Real gross domestic product (average estimate at 1975 prices) is officially estimated to have fallen by nearly ½ per cent in the last quarter of 1978. Nevertheless, the increase for the year as a whole was about 3 per cent. There were falls in the fourth quarter in the volumes of consumers' expenditure (–0.6 per cent), gross fixed investment (–1.8 per cent) and exports of goods and services (–0.8 per cent); public authorities' consumption, however, rose by 1.2 per cent and imports of goods and services fell by 2.3 per cent.

The index of production, like all the national accounts series, has been rebased onto an average 1975 price basis. The series has been fully rebased from 1973 onwards with the earlier figures simply being rescaled.

The most important changes in the world economy since our last forecasts were prepared at the beginning of the year have been the rise in the price of oil and the associated rises in the pound and dollar against other currencies and particularly against the yen.

Despite these developments we have not significantly changed our forecasts of total OECD output in 1979. This is partly because we seem to have underestimated in February the strength of the upward trend in the industrial sector especially in North America. But we do expect higher rates of inflation, partly as a result of the oil price rise, both this year and next, and in 1980 rates of output are likely to slow down rather more than we thought in February. This is partly because of the transfer of purchasing power from oil importers to oil exporters, who will spend only part of it on additional imports, and partly because governments in oil-importing countries must be expected, as in 1974–75, to compound the depressing effects by their policy reactions to faster inflation and higher import bills.

This article aims to compare the National Institute's forecasting performance with that of United States forecasters. Because of the magnitude of this task, we restrict our attention to forecasts of the principal output measure in each country—real gross domestic product in the United Kingdom and real gross national product in the United States. Results for US forecasters are taken from the detailed study by McNees (1975), where there is also information on the methods employed by these forecasters.

This article presents a methodology for decomposing ex ante forecasting error into exogenous variable error, data revision error, model error and judgement error. This methodology is applied to the forecasts made by the National Institute in February 1975 and February 1976. The first section describes the methodology including the NIESR forecasting procedure. Then the NIESR model (with some of its problems) is discussed together with the data used in the study. The methodology for decomposing the forecasting error to 1975 is applied and a similar analysis presented for 1976. Some conclusions and a summary complete the article.