ON THE AMOUNT AND DISTRIBUTION OF INCOME. 185 



a precision which can be assigned. In the schedules which were filled 

 up, the employers — except in the case of companies — included them- 

 selves as making over 160L, but it seems better to adjust the schedules 

 by subtracting 1 in every case from those receiving over 160L, and 

 then treat them as relating only to clerks and other subordinates. In 

 the small firms this is generally accurate, and in large ones it makes 

 practically no difference to the further argument. When this is done, 

 the proportion of those receiving over 160L to the total number em- 

 ployed is written down in each of the 102 classes. The result is given 

 in the table, p. 189. It now appears that the grouping of these pro- 

 portions shows the kind of regularity which is to be expected when 

 samples are taken at random from a definite group. Instances, in fact, 

 as shown in the table, agree fairly closely with the distribution shown 

 by the normal curve of error. Looked at from this point of view it 

 appears that from 1 in 5 to 1 in 3, on an average, of clerks in this 

 class may be expected to receive over 160L Noticing that in the larger 

 firms the proportion is on the whole smaller, and that when all the 

 firms for which we have returns are massed together the proportion 

 falls to 0'21, and aiming at a weighted average since the larger 

 firms have greater importance, we come to the conclusion that 

 the proportion is probably between 1 in 4 and 1 in 5, i.e., between 

 0'20 and 0'25. We will widen this margin for safety and say, then, 

 that 0'23 + 0'06 of male clerks and similar employes pay income-tax, 

 and that therefore 0'77 d: O06 of 406,000 (the whole number in this 

 class), i.e., 312,000 ± 24,000, are below the income-tax limit. The 

 average income for those for whom we have returns is 781., and we 

 assign the margin ± 10Z. 



As regards women, the proportion to men in our returns is nearly 

 the same as shown by the census. The number who pay income-tax 

 is very small, the average income shown is 45L We therefore adopt 

 the number 81,000 at 45Z. ± 10/, 



As will be seen from the detailed table, it would be possible to make 

 a carefully weighted average allowing different importance to the 

 different occupations, but after careful consideration it is considered 

 that no greater accuracy would be obtained. Of our 102 returns 27, 21, 

 and 14 are from the neighbourhoods of Manchester, Birmingham, and 

 London respectively, 10 from the rest of England and Wales, 21 are 

 from Ireland, and 9 from Scotland. London is represented rather in 

 Classes 17 and 18. 



17. Bank Clerks. — Thanks to the Central Association of Bankers 

 we have obtained detailed information as to about 30 per cent, of the 

 bank clerks of England and Wales. For Scotland we have information 

 as to about 80 per cent., and for Ireland we have a less detailed state- 

 ment, which is stated on good authority to be typical for the country, 

 and which simply gives the proportions of clerks who receive over 160Z. , 

 from 100L to 160L, and less than 100L The details, which are sum- 

 marised on p. 188, are highly confidential, and we cannot go into them. 

 They result in. the following numbers, which may be regarded as being 

 accurate within the limits stated: 24,000 ± 1,000 at 89L ± 51. This 

 corresponds to the receipt of an income of 160?. at the age of thirty-two 

 years on an average. 



