^{note (1) page 52} W. A. H. Godley and J. R. Shepherd, ‘Forecasting imports’, National Institute Economic Review No. 33, August 1965, page 35.

^{note (2) page 52} That is, imports of goods and services at constant (1958) prices.

^{note (3) page 52} Given that the discrepancy is a discrepancy between estimates of gross domestic product at factor cost, the coeffi cient should strictly speaking be compared with the coefficient which would be obtained for total final sales if the equations had been run throughout at factor cost. Had they been so run, the coefficients for total final sales in equations (1) to (3) would probably have been about 10 per cent higher. So the equation results do not in fact suggest a disproportionately high stockbuilding content in the unidentified expenditure.

^{note (4) page 52} The two equations recommended were as follows :

where M=Imports of goods and services; F=Final sales; S=Stockbuilding; t=Time trend; D=Difference between expenditure and ‘compromise’ estimate of gross domestic product; P=Public authorities’ current expenditure; I= Gross fixed investment; and X=Exports. Standard errors are shown in brackets beneath regression coefficients.

^{note (5) page 52} The constant term was positive. This implies that the average import content of total final sales at the beginning of the period was slightly higher than at the end of the period.

^{note (6) page 52} The marginal import contents implied were 30 per cent for consumers' expenditure, 7 per cent for public authorities' current expenditure, 1 per cent for gross fixed investment, and 24 per cent for exports. There is such a high degree of inter correlation between all the components of final sales that it is not surprising that regression analysis fails to distinguish the individual coefficients satisfactorily.

^{note (1) page 53} An extensive exercise in the use of this approach was carried out by J. Johnston and Margaret Henderson, Assessing the effects of the import surcharge, The Manchester School of Economic and Social Studies, May 1967.

^{note (2) page 53} The series used were those given in Economic Trends, October 1967.

^{note (3) page 53} The constant term in the equation was negative.

^{note (4) page 53} The 1963 coefficients used in this equation were not, however, very different from those for 1954 (shown in the equation in footnote (4), page 52.

^{note (5) page 53} Johnson and Henderson estimated a total import savings effect of £100-150 million.

^{note (1) page 54} The figures in this paragraph are in fact for the annual average changes from 1958-60 to 1964-66.

^{note (2) page 54} Harvests were better in the second period; increased importance of animal husbandry and modern methods made agriculture less dependent on the weather; with the spread of convenience (and, in general, more processed) food the ‘processing’ content, mainly performed in the United Kingdom, of expenditure has been growing; agricultural policy helped too (one example is the increase of the area under sugarbeet).

^{note (1) page 57} On a balance of payments basis.

^{note (2) page 57} Allowing for £20 million special stockbuilding of sur charged goods.