previous
next

I. The Statistical Situation

4.6. Any description of the taxes actually paid by some millions of taxpayers must consist of broad generalisations. Since income is the magnitude to which most of us first look in comparing the position of individuals and families, it would be convenient to be able to classify individuals into income groups and present for each group the average of all taxes paid by members of that group. Then from further information about the family situation and other particulars of individuals, it would be desirable to go on to present a summary picture of the average taxes paid by family groups of different size and composition, again classified by income. It is lamentable that estimates of the former kind cannot be made from the data available in Australia, and that the only source from which estimates of the latter kind can be developed is largely out of date and deficient in other ways. The extensive sample survey of family income and expenditure now being conducted by the Australian Bureau of Statistics may begin, in a year or two, to give results on which greater confidence can be placed. Meanwhile the Committee can only argue from ‘best guesses’.

4.7. A large array of statistics is inevitably generated in the administration of income tax, and in the annual volume Taxation Statistics a selection of the data is presented to the Australian Parliament by the Commissioner of Taxation, analysed in a variety of informative ways. These are, perhaps, the figures most inquirers turn to when considering the distribution of taxation, if only because they are virtually the only official ones available. Table 4.A, which presents a summary regrouping of data from the latest Taxation Statistics, reveals something of the distribution of individual net income as defined for statistical purposes—broadly, income less expenses of deriving income—and of average personal income tax paid by individuals in the income ranges indicated. But such information hardly serves as a basis for a general discussion of taxation policy. For example, it includes in the lower income ranges many persons who were unemployed for part of the year and omits many others who were not required to lodge returns; it excludes some income from trusts; it also excludes the undistributed income of companies and the bulk of social service grants; and of course such light as it throws on the distribution of taxation is confined to personal income tax, only about 40 per cent of total taxation. To use the Commissioner's figures in the present analysis would therefore be entirely misleading.




  ― 25 ―

TABLE 4.A: INCOME TAX: INDIVIDUALS, 1971–72 INCOME YEAR

                 
Number of taxpayers  
Grade of net income   Male   Female   Total   Net income   Net tax   Average rate of tax on net income  
’000  per cent  ’000  per cent  ’000  per cent  $m  per cent  $m  per cent  per cent 
0–2,999  849  23.4  1,418  68.7  2,268  39.8  4,124  18.1  338  9.4  8.2 
3,000–5,999  1,980  54.6  553  26.8  2,533  44.5  10,860  47.6  1,459  40.7  13.4 
6,000–9,999  642  17.7  75  3.6  716  12.6  5,255  23.0  970  27.1  18.5 
10,000+  155  4.3  20  1.0  175  3.1  2,589  11.3  817  22.8  31.5 
Total  3,626  100.0  2,065  100.0  5,691  100.0  22,827  100.0  3,584  100.0  15.7 
note  

4.8. For a fuller picture of tax distribution it is necessary to rely upon estimates originally based on the material contained in the Australian Survey of Consumer Expenditures and Finances, which was conducted in 1966–68 by Professors Drane, Edwards and Gates. This ‘Macquarie survey’ is a mine of information for students of social affairs and contains vital tax data unavailable elsewhere. Though it suffers from statistical defects fully acknowledged by its authors, is now somewhat dated and will be superseded before long, the Committee has found it, and the work done upon it on the Committee's behalf, indispensible.note

4.9. Figures in the Macquarie survey were collected on a household basis: indeed, with so much expenditure being shared, they could only have been collected on that basis if the results were to be meaningful. Taxes paid in cash are readily identifiable, but not taxes on goods and services paid indirectly: the latter have had to be estimated from expenditure information. To complete the picture company incomes, and the taxes on these incomes, have had to be assigned to particular income groups; so too have estate and gift duties. Thus in such an exercise a great many assumptions of a largely arbitrary kind have had to be made. No one, least of all the statisticians making them, would say otherwise. But if sets of alternative assumptions are made, each in its way reasonably plausible, it is at least possible to see how far the overall results differ and, when they mostly reveal much the same pattern, to attach some credence to the broad picture (if not the exact set of figures) that emerges. Over a hundred alternatives have been explored on the Committee's behalf and the patterns revealed are not in fact widely spread. One resulting estimate is shown in Table 4.B.

4.10. Though Table 4.B has to be interpreted cautiously, especially as regards the details of particular taxes, it suggests an overall distribution of taxation rather different from what most people would probably expect. Taxation, it seems, is quite sharply regressive at the lower end of the household income scale, nearly proportional in a wide middle band, and progressive only at higher levels. Indirect taxes are chiefly responsible for the large difference between this sequence and the progressivity apparently displayed in Table 4.A. (The surprising figure for company income tax paid by households with the smallest incomes is to be explained by the disproportionate number of elderly persons on low incomes living off investments.)




  ― 26 ―

TABLE 4.B: TAXES AS A PROPORTION OF INCOME, 1966–67

                           
Australian Government taxes  
Income range(a)  Personal income tax   Company income tax   Estate and gift duties   Indirect taxes   Total   State government taxes   Local government taxes   Total taxes  
$ per annum  per cent  per cent  per cent  per cent  per cent  per cent  per cent  per cent 
0–1,449  11.5  18.9  30.4  5.6  4.9  40.9 
1,450–2,899  2.9  7.2  13.0  23.1  4.1  2.8  30.0 
2,900–4,349  7.8  3.4  11.6  22.7  3.8  2.1  28.6 
4,350–5,799  9.4  1.9  9.9  21.2  3.3  1.7  26.3 
5,800–7,249  10.3  2.9  9.0  22.1  3.0  1.5  26.7 
7,250–8,699  11.1  2.5  8.6  22.2  2.9  1.4  26.5 
8,700–10,149  11.9  1.7  7.5  21.1  3.0  1.2  25.3 
10,150–13,049  13.2  3.1  7.4  23.7  3.3  1.2  28.2 
13,050–17,399  15.8  9.5  0.9  8.0  34.2  5.4  1.3  40.9 
17,400+  17.0  13.3  3.3  6.8  40.4  7.0  1.0  48.4 
note note  

4.11. However, as an account of the apparent impact of public finance upon households, these estimates represent only one side of the picture. The impact of transfers received from the government sector by way of age pensions, child endowment, sickness benefits and the like is ignored. Not surprisingly, these cash grants are heavily concentrated upon those families with small original incomes, and their inclusion substantially alters the overall result. The necessary calculations can be made in only a very rough and ready way, but some results are shown in Table 4.C. The importance of including transfers is apparent. When taxes and grants are considered together, progressivity is restored at the lower extremity of income distribution, but the system remains proportional in the middle range—a point of some importance in the Committee's argument. Despite the shortcomings of these estimates and the conceptual problems involved in constructing and interpreting ‘tax burden’ tables, the Committee is prepared to accept the validity of this conclusion as a broad statement of the apparent impact of the current tax-transfer system.

TABLE 4.C: DISTRIBUTION OF TAXES AND TRANSFERS ACROSS RANGE OF NET-OF-TRANSFER INCOMES, 1966–67 (a)

                             
1   2   3   4   5   6   7   8  
Income range   Original income   Transfers received   Total income (2 + 3)  Tax rate   Taxes (5×4)  Net taxes (6 - 3)  Net taxes as fraction of original income (7÷2) 
$ per annum  per cent  per cent 
0–1,449  194  1,417  1,611  41  661  -756  -390 
1,450–2,899  2,195  447  2,642  30  793  346  16 
2,900–4,349  3,679  283  3,962  29  1,149  866  24 
4,350–5,799  5,008  226  5,234  26  1,361  1,135  23 
5,800–7,249  6,412  232  6,644  27  1,794  1,562  24 
7,250–8,699  7,853  228  8,081  27  2,182  1,954  25 
8,700–10,149  9,368  189  9,557  25  2,389  2,200  23 
10,150–11,599  10,778  229  11,007  28  3,082  2,853  26 
11,600–13,049  12,142  207  12,349  28  3,458  3,251  27 
13,050–15,949  14,407  286  14,693  35  5,143  4,857  34 
15,950+  23,947  310  24,257  45  10,916  10,606  44 
note note  




  ― 27 ―

4.12. Even with grants included, the picture is far from complete, since no account is taken of government expenditure on goods and services and of the distribution of the benefits of such expenditure between families. Many of the benefits of direct government spending would, by any reckoning, be very unequally distributed, and at least in certain areas like education rather more heavily concentrated in the bottom half of the income scale. But ultimately there is no way of apportioning the benefits of public spending that is not completely arbitrary, and it seems best therefore to focus only on the tax-transfer operations of the budget.

4.13. It is impossible to assess the practicability, let alone the desirability, of any substantial alteration in the incidence of total taxation and grants without regard to the number of persons and families within each income range. Given that taxation in an advanced economy may have to be of the order of one-third of national income, it is simply not arithmetically feasible to secure a major part of it from the highest income groups if the great bulk of income is earned by those in the middle and lower ranges. Some estimates of family income distribution drawn from the Macquarie survey, and from a valuable pilot survey conducted by the Australian Bureau of Statistics in 1968–69, are shown in Table 4.D. They suggest that the top fifth of families in the second half of the 1960s received about 40 per cent of total income, the bottom fifth something less than 10 per cent, and the remaining three-fifths over 50 per cent—a very substantial bunching. It needs, however, to be borne in mind that high rates of tax at the top require reasonably high rates in the middle as well, if marginal tax rates are never to reach excessive heights: reducing the tax paid on income within the range from $8000 to $10,000, for example, necessarily reduces the tax paid on that slice of income from those whose incomes lie above $10,000. If the treatment lower down the scale is too generous, then the required revenue cannot be obtained without pushing marginal rates applying on upper incomes to extreme levels. It is the joint operation of revenue requirements and the constraint that marginal rates of tax be reasonable that compels the weight of tax in lower and middle income ranges to be significant.

TABLE 4.D: DISTRIBUTION OF ANNUAL INCOME, 1966–67 AND 1968–69 SURVEYS (Cumulative percentages)

                         
1966–67(a)  1968–69(b) 
Percentage of surveyed families   Percentage of income excluding transfers   Percentage of income including transfers   Percentage of income including transfers and imputed rent   Percentage of income  
10  0.1  2.4  2.9  2.2 
20  2.9  6.4  7.3  6.8 
30  9.2  12.8  13.8  12.8 
40  16.6  20.1  21.2  19.7 
50  25.2  28.5  29.6  28.2 
60  35.0  38.0  39.1  37.5 
70  46.1  48.7  49.8  48.1 
80  59.0  61.1  62.1  60.3 
90  74.6  75.9  76.7  75.2 
100  100.0  100.0  100.0  100.0 
note  




  ― 28 ―

4.14. These are annual figures and it is of the greatest importance in the present context to avoid being misled by them. It is tempting with annual statistics to argue as if families in each income group are there all their lives, which of course is not the case. Almost everybody's income starts, on entry to the work force, at a much lower level than it reaches later in working life and falls on retirement. Almost anyone, too, may have years of exceptionally low income during a normal working career because of sickness, unemployment or even extended holidays. Moreover, the figures in Table 4.E, drawn from the Australian Bureau of Statistics survey of 1968–69, reveal a further significant fact: the lifetime pattern of earnings varies greatly according to occupational level, at least in so far as this level is indicated by educational attainment.

TABLE 4.E: AVERAGE ANNUAL INCOME OF FULL-TIME MALE WORKERS, BY EDUCATIONAL ATTAINMENT AND AGE GROUP, 1968–69 PILOT SURVEY ($)

                       
Age group (years)  
Educational attainment   15–24   25–34   35–44   45–54   55 and over   All age groups  
University degree  (a)  6,940  8,910  10,320  8,920  8,170 
Non-degree tertiary  3,430  5,180  6,600  6,360  7,150  5,940 
Technician level  3,200  4,470  5,410  5,620  5,010  4,970 
Trade level  3,270  4,030  4,270  4,120  3,620  3,950 
Matriculation only  2,840  4,460  4,940  5,120  5,650  4,320 
Left school at: 
17  2,640  3,970  4,530  4,360  5,010  3,750 
16  2,290  3,750  4,380  4,630  4,210  3,510 
14 or 15  2,320  3,440  3,810  3,770  3,600  3,420 
note note  

4.15 The implication is clear that were it possible to ascertain the distribution of Australian income in terms not of annual income but of the probable average lifetime income of persons now alive, such a distribution would (after eliminating the effects of general economic progress) be considerably more bunched than any annual figures. The great majority of individuals in the lower intervals of annual distribution (the sick, unemployed, aged) would move into higher income groups; equally, many of those towards the top on the basis of annual statistics (persons at the temporary peak of their careers or old enough to have incomes from their own savings and from inheritances) would move into lower groups. There would be far fewer very rich and far fewer very poor; the bulk of the population would be even more concentrated in the middle ranges.




  ― 29 ―

4.16 It would have been desirable to round off this section with some arithmetical calculations showing the alternative patterns of progressivity in the overall taxation system consistent with a given total revenue. But available information on the distribution of income, even on an annual basis, is unfortunately too dated and too sketchy for such working estimates as the Committee has been able to make to deserve reproduction. A few brief observations of a qualitative nature must suffice:

  • (a) By manipulating rates on existing taxes and widening their bases, by their tighter administration, and by such new levies as a tax on capital gains, it would certainly be possible to raise more revenue from those at the top end of the income and wealth scale, were this considered desirable.
  • (b) The number of persons in those groups is so small that the additional revenue would inevitably be very modest in relation to total revenue requirements.
  • (c) Revenue needs are now so great, and the underlying distribution of income is so concentrated in the middle ranges, that the bulk of revenue must come, in one way or another, from those ranges.
  • (d) Within the middle ranges there is certainly scope for more or less progressivity in marginal and average rates as opposed to the near-proportionality now apparently exhibited, but the degree of variation practicable is much less than is often supposed.

previous
next