V. METHODS OF HANDLING THE DATA

The problem of securing an index of production is hardly solved merely by getting a result. Rather, it is necessary to know with some precision what the result means. We have examined the meaning of our formula. Besides this, it is necessary to determine how the result was secured from the data available. In doing so, we shall understand better why our general result contrasts with previous indices.

The question of which industries were to be included in our index was substantially settled for us by the scope of the census of production." This limits us to the fabrication of material goods, including capital equipment, and excludes personal and professional services as well as the enhanced utilities from the direct services of traders subsequent to manufacture. It is important, however, that the census does include the public utilities. The great increase in this branch was not measured by the Rowe index. The only important census group we omitted was that of railways-principally because the measure of production available was unsatisfactory, but also because the industry furnishes a direct service which is outside the scope of this index.

In general we applied our formula not only to every industry but to every commodity for which the two censuses provided a satisfactory quantity comparison.12 This goes far to explain why an index based on the census gives a higher result than the pre

"In geographical scope, the comparability of the census returns seems at first to be violated, since both Great Britain and Ireland were included in 1907 while Southern Ireland was omitted in 1924. The difficulty is more apparent than real, 'however. Southern Ireland has comparatively little industry, and the 1924 census reports provide a comparison excluding Ireland altogether wherever the discrepancy would be important. Such exclusions of product as there are in 1924 of commodities included in 1907 tend, of course, to lower the value of our index (cf. pp. 8, 12, supra). Notice that subsequent census reports on Northern Ireland merely tabulate separately the results already included here.

The 1924 reports contain many more returns by quantity than those for 1907, but they are useless for our comparison unless a 1907 base can be supplied. We are hampered also by some unfortunate changes in the unit of quantity reported. The change itself may be for the better, as when woolen piece goods were reported by square yards instead of linear yards (1924 Reports of the Census of Production, No. 29, pp. iv-vi, op. cit., note 5). But the value of censuses is largely destroyed if enumerators fail to ask the same question for at least two periods before they discontinue it.

the popular imagination that it was described as the "workhouse exodus."

The number of men who were workhouse inmates fell from 123,070 on January 1, 1913, to 62,611 in 1919, a decline of 60,459; the number of women declined from 80,073 to 60,738 in the same period; the number of children in workhouses or poor-law schools fell from 72,149 to 59,761.

Commenting on the fact that the pauper lists for indoor relief fell off more rapidly than the lists for outdoor relief in 1914-15, the London Times called attention to the large numbers of persons who left the workhouses as early as January, 1915, while the outdoor figures remained high. "Nobody watches the variation in the condition of the classes bordering upon pauperism more keenly than the workhouse inmates," said the Times, "and the opportunity of sharing in the distribution of the wealth which the war had brought was promptly taken advantage of, and many began to take their discharge."

The falsity of the old and shameful theory that people prefer public charity and the slothful existence of the workhouse to an independent life of hard work was never more completely exposed. The response to improved opportunity of employment was instantaneous; the demand for male labor was immediately met by the men who had been rejected in the labor market and driven into the workhouse in time of peace, but who seized the first opportunity to escape to normal conditions of life and work.

Pauperism declined during the war period in Scotland and in Ireland as well as in England. In Scotland as in England, a decline in the number of persons in receipt of relief took place after 1911, a decline brought about by the old-age pension acts; but the effect of the war was to cause a very sharp and continuous fall in the number of indoor paupers after 1914, and in the number of outdoor paupers after 1915. Comparing the year 1913 with the year 1918, the number of persons receiving indoor relief fell from 13,240 to 7,653, a decline of 42.2 per cent; and the number of persons receiving outdoor relief fell from 73,953 to 58,913.

1 Times (weekly edition), December 31, 1915. In England the "workhouse" is, of course, a poor-law institution, and not, as with us, a house of correction.

Sampling many commodities within an industry introduces, however, a problem of weighting. We have settled that weighting should be according to the value of commodities. Ideally, this should be a figure expressing value added by manufacture, for it is this value which determines the importance of any one commodity in the production of any one year. The British census has no such figure, but it does report the "net output"-sale value minus cost of all materials-and this is a reasonable approximation. Net outputs were, therefore, selected for our weights as between industries. But as between different commodities in the same industry, no such figure is available16 and sale value itself had to be used.

It is a second corollary of increasing fabrication that the proportion of raw-material value should decrease relative to the value of work done on the goods." Sale value includes the value of raw material. Therefore, the comparatively unfabricated goods in which a large proportion of the value is due to raw material will be overweighted and the higher fabricated goods will be underweighted. And since the fabricated types have increased faster than the rest, such a process will tend to lower the value of our final index. We see no solution for this difficulty except to limit the extent of the error by using net output weights between all industries which are separately reported. This furnishes an additional reason why even our index probably underestimates 1924 production.

Besides this general explanation of method, some explana

dex, since you must define the unit priced before you can construct a price relative that has meaning. When this has not been done, but instead the price index is based on a mere sample of a few commodities, the measure of fabrication afforded may be wholly illusory.

"This is partly for lack of detailed returns of costs which would be very expensive to collect and partly because of the sheer impossibility of getting anything but an arbitrary figure for separate commodities wherever these are affected by joint costs.

47

This indicates another reason why previous indices have been too low. So far as they have relied on data as to raw material consumed, they have obviously overlooked the additional utility bestowed on a given unit of raw material by increasing fabrication. Cf. Rowe's use of import returns on cotton ("Special Memo. No. 8," op. cit.).

pose of the committee was not merely to give temporary relief to the vagrant but to assist him, through a co-ordination of the voluntary agencies, to make a new start in life.

The first report of the London "Advisory Committee on the Homeless Poor" contains an account of the situation in London during the decade preceding the formation of the committee. There had been an increase in the number of homeless persons and vagrants beginning in 1901, after the South African War, and this increase appears to have been well maintained until the years 1912-13. The report of the committee contains the following statement (p. 6) regarding the situation in London following the South African War:

Many of these [homeless persons and vagrants] flocked to the Metropolis and swelled the ranks of those destitute persons who collected nightly in Central London. The Embankment had for years been the chief center at which such persons congregated. They were originally attracted to this spot by such shelter from the weather as the bridges afford, and probably to the locality generally by the proximity of the theatres and restaurants, where they could often earn a few pence by fetching cabs or carriages when those places of entertainment closed. The staircases and landings of houses have always afforded refuge to some, but their numbers are relatively few, and those making use of such shelter are usually found in the proximity of common lodging houses in which they would have slept at night, had they the few pence required to pay for a bed. These lodging houses are especially numerous in Whitechapel, and it is in this district, and in particular streets of it, where destitute persons thus sheltering themselves have been most frequently discovered when search has been made for them. It is here men and women, often with a child, take advantage of the open door of a tenement house, and there can be no question that their condition of destitution is extreme.

The census of "homeless persons," vagrants, and casuals taken annually for the London County Council on a winter night showed 6,644 persons enumerated February 18, 1910, in casual wards or other places where free beds could be found, or in the streets or sitting up in shelters, but the number had fallen to 2,881 on the night of February 13, 1914. These homeless persons were supposedly "able-bodied" men and women who could not pay

1 See the Report of this committee to Mr. John Burns, President of the Local Government Board, to December 31, 1913 (Cd. 7307), pp. 6–7.

for even the cheapest shelter and finally accepted, because they must, the shelter of the casual ward, for the casual ward is said to be a last resort for those who "sleep out," probably because of the compulsory bath and the labor task.

The decline, during the five years before the war, in the number of homeless persons may have been due in some measure to an improvement in the employment situation and to the effect of the new social machinery that had come into operation, labor exchanges, health insurance, and old-age pensions; but the change was certainly due in part to the improved methods of dealing with the vagrant class and, in particular, to the transfer of the casual wards to the Metropolitan Asylums Board in 1913.

From whatever cause, it is clear that vagrancy was declining in London before the war. The immediate effect of the mobilization was to accentuate this decline. The figures given in Table XII show the number of casual paupers in London reported at the close of each of the four weeks of August, 1914, in comparison with the same weeks for 1913.

The immediate decrease in vagrancy in London is of great interest because it occurred at a time when the number of indoor and of outdoor paupers was increasing. In London, apparently, the men who had been classed as beggars, "wastrels," and "won't works" must have joined the colors very promptly. The falling off in numbers could not, at this time, be explained on the ground that they had suddenly found employment outside of the army, for unemployment was increasing instead of decreasing during these weeks. The decline in the number of casual paupers con