⁽¹⁾ Errors measured as a percentage of the level in the base period.

⁽¹⁾ The forecasts of investment made at the beginning of each year were for the whole of that year, compared with the whole of the previous year; estimates of the likely level in the fourth quarter were then derived from the annual figures.

⁽¹⁾ The increase may have been accentuated by a slight change in the procedure followed in the questionnaire between February and June.

⁽²⁾ See National Institute Economic Review, no. 14, March 1961, pages 5, 16 and 17.

⁽¹⁾ There was a larger rise from the fourth quarter 1959 to fourth quarter 1960, but this was due to a hump in payments which appears to have no significance. See Economic Survey 1961, Cmnd. 1334, page 15.

⁽¹⁾ See ‘Britain's falling share of sterling area imports’, National Institute Economic Review, no. 14, March 1961.

⁽¹⁾ See National Institute Economic Review, no. 9, May 1960, page 8.

⁽²⁾ We changed our view because the employment figures for 1959 were revised upwards substantially in February 1960. Using the unrevised figures for November 1959 we had calculated in January 1960 that employment could rise 11/2 1 2 per cent, if the same proportion of the population in each age group was drawn into employment as in 1957; 1957 was the year when the labour force had been at its previous peak. So when the end-1959 figure of employment was revised upwards by 1/2per cent it reduced our apparent reserve of labour.

⁽¹⁾ Until July 1959, the only official stock figures published were in current prices; and they were not seasonally adjusted until October 1960. We adjusted them ourselves until the official adjusted figures were available. The change from our seasonal adjustments to the official ones accounts for most of the alterations in the figures between July and November 1960 (chart 7). Otherwise, the alterations are due almost entirely to revisions to the official series.

⁽²⁾ Part of the change back since the beginning of 1960 is accounted for by the change-over to official seasonal adjust ments.

⁽¹⁾ See conclusions to ‘Price stability and the policy of deflation’, National Institute Economic Review, no. 3, May 1959, page 29.

⁽²⁾ The rise from October 1959 to October 1960; no quarterly figures for earnings are available.

⁽³⁾ ‘Imports and expansion’, National Institute Economic Review, no. 2, March 1959.

⁽⁴⁾ Recent changes in methods of estimation have made big differences to the figures (see page 7). The surplus for 1959 is now estimated at only £51 million.

⁽¹⁾ For an analysis of the fall in invisible earnings see ‘The fall in Britain's invisible earnings’, National Institute Economic Review, no. 12, November 1960.

⁽²⁾ See in particular ‘British imports of manufactured goods’, Economic Review, no. 8, March 1960 and World trade in manufactures’, Economic Review, no, 10, July 1960.

⁽³⁾ Private forecasting of this kind was a more popular sport in 1959, when there was considerable slack in the economy and an expansionary policy could confidently be recom mended, than in 1960, when the amount of slack was fast declining and it was hard to know whether to recommend a restrictive budget or no change.

⁽¹⁾ See Appendix page 29.

⁽²⁾ Errors may also arise because all historical data are treated as if they were reliable. No allowance is made for the fact that the latest statistics are more likely to be revised than others.

⁽³⁾ A. Hazlewood, P. Vandome, ‘A Post Mortem on Econometric Forecasts for 1959’, Bulletin of the Oxford University Institute of Statistics, vol. 25, no. 1, February 1961.

⁽¹⁾ See, for example, H. Theil, ‘Who Forecasts Best?’, International Economic Papers, no. 5, 1955, and H. Theil, Economic Forecasts and Policy (North-Holland Publishing Company, 1958), chapter II.

⁽²⁾ The forecasts for 1956 and 1957 were not calculated on exactly the same basis. For example, the forecasts for 1956 made in the summer of 1955 was related to the revised estimate for 1955 made at the same time and not to the forecast for 1955 made in the summer of 1954 .

⁽¹⁾ See footnote

⁽²⁾ on page 27.