatory. A base forecast is provided each quarter with the latest version of the model, and the user can inspect numerically and graphically the judgements that lie behind this forecast. By changing these judgements, or by adding new data, the users can produce their own forecast. Alternatively by using the model in simulation mode the user can look at the global effects of events such as a US fiscal expansion or an oil price fall.

A comprehensive model manual lists individual equations and describes the theoretical basis of the model's relationships. A full data listing gives the definition and source for each model variable, allowing the user to update the forecast in between quarterly releases.

Hardware requirements

The package requires a 80386 based PC, such as the IBM P/S2 Model 70 or 80, with a minimum of 2 megabytes of memory (although we recommend 4 megabytes). A maths co-processor (80387) is also recommended. The computer should also have at least 20 megabytes of hard disk and a VGA graphics board.

The annual subscription is 94,500 plus VAT with a reduced rate of 2500 for academic users. The Japanese economy has been becoming more open over the last four to five years. This has been partly the result of pressure from overseas, and especially from America. Non price barriers to trade have been removed, and Japanese consumer goods markets have become more open. There has, as a result, been a significant rise in the level of imports into Japan in the last five years. Chart Al plots recent data for both total and non oil import penetration in the Japanese economy along with our forecast over the next three years. There has been a corresponding, but less marked change in Japanese export performance over the same period. A number of Japanese firms have been undertaking considerable programmes of direct investment overseas. They have in particular been setting up plants around the Pacific rim, to substitute for domestic production for both the home market and for export markets. Chart A2 plots the share of Japanese exports in World trade and Japanese direct investment outflows as a per cent of GNP.

The trade equations for Japan on GEM were last re-estimated in 1986-87, and we have suspected that the post sample period for these equations exhibited structural change. Chart A3 plots the equation residuals for both exports and imports of (non-oil) goods for Japan. There appears to be a downward trend in the export equation residual and a step change in the import equation residual. The combination of a belief that the structure has been changing along with the deteriorating performance of our equations has led us to re-estimate them. We have delayed doing this for some time in order to ensure that we had sufficient data for sensible stability testing.

(a) Japanese Imports

We re-estimated our equation for Japanese imports of non oil goods starting with the general form: - [Delta M] = a + bTFE(-l) + cRPM(-1) + dM(-1) + [Lambdal.(L).Delta.M] + [Delta.(L).Delta.TFE + [Phi.(L).Delta.RPM + [e.sub.t] and tested down by sequential elimination until the final parsimonious form was reached. This is reported in the first column of table B1. The equation is statistically acceptable. However, our prior is that there has been a change in the structure around 1986. Column two in the table reports the same equation estimated up to 1985Q4. There are two stability tests reported. The first is Chows predictive test, and the regression comprehensively fails. The second Chow test for internal parameter stability is much weaker in this situation, but the equation also fails it. Both the income and the competitiveness elasticities are lower in the equation estimated up to 1985Q4 then they are over the whole sample period.

Column three of table Bl reports the final equation we have chosen. We investigated a number of possibilities for structural change in the long run coefficients, and found that there was a case for allowing the impact elasticity of TFE to change as well as for allowing the long run elasticities on TFE and competitiveness to increase. In order to model these changes we included the lagged levels of TFE and non oil imports for the period 1986Ql to 199001. These dummied variables have slowed the effects of TFE and competitiveness as well as raising the long run effects. b) Japanese exports We also re-estimated our equation for Japanese exports of goods starting with the general form:AX = a + XTIME + b(X(- 1) - S(- 1)) + CRULT(- 1) + D(L)AX + E(L)AS + f L)ARULT + et and tested down in the usual way. Table B2 reports our results. The equation estimated over the whole time period has a competitiveness elasticity of 1.00, and market share grows by 2 per cent la year at constant competitiveness. The same equation estimated up to 1985Q4 in column 2 of table B2 has a lower competitiveness elasticity and a higher rate of gain in world trade share. Although it passes a predictive failure test it fails the usual Chow internal parameter stability test. After some experimentation with split time trends around 1986 we decided that the best description of the data generation process involved using a time trend for world trade share gain up to 1985Q4 and thereafter to have an equation that would display a constant world trade share at constant competitiveness. The resulting equation is reported in column 3 of table B2. The competitiveness elasticity is just below one, and before 1986 world trade share would have grown by 2.25 per cent at constant competitiveness.

(c) Consequences for the Japanese balance of payments

The new equations for Japanese trade change model properties. Although we had taken account of these perceived changes in structure by adjusting residuals in our forecasts, this adjustment would not have fed through to simulation properties. They now do so. in particular, at constant competitiveness we would now expect Japanese exports to grow around two per cent a year more slowly than they have in the past, and we would expect Japanese import growth to be around one per cent higher than it would otherwise have been. These changes are likely to reduce the Japanese surplus by a cumulating $9 billion per annum. They are a major factor behind our forecast of a fall of 1 per cent of GNP in the Japanese current account surplus over the next decade.