One route to an assessment of the long-run impact of the war on Soviet GNP per head is via econometric estimation of the long-run trend, if one can be found to exist; the war period can then be examined for signs of disturbance in the trend.

Index numbers of Russian and Soviet real GNP can be assembled from the work of Paul Gregory (net national product at 1913 prices, 1885–1913 and 1928), Moorsteen and Powell (GNP at 1937 factor cost, 1928–40 and 1950), and the CIA Office of Soviet Analysis (GNP at 1982 factor cost, 1950–85). Comparable population series are likewise taken from Gregory for the period before 1913, from recent Goskomstat revisions of the interwar population, and from official postwar figures. Tables N.1 to N.3 show the underlying data, and table N.4 shows the associated long-run index (see also figures N.1 and N.2).

Assembled in this way, the series embodies three major difficulties. First is the two substantial gaps, the first covering the period of World War I, the Civil War, and postwar recovery under the New Economic Policy, and the second covering World War II and postwar recovery. The reader will recall from chapter 5 that although for the second war period we have real GNP, we have no good annual population figures; we have annual GNP per worker, but this has little meaning, given the violent and anomalous structural changes of the war years.