Hostname: page-component-78c5997874-m6dg7 Total loading time: 0 Render date: 2024-11-15T09:16:09.587Z Has data issue: false hasContentIssue false
  • English
  • Français

Forecasting Imports

Published online by Cambridge University Press:  26 March 2020

W. A. H. Godley  and
J. R. Shepherd
W. A. H. Godley
Affiliation:
H. M. Treasury
J. R. Shepherd
Affiliation:
H. M. Treasury
Get access Rights & Permissions [Opens in a new window]

Extract

Two methods are used in the preparation of the official import forecasts in Britain. There is the detailed commodity-by-commodity approach, designed to bring in any special factors or information about particular commodities. There is also the aggregate approach, which attempts to find relationships between the volume of imports and various national expenditure series. This note describes a series of experiments conducted to investigate and, if possible, improve the second of these two approaches. It does not attempt to provide a full theory of import behaviour. It is simply concerned to find a method of import forecasting that works reasonably satisfactorily over a short period of 1-2 years.

Type
Research Article
Information
National Institute Economic Review , Volume 33 , August 1965 , pp. 35 - 42
Copyright
Copyright © 1965 National Institute of Economic and Social Research

Access options

Get access to the full version of this content by using one of the access options below. (Log in options will check for institutional or personal access. Content may require purchase if you do not have access.)

References

Notes

note (1) page 35 See ‘Short-term economic forecasting in the United Kingdom’, Economic Trends, August 1964, page X.

note (2) page 35 See table 1, page 5 of this issue.

note (3) page 35 This is the arithmetic mean of the error terms (calculated irrespective of sign). The estimated standard error is £18.6 million. (The standard error, amongst other things, takes account of the number of variables included, whereas the mean error does not.)

note (4) page 35 These were corrected for various factors, as described in the footnotes to table 1.

note (5) page 35 ‘Measuring national product’, W. A. H. Godley and C. Gillion, National Institute Economic Review, No. 27 February 1964, page 61.

note (6) page 35 However, this conclusion cannot be regarded as certain. The standard error of the coefficient is large enough for the result to be consistent with a true coefficient similar to that for final sales.

note (1) page 36 ‘Capacity’ output is defined in terms of ‘equilibrium’ unemployment—that is, the unemployment that would occur if employment were fully in adjustment to the current level of output. The estimate of ‘capacity’ in this sense is discussed in ‘Long-term growth and short-term policy’ by W. A. H. Godley and J. R. Shepherd, National Institute Economic Review No. 29, August 1964, page 26.

note (1) page 37 In fact, one of them (equation 3) shows quite a significant inverse relationship between the pressure of demand and imports, with imports falling as pressure on capacity rises. This must be a fluke result.

note (2) page 37 Excluding non-ferrous metals from semi-manufactures.

note (3) page 37 However, the standard errors for both the coefficients are large, and the results are consistent with import contents which approximate to those given in the Input-output tables.

note (1) page 38 Stocks are classified as ‘materials’ or ‘finished products' from the point of view of the enterprise which holds them. The materials of one industry will often be the finished products of another. The distinction, therefore, between stocks of basic materials, and of highly processed manufactures cannot be made at all sharply, and there is less reason than might appear at first sight for believing that the import content of stocks of’ materials ‘must be much higher than that of’ finished products’.

note (1) page 39 Measured by the ratio of the mean of non-food imports to that of final demand.

note (1) page 42 The mean errors of the equations with imposed coefficients (equations 20 and 21) are £12.4 million and £11.6 million, as compared with £11.7 million and £11.4 million respectively in the comparable equations with freely determined coefficients (equations 2 and 8). It is of some interest that, if the imposed coefficients are treated as determined entirely without reference to the data, the standard errors of these equations are no greater than (indeed in one of the two cases smaller than) those of the comparable equations with freely determined coefficients. Whether the standard error or the mean error is used as a criterion of goodness of fit, it should be appreciated that, when the equation is used for forecasting, the errors are always likely to be somewhat bigger on average than those in the period over which the equation was fitted.