Wages Drift and Salary Trends

K.J. Hancock , "Earnings Drift in Australia", Journal of Industrial Relations, VIII, 1966. Reprinted in J. E. Isaac and G. W. Ford (eds.), Australian Labour Economics: Readings, Sun Books, 1967, p. 222.

B. Hansen and G. Rehn , On Wage Drift, in 25 Economic Essays, Ekonomisk Tidskrift , Stockholm, 1956; S. Eskilsson , in A. Smith (ed.), The Labour Market and Inflation, I.I.L.S., Geneva, 1966, p. 104.

The Report of the Committee of Economic Enquiry, Commonwealth of Australia, Canberra, 1965 (here referred to as the Vernon Report), states (p. 129) that gross drift is relevant to the effect of award increases on price levels, while net drift is important for equity between those who do and do not receive over-award payments and also for consideration of average wage costs.

S. Weinberg in The Labour Market and Inflation, p. 227 (see 2.). For a further example of the American viewpoint see M.W. Reder , The Theory of Occupational Wage Differentials, A.E.R, Dec., 1955; reprinted in B. J. McCormick and E. Owen-Smith (eds.), The Labour Market, Harmondsworth (Penguin), 1968, p. 223. Reder refers to the "wage glide" in Sweden, Denmark, Germany and England as an expression of market forces circumventing regulatory maxima.

In general, a widening of the earnings gap is regarded as a sign of (price) inflation. An alternative has been proposed by Professor Lydall who regards it rather as a sign of potential deflation. He agrees that the gap widens when demand rises but does not regard these movements as correlated because he holds that it is not the increase in demand but the resulting rise in output which widens the gap. Providing there is productive capacity more will be paid for more work, and as long as the extra wage cost is less than the extra output the effect will be deflationary. See H.F. Lydall , Inflation and the Earnings Gap, Bull. Oxford Inst. Stats., August, 1958; also reprinted in the Penguin volume quoted under 4. It may be noted that his model is based on Britain and stresses the movement of piecework earnings which are probably less important in Australian industry.

F.W. Gould , "The Dynamics of Wage-Drift", Economic Record, XLIII, 1967.

Both series have some questionable features: e.g., for the M.W.R. it is stated that awards "which relate solely or mainly to salary-earners are excluded", but the Commonwealth Statistician neither defines salaries nor publishes a list of industries/occupations which are excluded (or included). Some of what usually are regarded as salary rates are in fact included, i.e., in the trade, services and communications sub-series; the composition of the last-named series in particular is a minor mystery. See the A.N.Z.A.A.S. paper quoted above under *, note 50, p. III/2 and pp. II/5 to 8. For the Average Earnings Series the assumed ratio of female to male earnings has been kept unchanged at 55 per cent since 1951, although there is evidence that it has declined to approximately 50 per cent since, due to increased employment of women in relatively low-paid or part-time jobs. This would have understated the rise in that series. There is also room for argument about the treatment of earnings of employees not subject to payroll tax. See the A.N.Z.A.A.S. paper quoted above, pp. II/3-5, and, for a description of the series, M. Keating , "Australian Wages, Employment and Average Earnings", Economic Record, XLIII, March, 1967.

The formula for drift is as follows, if E stands for Earnings, R for Award Rates, and for the second period:

The sources quoted do not comment on the significance of the recent decline or other short-term movements in the drift rate. Drift was actually negative in five out of sixteen years (1946/7-1962/3) covered in Table F.33 of the Vernon Report, when the "average" was 1.3 per cent p.a.; and also in five out of the twenty years (1948-1968) shown in Samuelson/Hancock/Wallace, when the average was 1.2 per cent p.a. For these and related technical points see the A.N.Z.A.A.S. paper quoted under * p. I/5-8.

The Vernon Report, for example, lists major points of the first group in Part I but relegates the other causes to a mention in Part II, p. 607. Samuelson/ Hancock/Wallace, p. 843, give a brief mixed list of reasons without comment on their relative significance.

The survey sample was based on establishment rather than employment, which explains why estimated coverage ranged from 75 per cent of factory employment down to 42 per cent of retail trade employment. The award rates for different industry groups, as shown, are from $2.50 to $8 above similar industry rates in the M.W.R., mainly because the survey, unlike the M.W.R., included loadings, such as shift allowances, penalty rates or dirt money, as award pay. Award rates in the 1965 survey included rates set by registered or unregistered industrial agreements. In this and other respects, the coverage was not the same as for a similar survey held in 1960, and the two surveys are not comparable for drift analysis. See K. J. Hancock, pp. 231-2 of reference quoted under reference 1 above.

See J.E. Isaac , "Wage Drift in the Australian Metal Industries ", Economic Record, XLI, June, 1965, portion of which is reprinted in Australian Labour Economics: Readings, and where it is also discussed in the subsequent article by K. J. Hancock, p. 231.

The definition of white-collar worker follows the Vernon Report, Table C.52, p. 520, and detailed in Table 6 here. The other categories are: blue-collar, farm and service (incl. recreation) workers. For a full discussion see H.R. Kahn , "Distinction between Wages and Salaries", Scottish Journal of Political Economy, June, 1956.

The sub-periods reflect leads in salary adjustments (or catching up with previous lags) in the middle fifties and lags in the later sixties. Since then (1969-70) they have been catching up again. If a sufficiently long period is chosen, the choice of base year does not greatly affect the long-term trend, as shown here.

There are obvious problems in using tax statistics for this sort of analysis. The table here uses actual income of taxpayers not subject to provisional tax, which covers most employees and includes only a very small fraction of non-wage income. The methods used lay no pretence to great accuracy but the results should reflect general trends. The main reference for methods (including refinements not used here) and analysis is H.F. Lydall , The Structure of Earnings, Oxford, 1968, and his submission to the Vernon Committee which is published in Economic Record, XLI, Dec., 1965. The more recent figures quoted in Table 9 generally confirm the trends mentioned by Professor Lydall.

See, for example, R. Perlman , Labour Theory, New York, 1969, which lists much of the recent work in this field.

There has also been a world-wide secular shift from "light" to "heavy" factory industries and a widening gap in their relative earnings. See T.S. Papola and V.P. Bharadwaj , Dynamics of Industrial Wage Structure, E.J., March, 1970. In Australia, male employment in the textile, clothing and food industries fell from 23 per cent of total male factory employment in 1953-4 to 16 per cent in 1967-8, and the 1965 Survey of Weekly Earnings shows their average earnings of $55 as well below the factory average of $58.50. The trend should have contributed to an upward shift in average blue-collar earnings, but it made probably little difference for earnings drift; overtime and other above-award payments, according to the 1965 survey, made up 24 per cent of total earnings in textiles and clothing and 22 per cent in food as against an overall factory average of 25 per cent.