^{note (1) page 35} R. R. Neild and E. A. Shirley, ‘An assessment of fore casts, 1959-1960’, National Institute Economic Review No. 15, May 1961, page 12.

^{note (2) page 35} Manufactures made up 84 per cent of British exports in the period 1955-66, with little variation.

^{note (3) page 35} See, for example, Fessey, [3].

^{note (4) page 35} National Institute Economic Review No. 39, February 1967, page 17; chart 7, page 18; and table 7, page 19.

^{note (5) page 35} This excludes exports to the United Kingdom, but includes United Kingdom exports to other industrial countries.

^{note (1) page 36} OECD Economic Outlook, July 1967; EEC, The Economic Situation in the Community, published quarterly.

^{note (2) page 36} G. D. N. Worswick and C. G. Fane,’ Goods and services once again’, District Bank Review, March 1967.

^{note (3) page 36} There are statistical problems in measuring the rise in productivity of many services; and employment is often used to measure output. This explains some of the absence of volatility of non-manufacturing production.

^{note (1) page 37} For the period 1950 to 1966, two ‘calculated’ series for the aggregate manufacturing production index were prepared : first, by adding together the national product series for the countries concerned (weighting them by their share in total imports of manufactures), and then converting this total national product series into an aggregate production index. Secondly, the conversion from national output to manufac turing production was made for each individual country, and the individual manufacturing production series were then added (weighting them again by each country's share in total imports of manufactures). When these two calculated series for industrial countries' aggregate manufacturing production were compared with the actual series (also reweighted), the average mean error, regardless of sign, was only 0.58 per cent for the disaggregated method, and 0.82 per cent for the aggregate method.

^{note (1) page 38} The aggregate relationships for all industrial countries, writing P_m for manufacturing production and G for gross national product, are :

^{note (2) page 38} In fact trade throughout is measured from the export side.

^{note (3) page 38} A. Maizels, Industrial Growth and World Trade, NIESR, table 4.1 and figure 4.1, page 81. From 1957-59 (where Maizels' figures stopped) to 1966, world production of manufactures has risen 62 per cent, and world trade in manufactures (by volume) 85 per cent.

^{note (4) page 38} Consequently there is no reason to put any confidence in the coefficient of the prices term in equation (3), which suggests that a 1 per cent rise in prices of manufactures would induce a 21/2 per cent rise in the value of trade.

^{note (1) page 40} There is other evidence for the significance of the relative prices term. For example, when it is included in the share equation (equation (2)), it improves the explanation, and the coefficient of the prices variable is significant and of the right sign :

^{note (2) page 40} A number of possible variables could not be tried because quarterly data are lacking, for instance, for stockbuilding or for unfilled orders.

^{note (1) page 41} In the United Kingdom, the import content of finished manufactures appears to be significantly higher in stock building than in other final sales. (W. A. H. Godley and J. R. Shepherd, ‘Forecasting imports’, National Institute Economic Review No. 33, August 1965, table 4, equation 13, page 39.) Estimates of the share of imports in fixed capital formation in machinery and equipment and in consumers' expenditure on household equipment and personal effects are given for the major industrial countries over the period 1956-62, in W. Beckerman et al., The British Economy in 1975, NIESR, table 5.6, page 162.

^{note (2) page 41} This was also true when the lagged effects of manufac turing production in individual past quarters were separately determined.

^{note (3) page 41} National Institute Economic Review No. 39, February 1967, table 6, page 16, column (2).

^{note (1) page 42} Conventionally, price-elasticities refer to the demand for the volume of exports. Estimates of this elasticity are given by subtracting unity from the regression co-efficient of relative prices.

^{note (2) page 42} Junz and Rhomberg [4] found a positive elasticity of substitution between British and other industrial countries' exports. Renton [5], after testing a wide variety of types of prices effect, failed to estimate a significant negative relation ship between the value of United Kingdom manufactured exports and relative prices. Ball, Eaton and Steuer [2] failed to estimate the effect of relative export prices on the level of United Kingdom exports though they estimate a significant negative effect on the United Kingdom export share. Smyth [6], using annual data, found a negative relationship between percentage changes in the volume of British manu factured exports and percentage changes in relative export prices (the partial elasticity is estimated at about — 1), but not between changes in the value of United Kingdom exports of manufacturers and changes in prices.

^{note (1) page 43} Smyth [6] found an asymmetry of this kind. Further, his results support the view that it is the rate of change of pressure and not just the absolute level which is the more important determinant of exports.

^{note (2) page 43} This is the percentage change in the United Kingdom's percentage share.

^{note (1) page 44} The effect of pressure on total exports of manufactures has been calculated by Ball et al. [2], page 516.

^{note (2) page 44} For example, in the medium-term export forecast for 1970 incorporated into the February 1967 Economic Review (No. 39, page 19) it was assumed that an improvement in competitive ness would slow down the rate of fall.

^{note (3) page 44} The relative pressure term is not explaining all the short- term variations around the trend in the United Kingdom export share.

^{note (1) page 45} Adams et al. [1] found that the effects of changes in relative demand pressure on the volume of total exports varied a good deal among industrial countries; the range of elasticities was from — 3.27 (United States) to — 0.46 (Italy), and the United Kingdom elasticity was estimated as — 1.75.

^{note (2) page 45} These forecasts were prepared before the sterling devaluation of 18th November, 1967 and exclude any effect on exports of the devaluation.

^{note (3) page 45} National Institute Economic Review No. 38, page 34, table 20.

^{note (4) page 45} National Institute Economic Review No. 41, page 17, table 10.

^{note (5) page 45} The residuals in the level relations between manufacturing production and gross national product and in the imports forecasting regression (5) showed some positive serial corre lation. An attempt was made to correct for this by calculating the forecasts from the first difference form of the equations.

^{note (1) page 47} The forecast rise is 2.7 per cent. This is weighted by 1965 import shares; consequently the aggregate forecast, though not the country forecasts, differs from that given on page 23, table 15.

^{note (2) page 47} Intra-trade between industrial countries, including non- manufactures as figures excluding them are not yet available, was running in the first half of the year about 6 per cent higher than the 1966 average.

^{note (3) page 47} Another possibility is that there are ‘cut-out’ points for the effect of pressure on demand; beyond certain levels of very high and very low pressure, further changes have little extra effect on exports.

^{note (4) page 47} This is the sum of the changes in the individual gross national products given on page 23, table 15, weighted by 1965 import shares.

^{note (1) page 48} An actual outcome in 1967 greater than the expected 3 1/2 per cent would, of course, give a projection for 1968 of less than 7 per cent.