Printer Friendly

The home economy.

CHAPTER 1. THE HOME ECONOMY The forecasts were prepared by Andrew Britton, Paul Gregg and Michael Joyce, but they draw on the work of the whole team engaged in macroeconomic analysis and modelbuilding at the Institute. On this occasion we have adopted a rather different format for this chapter from the customary one. Part One begins with an analysis of some of the most important developments of the past few years, with notes on the deterioration in the balance of payments, on the fall in the savings ratio and on the acceleration of inflation. Next we discuss some of the problems associated with economic forecasting. We analyse the errors made last year and compare them with the error margins normally associated with short-term forecasts of this kind. We look at the behaviour of the economy at the corresponding stage of previous economic cycles. And we consider the best way of forecasting GDP when there are discrepancies between the measures of its growth in the past. Our central forecasts for 1989 and 1990 are described briefly in the text of Part Two, and more fully set out in the usual tables. We end in Part Three with a discussion of alternative scenarios for the medium term, with particular reference to their implications for interest rates and the exchange rate. An appendix describes the regional pattern of unemployment and the way it has changed since the early 1980s.

PART ONE Background to the forecast

The deterioration of the balance of payments In 1985 the current account of the balance of payments was in small surplus, but last year there was a deficit of 14 billion pounds. The turnround in three years amounted to 4 per cent of GDP, or 21 per cent of visible exports. It has been suggested by the Treasury that the size of the deficit may be exaggerated, but we do not believe that any convincing argument has yet been put forward for changing the recorded figures on the basis of present information. (If there are revisions they are more likely to affect the level of the balance than the change from year to year.) We therefore take the data as broadly correct and seek to understand their significance. Can the turnround since 1985 be explained simply by the cyclical expansion of demand in the UK, or is there a reason for concern about the underlying trend as well?

The invisible balance has not changed much over these three years, and its measurement is in any case very uncertain. The deterioration is all in the visible balance, which is probably quite well recorded. The oil surplus has been reduced from 8.1 billion pounds to only 2.3 billion pounds in 1988, because of the fall in oil prices and the fall in North Sea production. But the most significant deterioration has been in the balance of trade in manufactured goods, where the deficit has risen from 3.2 billion pounds to 14.3 billion pounds over the same three years.

The volume of manufactured exports rose in those three years by nearly 18 per cent, probably about the same rate of increase as in total world trade. There was a significant improvement of relative price competitiveness in 1986, but since then the gain has been more than offset by subsequent exchange-rate increases. On balance competitiveness should have made little difference to the three year comparison. Some of the increase in exports last year in particular is unexplained by our equation and, to that extent, performance might be characterised as unusually good, subject to the uncertainty remaining over the measurement of most of the explanatory variables.

The rise in the volume of manufactured imports was far higher, 37 per cent over the three years, with much of the increase coming in 1988. The import equation on our model fits this period rather well. It predicted a rapid rise in import volume last year because of the acceleration in the growth of demand and because output was rising faster than capacity. Thus the deterioration in the balance of trade in manufactures can, it seems, be adequately explained by our model. The implication for the forecast is not altogether encouraging. It means that there is no obvious case for making any special adjustments to the model's predictions when the pressure of demand in the UK falls away.

But does the model itself, based as it is on experience over two decades, imply a deteriorating trend in the trade balance even in periods when demand in the UK is not excessive? If world trade volume grows at 5 per cent a year and competitiveness is unchanged, our equation predicts that UK manufactured exports will rise by 3.4 per cent a year. We have nevertheless taken a more optimistic view, consistent with experience over the last few years, that UK exports now rise broadly in line with world trade, other things being equal. On the imports side, our model implies that the volume of manufactured imports will rise in line with domestic demand plus a term which measures the trend towards greater international specialisation in production. With constant competitiveness and output growing in line with capacity in the UK, therefore, import growth will be about 2 per cent above the growth of demand.

The implication of these calculations is that the total demand for manufactures in the UK (including exports) can only rise at best by about 3 per cent a year, if imports and exports of manufactures are to keep a constant ratio. (The implication for the growth of GDP is more difficult to compute, and it must depend on the composition of output as well as its overall level.) The norm is now for the total demand for manufactures to grow faster than 3 per cent a year. In that case the balance of trade in manufactures will tend to deteriorate, or the level of price competitiveness will have to improve. Thus, the performance of the last three years, although it was due in large part to the boom in domestic demand, should be seen as part of a continuing trend towards larger balance of trade deficits in the UK. We explore the implications for the medium-term forecast below.

The fall in the savings ratio Consumer spending rose in volume last year by about 6 per cent. This follows two previous years in which the increase was almost as fast, 5.8 per cent in 1986 and 5.6 per cent in 1987. The growth of real personal disposable income over those three years has been moderately fast, about 3 per cent a year, and the savings ratio has fallen from 9.5 per cent in 1985 to only about 2 per cent last year. This makes a sharp break from the previous behaviour of consumers and cannot be explained satisfactorily by the models estimated during the 1970s and early 1980s. Somewhat similar developments have occurred in other countries, including the United States and in the Scandinavian countries, where savings ratios are now actually negative.

The main reason for the change in behaviour in the UK appears to be the liberalisation of financial markets, especially the provision of consumer credit and mortgage lending. In the consumption function in our model this works in two distinct ways. The growth of consumer credit adds directly to the resources available to households who would otherwise have to delay purchases, especially durable goods. This effect accounts for much of the fall in the savings ratio. Over the last three years the stock of consumer credit outstanding has risen in real terms at an average annual rate of about 12.5 per cent. The ratio of consumer debt to personal disposable income, which was 10 per cent at the start of 1985 is now about 14 per cent. The main reason seems to be an increased willingness of banks and other financial institutions to lend for purposes of consumption; indeed there has been intense competition amongst them to do so.

The second route by which liberalisation of financial institutions has increased consumption is by making the personal sector more liquid. The identified liquid assets of the sector have risen over the past three years on average by about 7 per cent a year in real terms. Even though the sector has over those three years been in financial deficit, its stock of bank deposits and other readily `cashable' assets has been growing fast. The main reason why this has happened has been the growth of mortgage lending, far exceeding that needed to finance new house purchases.

We do not believe that the rise in house prices, on its own, was an important reason for the fall in the savings ratio. It added to the spending power of those who inherited homes they did not want to live in, or those `trading down' from larger to smaller properties, but it also reduced the spending power of first time buyers, who did not have access to 100 per cent mortgages. The recent levelling-off of house prices will not necessarily lead to a marked change in savings behaviour.

This rapid growth of household borrowing, and the relaxing of the criteria used by banks and building societies to assess the creditworthiness of their customers, have inevitably increased the number of bad debts and repossessions, but there is no sign that this is deterring either borrowers or lenders. The `natural' limits to credit expansion have not been reached. Higher interest rates will slow down the growth of consumer credit, indeed there are signs of this happening already. But the shift to a higher household debt-to-income ratio is probably irreversible, and the trend might even go further if interest rates are not kept high enough to prevent it.

The acceleration in inflation The rise in the retail price index exaggerates the present level of inflation, and especially its acceleration in the course of last year. The increase in the wholesale price index year-on-year was 4.4 per cent in 1987 and 4.8 per cent in 1988; the GDP deflator rose by about 5 per cent in 1987 and probably about 6 1/2 per cent in 1988. These figures should give a better estimate of underlying inflation in the UK.

The index of total costs used in our model is constructed by adding together the rise in unit wage costs and the rise in import prices. This shows an increase of 3 per cent in 1987 and 3.5 per cent in 1988, suggesting that profit margins increased in both years. Our equation for wholesale prices takes account of effects both from the growth of demand and from the level of capacity utilisation, but under-predicts slightly the rise in prices over the last two years.

The rise in costs last year understates the danger of accelerating inflation for two reasons. Import prices actually fell slightly year-on-year, thanks to the fall in world oil prices and the appreciation of the exchange rate. Moreover, the rise in earnings was partly offset by an increase in productivity well above the sustainable trend. There are good grounds therefore for concern about inflationary pressures in the economy even though the actual acceleration in price rises so far has been modest. Both earnings and consumer prices have been rising rather faster than a normal relationship with costs and the pressure of demand would suggest. If this behaviour were maintained it would amount to an increase in the sustainable level of unemployment, but it would perhaps be premature to draw that conclusion from the evidence we have to date. It remains to be seen what effect a tighter monetary policy and slower growth in economic activity will have on wage settlements, on productivity and on profit margins.

Forecast error margins Last year was not a good one for macroeconomic forecasters. The mistakes made by the official forecasters at the Treasury have received particular publicity, but they were shared by most independent forecasters, ourselves included. It may be useful to set those errors in the context of measures of forecast accuracy calculated over a much longer period.

The last assessment of our output forecasts was published in the National Institute Economic Review, August 1983. Absolute errors were calculated for Institute forecasts published in the February edition of the Review each year from 1959 to 1982. The growth rate of GDP through the year was assessed, that is the fourth quarter on fourth quarter growth rate for the year of publication. Compared with the first published estimate of the outturn the average absolute error was 1.8 per cent; compared with the latest estimates available at the time the exercise was carried out the average absolute error was 1.4 per cent. A similar calculation for the years 1983 to 1987 gives corresponding figures of 1.4 and 2.0 per cent.

Our own estimates now of the level of GDP in the fourth quarter of last year suggest that our forecast in last February's Review was too low by about 2 per cent, which would be quite a large error but not outside the expected range.

A similar review of inflation forecasts was published in the Review in February 1984. The absolute error of consumer price index forecasts was calculated using fourth quarter growth rates in February Reviews for the period 1964 to 1982. Using first published estimates of the outturn the average absolute error was 1.7 per cent; using the latest data available at the time it was 2.5 per cent. For the period 1983-7 the average absolute errors were a little smaller than that; 1.6 per cent using first published outturns or 1.5 per cent using the latest data available now.

Our forecast of consumer price inflation in the February Review last year was 5.0 per cent. No outturn is yet available, but information from other price indices suggests that our forecast was less than 1 per cent out. In that respect 1988 was in fact a relatively good year for forecasters.

The record of balance of payments forecasts has never been analysed in the same way. We have calculated, however, the average absolute error made each year from 1970-87 in the February Review when forecasting the current account for that year. The figures were rescaled in proportion to the value of GDP at current prices and averaged to produce an absolute error of 2.3 billion pounds at 1985 values. The largest absolute errors (rescaled in this way) were 7.0 billion pounds in 1979, 6.3 billion pounds in 1980 and 8 billion pounds in 1972. The error for 1988 will probably be in excess of 8 billion pounds at 1985 values, making it (by a small margin) the largest absolute error over the period.

These notes provide some indication of the margin of error to be placed around our forecasts for 1989. Our forecast of the growth rate through the year, fourth quarter to fourth quarter, is given as 1.5 per cent. A realistic interpretation of that would be that the growth rate will probably be in the range 0-3 per cent. Similarly for inflation we would wish to describe our forecasts for the rise in the consumer price index, not as the point estimate of 5.4 per cent but as a range from about 4 per cent to about 7 per cent.

It is much more difficult to put a range round the forecast for the current account of the balance of payments. The Treasury last year quoted an average error of 3 billion pounds for their forecasts (which are published later in the year). That would be in line with our own forecast record, but takes no account of the unusually large errors we both made last year. Our central forecast is a deficit of 14 billion pounds; a reasonable interpretation of that might be a range from 10 billion pounds to 20 billion pounds, with the added warning that the outturn could be very different if there are substantial data revisions to invisibles during the year.

Is the economy at a cyclical turning point? Since the war the British economy has followed a relatively well-defined business or trade cycle with an unusually regular time period peak-to-peak of 4-5 years. We have come to expect good years of rapid growth, such as we have seen in 1987 and 1988 to be followed by years of especially poor growth or even recession. Do current conditions point to a sharp cyclical downturn in the next year or two? Before turning to our own short-term forecasts we look briefly at a much simpler approach based only on the time series properties of output itself.

Counting the years between cyclical peaks is unlikely to be a reliable method of forecasting unless the cycle is driven by some external force which follows a strict chronology (such as the four-year gap between elections in the US). A less suspect method would be to use an autoregressive equation relating activity in one year to activity in the two preceding years. (An equation of this sort will reflect the periodicity of the output cycle over its estimation period.) For example such an equation estimated for industrial production over the period 1958 to 1982 is as follows:(1) y = 2.4 - 0.0031t + [0.195y.sub.-1] - [0.393y.sub.-2] + 0.692w where y is (the log of) industrial production, t is a time trend and w is (the log of) world industrial output.

Because of the large, negative, coefficient on output lagged two years this equation applied to current data, would predict a sharp slowdown in 1989 (reflecting the acceleration in growth in 1987) and stagnation or even recession in 1990 (reflecting the further acceleration of growth in 1988). However, further investigation of the mechanisms generating the cycle in the past suggests that such a forecast might be unduly pessimistic.

The behaviour of the output cycle through the 1960s and 1970s is better explained by reference to additional variables including inflation, interest rates and the exchange rate.(2) Thus the sharp downturn in 1974 probably resulted from the sharp rise in interest rates in the preceding years 1972 and 1973, as well as the downturn in world activity and the acceleration in inflation. A sharp increase in inflation and a marked loss of international competitiveness probably accounts for much of the severity of the downturn in 1980. These conditions are unlikely to recur in the forecast period.

Although interest rates have risen sharply since the summer of last year, the significance of that increase should not be exaggerated. It follows a period when the trend in interest rates was generally downwards. This contrasts with the larger and more sustained rise in interest rates between early 1972 and the end of 1973, which preceded the 1974 recession. Moreover, the increase in inflation, and the loss of competitiveness during 1988 are both very small compared with the changes that occurred prior to the recession of 1980-1. The current situation is in some ways more like that of 1985 when interest rates had risen quite sharply, and so had the rate of inflation, but the growth of output was nevertheless maintained. The behavioural relationships which seem to lie behind the relatively regular behaviour of the post-war trade cycle in Britain would now suggest a slowing down in the growth rate of output rather than a recession. This is consistent with the forecasts described below.

The discrepancies in the measurement of GDP The three measures of gross domestic product, based on output, expenditure and income data, should in principle be equal, but are often in fact very different, especially when the first estimates are published. Comparing the first three quarters of 1988 and 1987 gives the following percentage growth rates for the three measures: output 5.0, expenditure 1.8, income 3.8. These are unusually large discrepancies and have provoked a review of some of the procedures by which the data are collected and processed.

On this occasion there may be some question whether the estimate of the output of manufacturing (which is partly based on gross output indices) takes full account of the unusually rapid rise in imports in the latter part of last year. Nevertheless we have persisted in our view that the output estimate provides the best guide we have. This leaves us with the problem of deciding how to treat the error in the expenditure estimate when we make our forecast of growth next year. Since we build up our forecasts of GDP by projecting the categories of expenditure this is no mean problem. Indeed, given the size of the discrepancies in the latest data, the size of the effect on the forecast of this one judgement dwarfs most of the other uncertainties.

If the discrepancy in the past data were allocated to different expenditure categories, the projection would not in general be the same. The implication for imports would be different, and so would the effects on domestic incomes and hence on consumer spending, stockbuilding and so on. But these `second round' considerations are generally small compared with the uncertainty of the direct expenditure effects themselves. These depend on the dynamic structure of the equations for the expenditure categories chosen and on the treatment of the residuals in them. This can be illustrated with two examples, fixed investment by manufacturing industry and stockbuilding by manufacturing industry. The equation for investment includes the following relevant terms: [Delta] InQDKMFA = -0.38 [Delta] In [QDKMFA.sub.-1]

-0.20 In [QDKMFA.sub.-1] which can be expressed as: y = [0.42y.sub.-1] + [0.38y.sub.-2] + u where y is the log of the level of spending. Suppose that ex-post, the level of y in the past year is raised by 1.0 units per quarter. The calculated values after the first quarter of last year will also be raised, because of the influence of lagged values of investment. Hence the effect on the level of the equation residual in the past year is less than 1.0 units per quarter. If an `automatic' forecasting procedure is followed for this variable, the equation residual will be raised by 0.5 in all subsequent quarters. The calculated values for the future will also rise and this effect on the calculated forecast values is additional to the effect on the projected residual. Thus the forecast value of investment will be raised by more than 1.0 if that addition is made to investment in the past. The average for the first four quarters is about 1.5. A similar calculation could be made using the equation for the level of stocks held by manufactures. The net effect is to raise stockbuilding in the forecast period, by 0.8 in the first quarter, 0.7 in the second, gradually falling towards zero. The average for the first four forecast quarters is a little over 0.5.

The discrepancy between the output and expenditure estimates in the first three quarters of 1988 averages a little over 3 per cent of GDP, widening through the year. If we added 3 per cent of GDP to investment in 1988 the calculations above suggest we would raise GDP in 1989 by about 4.5 per cent. If we made the additions to stockbuilding instead, we would raise GDP in 1989 by only about 1.5 per cent. The procedure we have actually followed is to hold the residual roughly constant (starting from a level rather higher than the average of the first three quarters of 1988). This adds about 3.5 per cent to GDP in 1989. It implies that this year, and subsequently, the growth rates of the output and expenditure measures of GDP are very similar.

PART TWO The short-term forecast

Policy assumptions In the Medium-Term Financial Strategy presented at Budget time last year the assumed `fiscal adjustments' (which are assumptions about the scope for future tax cuts) were 3 billion pounds in 1989-90 and 1 billion pounds a year subsequently. This was then thought to be consistent with a balanced budget, or zero PSBR. In the event the public sector will be repaying debt to the extent of about 2.5 to 3 per cent of GDP in the current financial year, with the prospect if tax rates are unchanged of even larger surpluses in 1989-90 and beyond.

Nevertheless the deficit on the balance of payments means that tax cuts on a large scale are unlikely. The Chancellor could argue, and probably will argue in the Budget Speech this year, that it is appropriate for the public sector to run a surplus if the private sector is in deficit, and that his previously stated aim of balancing the Budget is meant to be no more than a rough guide to the position on average over a long run of years. We assume that income taxes will be cut by 3 billion pounds this year and by 2 billion pounds a year subsequently, which is only marginally different from the figures in the FSBR last year.

For monetary policy the choice of assumptions is more difficult. It is now well established that priority in the setting of interest rates is being given to reducing the rate of inflation. One might suppose therefore that the immediate aim would be to keep the exchange rate strong and to prevent any reversal of its recent appreciation. However for reasons which are spelt out in Part Three, we do not think that interest-rate movements on their own can be used to hold the exchange rate at its current level. The present level of UK interest rates relative to those abroad implies that the markets expect a sterling depreciation of about 4 per cent a year. We take, as our central assumption, the case where the present level of interest rates, a 13 per cent base rate, is maintained throughout this year and next. Variants involving lower interest rates are considered in Part Three.

Output and expenditure Our forecast for GDP growth this year is very similar to that in the last (November) Review, about 2.5 per cent. This is consistent with growth through this year, fourth quarter on fourth quarter of only about 1.5 per cent. As explained in Part One the treatment of the discrepancy between the output and expenditure measures is the key judgement on this occasion.

Year-on-year consumer spending is a leading sector of domestic demand, but through the year its growth is much slower. Investment growth also slows down in response to the rise in interest rates. Export volume accelerates (partly because oil production recovers) but there is a small negative contribution to output from net overseas demand.

We include 1990 for the first time in our short-term forecast on this occasion. At this stage, it seems likely to be a year of very slow growth, although not quite a recession in the strict (or old-fashioned) sense of the word. Domestic demand may not grow at all, since both consumption and investment will still be very depressed, but there should be some continuing impetus from exports. Obviously it would be possible for the growth of the economy to recover in 1990 if interest rates were brought down or the Budget in that year were more expansionary. At the present time however that would seem incompatible with a significant further reduction in inflation and with reasonable progress back towards balance of payments equilibrium.

Consumer spending There is some indication from the figures for retail sales at the end of the year that the increase in spending was already moderating. Nevertheless we begin the forecast period with a `positive residual' on our consumption equation, that is an outturn over the model's prediction. We treat the excess as transitory, a judgement which contributes significantly to the overall profile of expenditure and output in our forecasts this time. We also expect the growth of consumer credit to slow down quite abruptly. Even so, the level of the savings ratio remains very low, just 1 per cent on average this year, rising to 1.5-2 per cent in 1990. As explained in Part One, the factors which have led to progressive falls in that ratio during the 1980s seem unlikely to be quickly reversed.

Fixed investment and stockbuilding We start with the problem of estimating investment and stockbuilding in the latter half of last year. We have not attempted to allocate the discrepancy in the national accounts to the various categories of investment or stockbuilding where it may well belong. The equations in our model do not suggest that the recorded figures for any of these categories of expenditure are particularly implausible.

If the rise in interest rates is to succeed in slowing down the growth of domestic demand and of imports, it must do so in large part by cutting back on industrial investment. Given the time lags involved in investment spending decisions, the effect is likely to be more pronounced in 1990 than this year. Thus our forecast for manufacturing investment shows a rise of about 10 per cent this year, followed by a fall of nearly that magnitude next year. A smaller rise followed by a smaller fall is likely for spending by distribution and services. The downturn in private housing investment, already signalled by the behaviour of house prices may come a little earlier.

In the early stages of a downturn in economic activity involuntary stockbuilding may moderate the effect of demand on current production. Nevertheless we would expect industry to be reluctant to hold more stocks than necessary given the high cost of borrowing.

Exports and imports at 1985 prices We are taking a relatively optimistic view of manufactured exports. Through the current year, fourth quarter to fourth quarter, the projected increase is over 5 per cent. This growth rate is above that of world trade, despite the loss of competitiveness amounting to more than 4 per cent over the last twelve months. On imports of manufactures we are also taking a relatively sanguine view, in that we are not projecting into the first half of this year the small positive residual in our equation at the end of 1988. The implication is that imports of manufactures are falling back slightly in the current quarter.

The oil balance can be expected to recover a little during this year as production losses following the Piper Alpha disaster are made good. Trade in other non-manufactured goods could also help to improve the overall balance of payments. We expect, for example, to see an early rebound from the low level of exports of food and basic materials at the end of last year.

Employment and unemployment The path of output described above should be consistent with roughly constant employment. Productivity growth, which has a trend of over 2 per cent a year, will slow down in response to the cycle. The fall in manufacturing employment will become quite rapid once more, but we still expect some increase in employment elsewhere in the private sector, notably in services.

Despite this we expect unemployment to go on falling for the rest of this year. The count has already been brought down to below the 2 million mark by another substantial fall in the last quarter of 1988, and yet another change in definition. On the latest definition we expect the count to be about 1.8 million by the end of this year, at which point it may level off in the absence of further changes in benefit administration, or in the definition of the count.

Wages and prices The forecast period begins with wage settlements averaging around 8.5 to 9 per cent, considerably more than our equation would predict, or the prospects for the economy would seem to make sustainable. There is little doubt that the rate of increase of earnings will slow down substantially, but it is difficult to say how soon this will happen. Our judgement is that the rise through this year, fourth quarter of fourth quarter, must now amount to about 8.5 per cent, and that during 1990 the rate of increase will slow down a little further. Left to itself (with zero residuals) our equation would predict a more abrupt slowdown, because of the turndown in productivity growth and because the fall in unemployment is coming to an end.

Despite the relatively high rate of growth in unit labour costs, and the rise in import prices due to exchange-rate depreciation, our forecast of inflation at the end of this year is below the outturn for the end of 1988. In the case of the RPI this comparison is helped by the profile of interest rates, but a similar point can be made about the GDP deflator, and probably the CPI as well. Profit margins, whose contribution to price increases during the boom has been described in Part One above, should narrow again this year and next, as demand and output stagnate.

The balance of payments The relative growth of import and export volumes in our forecast is consistent with a slight improvement in the non-oil visible balance this year and next. The prices of non-oil exports should rise by 5 per cent this year, that of non-oil imports by about 4 per cent. In 1990, thanks to the depreciation of sterling the terms of trade may deteriorate a little.

The oil balance was only just in surplus in the latter part of last year, but should improve when oil production is higher and when prices recover too in line with the world oil market.

The surplus on invisibles is likely to remain broadly constant this year and next. The service balance this year will be adversely affected by the rise in the pound last year. The most difficult component of all to forecast, is the balance of net property income. The accumulation of large deficits on the current account should mean that the stock of net assets held by UK residents is gradually run down, implying larger outflows or smaller inflows of interest, profits and dividends.

Overall the implication is that the current account deficit will not change much year-on-year either in 1989 or 1990, although the very large deficits of the latter half of last year may not be repeated. Inevitably the figures will vary greatly from month to month, but some movement back towards balance should be perceptible over the forecast period. In the medium term, as explained below, we assume that an improvement in the balance of payments by one means or another is inevitable.

Public sector finance The large surplus on the public sector accounts will continue for the next two years, unless our fiscal policy assumptions are radically mistaken. The growth of public spending will remain relatively slow. Current expenditure on goods and services is forecast to rise by about 1 per cent a year in volume, or about 6 per cent at current prices. Spending on current grants is projected at 4 per cent next financial year, perhaps rather faster thereafter as unemployment turns up again. Debt interest payments are already falling.

On the revenue side the increases will again be rapid this year, although subsequently they will slow down as the economy itself slows down. Profits are taxed with a lag of at least a year, whilst personal income tax payments and national insurance contributions are both related to wages, which will remain buoyant for some time yet. In the medium term revenue will be curtailed by the slow growth of incomes and, more especially, of consumer spending. Nevertheless we foresee large surpluses for the next two years, and probably even beyond that horizon.

PART THREE The medium term Since 1980 the Treasury has presented each year, at Budget time, a statement of the Government's medium-term financial strategy. The exact coverage of the statement and its accompanying tables has varied from year to year, and now includes annual projections not only for `money GDP', but also for real GDP and inflation up to the year 1991--2. Figures were also given for government spending, tax receipts and for the PSBR, as well as `illustrative ranges' for the growth of the monetary base MO. The projections we present in the remainder of this chapter should be understood as an exercise not unlike that underlying the MTFS, although a fuller range of information is provided. They are not forecasts in the normal sense of the word, but projections based on a consistent view of the world that is embodied in our macroeconomic model.

Analysis of the medium term emphasises some different aspects of economic behaviour from those which dominate short-term forecasting. We do not attempt to produce a cyclical path for economic activity, but concentrate rather on the determinants of trend growth. No great significance should be attached to projections of inflation, unemployment or the balance of payments in particular years. The aim rather is to see how these variables might evolve in general terms during the early 1990s.

The tables reported here run from 1988 to 1997 and in every case the current account of the balance of payments is close to zero in the final year. It is also assumed in every case that the rate of depreciation of the exchange rate is approximately equal to the uncovered differential between UK and overseas interest rates. In combination these conditions ensure that the model treats the current exchange rate at the beginning of 1989 as a market clearing rate. If these conditions did not hold, the model would forecast an immediate `jump' up or down in the rate, implying that the expectations now held by the market are incorrect. We have chosen instead to concentrate on alternative scenarios all of which are consistent with the view that current exchange market expectations are `rational'.(3)

The base run will be described first, then a variant on the assumptions about monetary and fiscal policy and finally a variant on investment which improves the `supply-side' of the economy.

The base case Perhaps the most striking feature of the base case is the slow growth of GDP. Over the nine years 1989--97 the growth rate is 1.7 per cent, compared with an average 3.9 per cent over the five years 1984--8. The best way of accounting for this change of trend is in terms of the required improvement in the current account of the balance of payments. The rapid growth of output in the mid-1980s has been accompanied by a sharp deterioration in the current balance; from now on that balance must gradually improve. This change of trend must be brought about either by spontaneous changes in private sector spending or by tighter fiscal and monetary policies. The calculation of the permitted growth rate might, adopting a term familiar to American economists, be called `some unpleasant Keynesian arithmetic'.

The proposition underlying all three medium-term projections is that the current account of the balance of payments must return to zero, or some relatively small magnitude, at least by the latter half of the 1990s. So long as the account is in deficit UK residents in aggregate must be running down their financial assets or borrowing from abroad. In a growing, or an inflationary, world economy some net flows of borrowing and lending between countries may persist in a state of equilibrium, but typically they will be small relative to the incomes of the countries concerned. At the present time the UK is believed to have a positive net asset position relative to the rest of the world. Other things being equal, one might expect that net stock of assets to grow in line with national income, although this need not be true of every year, or even of every decade. Changes in relative asset prices complicate the argument, but it suggests that the long-term equilibrium for the UK might involve a small surplus on the current account rather than the present large deficit. A similar conclusion could be drawn from the argument that the liberalisation of exchange control is a relatively recent change, to which UK financial institutions have even now not fully adjusted. Moreover, pension funds in particular may increase the proportion of their assets held abroad because of a shortage of UK gilts in the market as the national debt is reduced.

The time scale over which it is necessary to restore equilibrium in the balance of payments is a matter of conjecture. Until recently it would have been generally assumed that deficits of around 3 per cent of GDP could not be financed for more than a year at best. Now that capital flows are liberalised and more abundant, the situation is very different. American experience certainly points to the possibility of financing a large deficit for many years--although not painlessly.

It is rather doubtful whether existing channels of finance would transfer funds between the overseas sector and the UK private sector on the scale implied by our projections. The fact that the UK public sector is in huge surplus makes the situation doubly strange. There is a problem here, but we have proceeded on the basis that it is a problem that can be solved by innovation in financial markets.

It was suggested in Part One of this chapter that there is a tendency for the balance of trade in manufactures to deteriorate if price competitiveness is constant. Over the forecast period it is necessary not only to halt that trend but actually to put it into reverse. We are constructing all our projections on the assumption that the current account of the balance of payments is close to zero in 1997. In the base case it is cut ot about 1 per cent of GDP by 1993 compared with over 3 per cent in 1988, a rather slow but sustained improvement over the five years. Moreover, the oil surplus will be virtually eliminated within five years even if we assume oil production in line with the top of the Brown Book range. Thus the `permitted' growth of total demand for manufactures in the UK is about 2 per cent a year, rather than the 3 per cent trend calculated in Part One from the trade equations of our model.

The permitted growth of the economy would be faster if we assumed a large and sustained improvement in competitiveness. The assumption that the exchange rate depreciates at the rate implied by uncovered interest arbitrage in fact seems consistent with only a slight improvement in relative export price competitiveness, averaging about 1 per cent a year for the five year period to 1993.

The implications of slow growth for the achievement of policy objectives are also spelt out in the projection. Inflation falls to about 3 per cent a year by 1992, which is similar to the medium-term world rate of inflation projected in chapter 2, but rises again later on. Unemployment after turning up next year as described in Part Two, rises slowly to about 2.4 million in 1993, and over 3 million in 1997.

Capacity utilisation in manufacturing which falls sharply in the period covered by the short-term forecast, might actually rise in the early 1990s in the base case projection. Just as manufacturing output begins to rise more in line with its trend in the mid-1990s, the growth of capacity is likely to slow down. The increase in capacity this year is partly due to extra investment now coming on stream, but also it seems to a reduction in scrapping. Our model associates this with changes in relative prices. Maintaining the level of capacity nowadays requires a higher ratio of investment to output than it did in say the 1970s. This is a necessary consequence of the trend towards higher capital to output ratios resulting from technical progress and the rise in real labour costs. Thus the levels of investment projected into the future once the current investment boom is over seem inadequate to keep capacity growing even in line with a very slow rate of output growth.

The implications of this calculation, if correct, are clearly important. The growth of output in the economy could be constrained in the medium term not only by the requirement to restore external balance, but also by a shortage of physical capacity. A higher growth scenario on our model requires not only an easing of the external constraint, but also a better rate of industrial investment. The obvious way to bring that about would be to reduce the level of interest rates.

Another feature of the projection which calls for comment is the pattern of sector surpluses and deficits. Obviously this can only be a matter of conjecture, but even in five years time the savings ratio could still be so low that the personal sector is in net financial deficit. The discussion of financial deregulation in Part One implies that consumer borrowing will remain high by any but the most recent standards. It is true that as long as the sector remains in deficit, it will be drawing on its stock of net financial wealth to finance its spending. But this process can continue almost indefinitely if the value of the sector's outstanding assets is rising. Our model takes all these effects into account and comes up with the answer that the savings ratio in 1993 will be about 3 1/2 per cent, only a little higher than it is now. The likely consequence is that the public sector will still be in financial surplus, and the national debt by then substantially reduced--but clearly such a conclusion must be subject to every kind of uncertainty.

Fixing the exchange rate by changing the policy mix In the first of our variant cases we keep all the judgements about the behaviour of the economy involved in constructing the base case unchanged, but assume a different policy mix. As explained above, we do this in such a way as to leave today's exchange rate unchanged.

In our main forecast the exchange rate depreciates in line with the differential between UK and foreign interest rates, implying that the yield from holding assets in sterling and foreign currency is much the same. An alternative view would be that monetary policy will be conducted over the forecast period in such a way as to keep the exchange rate broadly constant at its present level. The result of such a policy would be considerably lower inflation over the forecast period, and it could for that reason be thought a better representation of policy as it has recently been enunciated in ministerial speeches and statements. The alternative is also of interest to those who advocate full membership of the EMS. What path for interest rates is necessary to keep the exchange rate constant at its present level?

The answer to that question is rather a surprising one. Any constant exchange rate must imply a much lower interest rate than the present level. Our interest rate must be similar to that of other countries, since only then will the yields on UK assets be similar to those abroad. But the catch is that on the announcement of lower interest rates the exchange rate would immediately jump down to a level which is compatible with long-run equilibrium on the current balance. In other words a constant exchange rate brought about by cutting interest rates would have to be lower than the present exchange rate. That would mean an extra burst of inflation over the next few years as our price level adjusted to the higher sterling prices of imports.

If the aim is stability at the present exchange rate, then some other instrument must be used as well as the interest rate. If fiscal policy were kept sufficiently tight then current account equilibrium could in principle be restored in the long run, however high the exchange rate may be. Thus the present exchange-rate level could be held by means of a policy switch, relaxing monetary policy by bringing nominal interest rates down, whilst tightening fiscal policy with tax increases or spending cuts.

In the context of our model that seems to be the only way a fixed exchange-rate path could be maintained. There might seem to be an alternative, which would consist of raising interest rates so as to prevent the depreciation expected in the base forecast. Arguably that is the way monetary policy has been conducted over the last year, but it is unsustainable. It involves repeated surprises, that is, increases in interest rates which the market never learns to anticipate: the interest rate must rise higher and higher in each successive period, implying more and more depreciation at some time in the remote future. There must be a limit to the credibility of the market. At some point the pressure must be released by a `crash' in the exchange rate, followed by a drop back in interest rates to a `normal' level again. (Alternatively both interest rates and exchange rates might oscillate violently as both overshoot above and below their equilibrium level).

The behaviour of our model in these simulations turns on a computation of the prospects for the current account of the balance of payments a long way into the future. But not modelbuilder can be confident of representing correctly behaviour such a long period ahead. The effect of exchange-rate appreciation now on the current account in many years' time is obviously very uncertain. It is possible to argue that there will be no effect at all, because the current balance will have been brought back into equilibrium by some other market mechanism irrespective of the path of the exchange rate. In our model this is not true over a ten year horizon, but we cannot rule out the possibility that it is true in the real world over the horizon relevant to market expectations.

In our policy variant we bring interest rates down straight away by 3 percentage points so that they are in line with world rates, and therefore consistent with a fixed exchange rate. To prevent the exchange-rate jumping in the initial period we offset this change to interest rates by assuming a tighter fiscal policy. Specifically we assume that the tax cuts in the base case are all reduced: in this year's Budget the tax cuts are only 1.0 billion pounds, followed by 0.75 billion pounds each year subsequently, up to 1995.

Initially the effect of this combination of policies is to raise output, but after 1992 the level of output is lower than in the base run, and the level of unemployment is higher, because the effect of the tighter fiscal policy is stronger than the effect of the easier monetary policy.

The effect on inflation, however, is more striking. The exchange rate is now held constant at its present level, removing the main cause of inflation in the base case. With the exchange rate remaining strong the aim of price stability could, it seems, be achieved in the early 1990s. This is the implication of our medium-term projection, in which output growth is kept very low by the need to correct the balance of payments deficit. Given such a poor prospect for activity and employment it may be realistic. On the other hand it has to be acknowledged that there is no recent experience in this country of inflation rates at or near zero. It is difficult to judge in particular, the extent of resistance to cuts in money wages. Our model, being estimated over a period of persistent, and usually fast, inflation will not take any account of this kind of `non-linear' effect. Other countries with recent experience of zero or near zero inflation, Japan or Germany for example, are not very helpful in this judgement, as in those cases their exchange rates have been appreciating sharply. Our projections are more remarkable in suggesting that zero inflation could be made compatible with a constant exchange rate, in a world environment where prices are still rising, if only at a modest pace.

By construction the effect on the current balance is zero right at the end of the solution period in 1997. Throughout the period shown in the table, however, the current balance is adversely affected. The loss of competitiveness at its maximum is about 6 per cent. Nevertheless it is possible that the exchange rate would be sustainable as shown, provided that the prospect of continuing fiscal stringency is believed. By the end of the forecast period the level of competitiveness in the variant is no different from that in the base run.

A variant on the supply side In the course of last year we completed a thorough overhaul of the equations in our model which describe the behaviour of manufacturing industry. The result were described in the Review last November(4). Particular attention was paid to the way in which fixed investment adds to the growth of capacity, thus reducing both the volume of imports and the prices of manufactured goods.

The base run projections include a low level of investment, partly because interest rates remain high. It is possible that investment will be much more resilient, and rise rapidly now despite the high cost of borrowing. This could happen because firms expect a higher growth of demand for their products than the base run shows, or because they know that they need to undertake new investment so as to respond to the opportunities or threats of a closer integration of the UK with the rest of the European market. It is not unlikely, indeed, that the level of investment last year was higher than the recorded data so far indicate.

In the variant we raise fixed investment in the manufacturing sector quite substantially; by 15 per cent relative to the base in 1989, by 30 per cent in 1990, 20 per cent in 1991 and 10 per cent in 1992. The effect of this extra investment is to raise the capacity of the sector by 4 per cent.

Initially the main effect on output comes from the extra spending on investment goods, together with the `multiplier' consequences that has for consumer spending. But in the later years of the variant the `supply side' consequences dominate. Consumer spending is higher because the lower level of prices enhances the real value of incomes and of wealth. Exports are increased in volume as their prices are reduced, and the import propensity is lowered because there is more productive spare capacity at home. The addition to GDP by 1997 is over 1 per cent, and the effect on manufacturing output is over 2 1/2 per cent. Employment is almost unchanged, because productivity is enhanced by using more and newer machines. The effect on the current balance by 1993 is zero, having been unfavourable in the earlier years.

The effect of this `supply side' variant on the current balance is of some interest in the present context. It is widely hoped that the scale of the present deficit may be reduced, not only by cutting back the growth of domestic demand, but also by improving industrial performance. Obviously this would be helpful, but the extent of the help in our model simulation is rather modest. It is worth spelling out why this is the case.

Part of the problem is that much of the investment needed to improve industrial performance must be imported. It is inevitable therefore that the current account should get worse in the early years. That in turn affects the stock of financial assets outstanding, and subsequent net interest flows.

The other point to make is that much of the improvement in trade performance shown in the model simulation is transmitted through a reduction in prices. Lower prices make UK goods more competitive, but they also make the terms of trade worse. A fall in the UK price level does not work just like a gain in competitiveness brought about by exchange-rate depreciation. Export profit margins for example may be made worse, not better; and the effect on the sterling value of interest payments is quite different. Moreover a fall in the UK price level makes UK consumers better off, whilst exchange depreciation makes them worse off. Thus a lower price level raises consumption and imports offsetting much, if not all, of the effect on the current account of improved competitiveness and better supply-side performance. The corollary is that monetary or fiscal policy might actually have to be tightened rather than relaxed in this case.


REGIONAL UNEMPLOYMENT Unemployment has fallen in all regions since the national peak of mid-1986. Indeed it has fallen fastest in the West Midlands, Wales and the North West regions which were hard hit in the early 1980s. But, as unemployment rose in those recessionary years, the unemployment rate rose fastest in the regions of already relatively high unemployment, so it would be expected that as unemployment recedes so it should return to rates similar to those last seen when national unemployment rates were last at the current level. This measure rose rapidly from the end of 1979 to the end of 1982 and then more slowly, peaking in mid-1987--since when it has fallen slowly. What is noteworthy is the failure of this measure to fall back to the levels of early 1981, as has unemployment. If a snapshot is taken of the unemployment rates now and early 1981 when nationally they were identical a divergent picture emerges. The regions with the largest decrease in unemployment relative to 1981 are East Anglia, the South East excluding Greater London and the West Midlands while Northern Ireland, Scotland and interestingly Greater London suffer relative rises in their unemployment rates. This picture concurs much better with the reports of tight labour markets in suburban and rural South East England and East Anglia, while much of the northern part of the UK has remained slack. The exceptions to this are the West Midlands, which has done relatively well since 1981 perhaps reflecting the very sharp decline of the car industry prior to 1981, and Wales which has held its own.

The snapshot approach while showing clear evidence of a worsening in the concentration of regional unemployment between 1981 and 1988, only usefully describes preiods with similar unemployment levels. It is less easy to choose the most appropriate way of measuring the regional variation over time comparing dates when the average level of unemployment is different. One possibility is to sum the absolute deviations of the regional unemployment rates from the national. This may be an appropriate way of measuring the consequences for the labour market nationally and the pressure for wages increases if the effects of unemployment are non-linear. If the effect on wages of each additional percentage point of unemployment declines, as is often found to be the case nationally, (but the regional picture is less clear), then for the same national rate greater deviation in the regional rates means less downward pressure on wages. The same would be true if the regions of low unemployment were disproportionately important in the wage setting process. However, all regions have the same weight in this measure, whereas their importance to the national picture is less uniform. An alternative is to use a measure of the differences between the proportion of the unemployed in each region and the proportion of the total workforce. This is akin to a measure proposed by Jackman and Roper, but substitutes the workforce for vacancies. As this is a proportionate measure it is not directly related to the overall level of unemployment, if unemployment doubles in all regions the measure would remain unchanged. What is poignant is that as unemployment has again receded this measure has risen to a level not seen since mid--1974, when it stood at only 0.5 million.

All this evidence then shows a deteriorating regional situation. For as unemployment is falling it is doing so disproportionately in the already relatively tight labour markets of southern England outside London. This, if it is true, would point to rising wage pressure if the south is a leading area in wage setting or if the effects of unemployment are not linearly related to the rate. But there is some doubt as to the reasons for the fall of unemployment in all regions over the last two years. There has been a tightening of the availability of work criterion and Restart interviews have resulted in some claimants losing eligibility to benefit. If the success of this clampdown is through weeding out people working in the black economy, or with dependents, illnesses etc. that prevent them from working, these people will be identified more easily where work is available, and where benefit offices have fewer claimants to cope with. Thus the impression of a worsening regional problem could be partly spurious. REFERENCES Jackman, R. and Roper, S. `Structural unemployment', Centre for Labour Economics Discussion Paper No. 233. NOTES (1)See Andrew Britton(1986), The trade cycle in Britain, 1958-1982, Cambridge University Press p.71. (2)See Andrew Britton, op. cit. p.77-81. (3)For an account of exchange-rate estimation and the Institute's model see R. Barrell et al. `Three forward-looking exchange-rate equations and their properties', NIESR, mimeo. (4)See Simon Wren-Lewis, `Supply, liquidity and credit: A new version of the Institute's domestic econometric macromodel', National Institute Economic Review, November 1988, No. 126 , pp.32-43.
COPYRIGHT 1989 National Institute of Economic and Social Research
No portion of this article can be reproduced without the express written permission from the copyright holder.
Copyright 1989 Gale, Cengage Learning. All rights reserved.

Article Details
Printer friendly Cite/link Email Feedback
Title Annotation:Great Britain; Chapter 1
Publication:National Institute Economic Review
Date:Feb 1, 1989
Previous Article:Summary and appraisal.
Next Article:The world economy.

Terms of use | Copyright © 2016 Farlex, Inc. | Feedback | For webmasters