While construction is always a volatile sector, its performance in the interwar period was extreme by any standards. Cornwall (1977) has attributed the economic stability of the 1920s primarily to the experience of the construction sector. Schumpeter (1939: 743) felt that if input sectors were properly accounted for, construction was the largest source of growth in that decade. Soule (1947:170) concurred. Construction comprised over 8% of GNP each year 1923–1927. This is the highest ratio for an extended period seen in the twentieth century (Gordon, 1981), though higher ratios had been known in the nineteenth century. Construction expenditure, however, declined from 1926–27. Zevin (1982) and others have attributed much of the original downturn into Depression to the decline in construction; he argues that stock market speculation supported consumption and investment for the intervening two years. Certainly the 40% drop in employment 1928–1933 was worse than the decline in any of the other eight major divisions (Henderson, 1961:45). New house construction fell from 937,000 in 1926 to 509,000 is 1929 to 93,000 in 1933. Expenditure in construction fell from $11.2 billion in 1929 to $3.7 billion in 1933. The Great Depression, like the two wars, would see net disinvestment in housing. Recovery would be painfully slow, lagging far behind all other sectors. The ratio of house construction to output would not regain half of its 1925 value even in 1940.