In his report already referred to, Sir Robert Giffen, after detailing the average rates of the above table, says (p. xxxiii): "Thus in nearly all these trades the average rates are about the same as the average rate in the Census of Wages Summary." But the table does not include the badly paid agricultural labourer, the largest group of all, and the figures for seamen, etc., are, it should be observed, swollen by estimates of the value of board and lodging.

Finally, Sir Robert Giffen arrived at the general conclusion that "the broad results shown by the census summary would not be sensibly modified by including the great mass of other employments not comprised in that summary."

In January 1893 Sir Robert Giffen gave evidence before the Labour Commission and submitted the facts I have detailed. He prepared a general estimate of the proportion of the national income then taken by the wage-earning classes, and his evidence on this point (questions 6909 to 6914) is summarized in the following table:—

EARNINGS OF MANUAL LABOURERS IN 1886
(Sir Robert Giffen's estimate for the Labour Commission)

There can be no question that this estimate of Sir Robert Giffen's somewhat exaggerated the actual earnings of manual labourers as a whole. In the first place, it was too much to assume that the 24s. 9d. per week or £64 per annum was representative of the whole of adult male labour. Without introducing agricultural labourers (the largest group in the country), general labourers, postmen, and other ill-paid workers, the unweighted average of the table on page 24 is £62. If £60 per annum had been given as the average rate of wages of all the adult male workers in 1886 it would probably have been an exaggeration. It was not given as a rate of wages, however, but as the actual earnings of the men after all allowance made for short time, unemployment, sickness, accidents, strikes, lockouts, stress of weather, etc. Sir Robert Giffen appears to have assumed that all the adult male workers of the United Kingdom were employed on the average about 50 weeks out of 52, and were paid at the average rate of £64 per annum!

In 1866 Leone Levi, in estimating the manual workers' earnings, assumed that four weeks per annum were lost. Dudley Baxter in 1867 pointed out, in criticism of Leone Levi, that if four weeks' "play" were all that need be allowed "England would be a perfect Paradise for working men."[6] Dudley Baxter, in view of the circumstances of his day, allowed ten weeks for "play" in making his estimate, and there can be no question that he was nearer the truth than Levi. At the present day the level of employment is very much the same as it has been for the past forty years, while sickness, accidents, and the weather are still with us. We need not wonder, then, if Professor A. L. Bowley, who has given the subject of wages so much attention, bases his estimates upon the loss of six weeks' work per annum through sickness and holidays, and makes an additional allowance for unemployment, while also assuming that 10 per cent. of the working population only get casual or irregular work, bringing them in about half the amount shown in the Wage Census.[7]

If the estimate given to the Labour Commission had allowed for six weeks' "play," the average earnings of men, women, boys and girls would have come out at £40. 5s. per annum instead of £48, and the aggregate earnings, therefore, at much less than £633,000,000. Leone Levi's estimate for 1884, allowing for only four weeks' play in the year, was £521,000,000. This figure is too large, but it is over £100,000,000 less than that of Sir Robert Giffen.

I now take the Wage Census figure of 1886 as a basis and correct it for the upward movement of wages since that date by the wage index numbers of the Board of Trade (Cd. 4954, which slightly corrects the index numbers of Cd. 1761, used in "Riches and Poverty," 1905 edition, p. 24), which are based on the mean of over 150 rates:—

* The meaning of this column is that, if the average wage of 1900 be represented by 100, the average wage of 1886 is represented by 82·86 and that of 1908 by 101·02.

We thus arrive at 21s. 3d. as the average weekly wage of the manual workers in 1908. There is much reason to believe that this estimate errs on the side of liberality. It is unfortunate that we have not a compulsory wage census, and the method of estimation used here can pretend to no more than approximation. It neglects the important fact that between 1886 and 1908 the ranks of women and child workers have swollen at the expense of adult male workers. The 15,500,000 (estimated) manual workers of 1908 consisted as to a larger proportion of women and children than the 13,200,000 (estimated) manual workers of 1886. I regard the 21s. 3d., therefore, as the most liberal figure that can be put forward as the average earnings of the men and women and child workers of the United Kingdom in 1908.

We have now to decide what allowances should be made (1) for the great army of casual, incompetent, and aged or ageing workers who figure in the census returns as following definite occupations, and (2) for the loss of time through unemployment, sickness, accidents, stress of weather, strikes, lockouts, "bank" and other holidays, etc., in the case of the remaining workers.

With regard to the first item, I do not think we are justified in estimating the incompetents and casuals at less than 1,000,000 out of the 15,500,000. For the purposes of the present estimate, I assume that these 1,000,000 workers earn, on the average, £25 per head per annum, or an aggregate of £25,000,000. My view is that this is a liberal estimate of the earnings of what may be termed the camp-followers of the industrial army.

With regard to the remaining 14,500,000, we have to form an estimate of the amount of time lost per annum through voluntary or enforced leisure. No certain information exists, and the widest differences of opinion have been expressed on the subject. As I have said above, Dudley Baxter took ten weeks; Leone Levi took four weeks; Mr A. L. Bowley takes six weeks plus a further allowance for unemployment.

The Board of Trade, in their recent examination of fluctuations in employment, made an analysis from the records of the Amalgamated Society of Engineers, combined with information supplied by employers, of the time lost in the engineering trade. They came to the conclusion that, in an average year, perhaps 8 per cent. of working time was lost from all causes, and expressed the opinion that in a good year the loss might fall to 4 per cent. and in a bad year rise to 15 per cent. or more (Cd. 2337, p. 101). This would mean, for the engineering trade only, a loss of time varying from only two weeks in the year to as much as eight weeks or more.

In other employments the widest variations exist. There are the quite regular employments, such as the army, the navy, the postal service, the police service, and, for the greater part, the railway service. There are violently fluctuating employments, such as the building trades and the shipbuilding trades. In all alike, sickness takes its toll, and unemployment arises from accidents, from disputes, from "drink," and from seasonal influences and depression, while, on the other hand, overtime occasionally goes to swell the aggregate earnings.

I make the assumption that the average working year of the 14,500,000 remaining wage-earners consists of 44 weeks. Applying the average wage already arrived at (21s. 3d. per week), we get an average annual earning of, say, £46. 15s., which gives us £678,000,000 as the probable aggregate earnings of the 14,500,000 workers. Adding the £25,000,000 assumed to be earned by the remaining 1,000,000, we arrive at £703,000,000 as the total earnings of the manual labourers in 1908.

It is probable that this calculation does not take sufficient account either of the changes of occupations since 1886, or, as has been already pointed out, of the changes in the respective proportions of men, women and children employed. The average wage of the 1886 Census, taken as the basis of the calculation, was, it is necessary to insist, exaggerated by the omission of the most ill-paid workmen, while the returns upon which it was based, framed as they were by employers, are only too likely in a proportion of cases to have put the wages paid in the most favourable light. The employers again, who filled in the forms, were only some 75 per cent. of the firms applied to by the Board of Trade, and it is a fair inference that those who neglected to reply had no excessive pride in the records of their wage-sheets. I submit, therefore, that as the 1886 average wage figure is a liberal estimate,[8] the figure which I have deduced from it does not, in all probability, err on the side of under-estimation.

Professor Bowley estimates the total paid in wages in 1901 as £705,000,000,[9] and the Board of Trade in the Fiscal Blue Book of 1903 (Cd. 1761) say:—

"From investigations based on the Board of Trade Census of Wages (1886) combined with the recorded changes of wages since that date and the distribution of the working population among various industries as shown in the census returns, the total wages bill of the United Kingdom has been estimated at between £700,000,000 and £750,000,000, according to the state of employment."

The estimate which I have given, therefore, differs but little from those of Professor Bowley and the Board of Trade.[10] I prefer to use the smaller figures on several grounds. In the first place, the allowance for "play" is a conservative one. In the second place, I have the gravest doubts as to the propriety of including in the estimates of the wages of domestic servants, sailors, and others, an allowance for the value of "lodging," as is done in the figures used. To include so many shillings a week for the accommodation afforded by a seaman's bunk or a general servant's fraction of an attic is to flatter "earnings" out of all resemblance to the truth. The free cottages and other allowances to agricultural labourers are often of a scarcely marketable character. We may be justified in valuing an unhealthy hovel at 1s. 6d. per week, in view of the fact that the labourer, if he had it not, would need to pay rent elsewhere, but in too many cases the "cottage" is fit not for inhabitation but for demolition. In the third place, no allowance is made for the excessive rents paid by workmen in London and other large towns. These rents are really part of the working expenses of the wage earners, and there is as good ground for making deductions on account of them as there is for deducting wear and tear of machinery in the case of income-tax incomes.

We can now arrive at an approximate estimate of the National Income as a whole in 1908-9 (say 1908).

THE NATIONAL INCOME IN 1908

It will be seen that the income tax exemption limit of £160 per annum splits the national income into two almost equal parts. Of a total income amounting to £1,844,000,000 in 1908, those with over £160 per annum took £909,000,000, while those with less than £160 per annum took £935,000,000.