In a consideration of railway labor I think that the fundamental thing is an examination of the percentage that it is, and has been, taking out of the national income. This appears in the following table:
| Calendar year | Average number employed | Total compensation | Average compensation per employee | Index of wages[t 1] | National income in millions of dollars | Percentage of railway labor compensation to the total income of U.S. |
| 1916 | 1,647,097 | $1,468,576,000 | $891.61 | 115.7 | $45,400 | 3.23 |
| 1917 | 1,732,876 | 1,739,482,000 | 1,003.81 | 130.2 | 53,900 | 3.23 |
| 1918 | 1,841,575 | 2,613,813,000 | 1,419.34 | 184.1 | 61,000 | 4.28 |
| 1919 | 1,913,422 | 2,843,128,000 | 1,485.89 | 192.8 | 66,000 | 4.31 |
| 1920 | 2,022,832 | 3,681,801,000 | 1,820.12 | 236.1 | 72,000 | 5.11 |
| 1921 | 1,660,617 | 2,765,236,000 | 1,665.19 | 216.0 | 55,000 | 5.03 |
| 1922 | 1,579,000 | 2,634,717,000 | 1,668.60 | 216.5 | 59,000 | 4.47 |
- ↑ Average wage of employees in fiscal years 1912–14 is taken as the base = 100.
The above table begins with 1916, for it was then that the Interstate Commerce Commission first began to report for calendar years. The estimates of the national income are those of the National Bureau of Economic Research for 1916-19, both inclusive. Those estimates underwent the scrutiny of, and were approved by, the board of directors of the Bureau. The estimates for 1920–22, both years inclusive, are by myself.
Estimates of the national income can not be made with precision. The National Bureau of Economic Research has expressed the opinion that the margin of error in its estimates “is probably less than 10 per cent.” My own opinion is that they are within 5 per cent. My estimates for 1920–22 are made by rougher methods than those of the Bureau, but I believe them to be inside of 10 per cent, plus or minus. In considering the
