RDP 8303: The 1930's and the 1980's: Some Facts The 1930s and the 1980s: Data Sources

The data used in the graphs are grouped appropriately at the rear. The sources of data shown in the graphs and tables in the main paper are presented below in the same order as they appear in the paper. The catalogue number of each ABS publication is given in the first reference made to it; only the name is given thereafter. Where raw data, or basic manipulations of raw data, have been shown, the source is simply listed (with comments where appropriate). Where the computations are more involved, details are given. Where possible, an attempt has been made to point the reader to some discussion of the quality of the data and alternative sources. All projections for 1983/84 to 1985/86 are from Scenario A of National Economic Summit Conference (1983).

Graph 1 – Unemployment Rate

The most comprehensive source of data on the labour market for the first sixty years of the twentieth century, is Keating (1973). Keating used information from population censuses as bench-marks and interpolated between them using data from various other sources, to arrive at estimates of the workforce. One approach was to estimate employment directly in the various sectors, and then aggregate. The other approach was to use the census data to derive bench-mark estimates of labour force participation rates and then estimate population and participation rates in intervening years to arrive at estimates of the total workforce.

Estimates of unemployment, based on returns supplied to the Bureau of Census and Statistics by Trade Unions, were then incorporated to give an estimate of the total workforce and the unemployment rate. The series in graph 1 uses Keating's data from the first of the two approaches above, as presented and updated in Butlin (1977).^[1]

The procedures involve a number of assumptions and a good deal of estimation. Among the problems, as pointed out Dy Keating (1973) and Forster (1965):

• the exclusion of certain industries from the Trade union unemployment estimates;
• the under-representation of women and part-time employees, and the exclusion of young people who had never wurked, from the Trade Union estimates.

From 1966/67 to 1981/82, the data are from The Labour Force (Annual and Monthly), ABS Cat. No's 6203 and 6204 respectively.

Graph 2 – Share Prices

The series for Australia in the 1920's and 1930's is the Commercial and Industrial Index compiled by the Sydney Stock Exchange. This series does not cover financial and mining shares but is often used to extend the Sydney All Ordinaries Index pack to 1875. The New York series is series Pl presented in Alfred Cowles and Associates (1938) PP66–7. It is an antecedent of the Standard and Poors Index which has been used for the 1970's'and 1980's. The latter was taken from IMF International Financial Statistics. The Australian data for this period is the Australian All Ordinaries Index from Reserve Bank of Australia Bulletin, Table H.2.

The points shown are June averages but the peaks and troughs in the late 1920's early 1930's and 1980's are also shown.

Graph 3 – Growth in Real GDP

The primary source for the expenditure-based estimates of gross domestic product is Butlin (1962). Butlin (1977) made minor adjustments to these figures. These later estimates have been used as the source of all National Accounts data for the 1930's. They are, as far as possible, consistent with the basis on which the National Accounts are presently prepared. It should be noted that there are at least two other sources of estimates of GDP in the 1930's, viz. Clark and Crawford (1938) and Clark (1951).

For the 1980's the source is the National Accounts (Quarterly Estimates of National Income and Expenditure, ABS Cat: No. 5206).

Graph 4 – Prices and Wages

For the 1930's, estimates of average weekly earnings and prices are available in Butlin (1977). Once again, there is some question as to the reliability of early data. Butlin's estimates are based on earnings for male factory workers in Victoria. These data therefore cannot allow for differences in earnings between States, between manufacturing and other sectors, or between manual and non-manual workers. Alternative data for N.S.W. are available from Campbell (1956). A study of wage rates and unemployment in Australia up to 1930 is found in Pope (1982).

For the 1980's, sources are Consumer Price Index and Average Weekly Earnings, ABS Cat. Nos 6401 and 6302 respectively. Prior to September 1982, earnings of males were adjusted to give an estimate of earnings per person. This series was then spliced to the new, post-September 1982 series for ordinary-time earnings of persons.

Graph 5 – Output, Employment and Productivity

For the 1930's, estimates of real GDP and employment are from Butlin (1977). Productivity is estimated as the ratio of the two. For the 1980s, non-farm GDP is from the National Accounts. Employment is employed persons, from The Labour Force, and per hour productivity is real non-farm GDP is divided by total hours worked by non-farm wage and salary earners. The latter component is discussed in more detail with Graph 6.

Graph 6 – Real Unit Labour Costs

Conceptually, real unit labour costs can be calculated by dividing average earnings by a price index, to get real labour costs, and then dividing this by productivity to arrive at real labour costs per unit of output. For the 1930's, the calculation is done as above, using Butlin's (1977) estimates for average weekly earnings, the GDP deflator, employment and real GDP.

For the 1980's, the calculation is more involved. Payroll taxes, for example, are an additional cost to the employer, and thus must be taken into account. Furthermore, changes in the number of hours worked affect unit labour costs. The series used in this graph is as follows (the sources for components are given in brackets):

The comparison of graphs 4 and 6 in the main paper reveals an interesting characteristic of the 1920's. Early in the decade, earnings were rising strongly, whilst consumer prices were falling. This might be expected to imply a large rise in real unit labour costs. But graph 6 shows real unit labour costs falling in this period. This is because they are calculated using the GDP deflator (rather than consumer prices) which rose during this period, mainly due to a sharp rise in export prices. This illustrates the large impact that the terms of trade had on the economy at the time: a large rise in real wages (in terms of consumption goods) could be sustained because product prices rose sufficiently to allow profitability to be maintained. Another way of expressing this would be to say that the increase in the community's well-being as a result of favourable movements in the terms of trade was mostly taken in the form of higher real wages.

Conversely, in the late 1920's and early 1930's, when there was only a relatively small rise in real wages (shown on graph 4), real unit labour costs soared partly as a result of falling productivity, but mainly as a result of the disastrous deterioration in the terms of trade.

The graph below shows real unit labour costs calculated using both the GDP deflator and consumer prices for the 1930's, and the deflator for gross national expenditure for the 1980's, to deflate average earnings. The difference between the two curves reflects (mainly) these terms of trade effects.

Graph 7 – Commodity Prices

Commodity prices for the 1930's are from Commonwealth Yearbooks. In the case of the prices of wool and wheat, the Yearbooks quote prices in £A for exports. Gold prices are derived by dividing the value of gold exports (in £A) by estimates of the physical quantity exported.

For the 1980's, wool and wheat prices are the average prices (in $A) for Australian exports, from Exports of Major Commodities and Their Principal Markets, ABS Cat. No. 5403. For metals, The Economist Metals Index (in $US) is used, converted to $A at the average US$/$A exchange rate for the year.

Graph 8 – Import Prices

For the 1930's, the implicit price deflator for imports from Butlin (1977) is used. For the 1980's, implicit price deflators from Exports and Imports (Balance of Payments Basis) at Constant Prices, Australia. (ABS Cat. No. 5332).

Graph 9 – Terms of Trade

For the 1930's, both price deflators are from Bambrick (1970) as presented by Butlin (1977). Among the problems inherent in these series is their measurement of imports and exports on an f.o.b. basis, which means that the cost of invisibles is ignored. This problem is not encountered with the data for the 1970's, which are from the National Accounts.

Graph 10 – World Trade

World trade volume for the 1930's is from Lewis (1952), p.107 and p.118. Lewis gives, on p.107, index numbers for the value of trade in primary commodities and figures for the percentage of total trade accounted for by manufactured products. With the assumption that these two components make up the whole of trade, this implies an index number for trade in manufacturers, and for total trade, which can then be deflated by the price deflators on p.118.

For the 1980's, the data are “world exports” in the IMF's International Financial Statistics. For 1982–84, OECD (1983) forecasts of growth rates for trade in the OECD area have been used to extrapolate the index.

Graph 11 – Interest Rates

Until 1925/26, Lamberton (1960) is Butlin's (1977) source for data on the long-term bond rate. From 1926/27, Butlin uses Reserve Bank data for the yield on bonds maturing in ten or more years. It should be noted that the bond market was not highly developed in this period.

The London Bank Rate is taken from The Economist. The Bank Rate is, of course, a short-term rate, and it would be preferable to use a longer-term rate for the comparison in this graph. Unfortunately, no such series is available.

For the 1980's, the source for both the US and Australian bond rates is the Reserve Bank Bulletin.

Graph 12 – Balance of Payments

The data in Butlin (1977) for the current account have a long history. Early estimates for 1904 to 1927/28 were made by Wilson (1931). From 1928/29, estimates were prepared by the Bureau of Census and Statistics (1950). Further work was done by McLean (1968) and Swan (1968), and the result is reported in Butlin (1977).

The series for the change in international reserves used by Butlin (1977) until 1926/27 originated in Butlin, Hall and White (1971), and is based on annual data for the foreign currency holdings of Australian banks converted to £A. Butlin (1977) believes them to be less reliable than the post-1928 data, which was originally from Giblin (1957).

For the 1980's, the exports and imports figures are from Balance of Payments, annual and quarterly releases (ABS Cat. Nos 5302 and 5303). The figures for the change in international reserves are the movements due to transactions in Official Reserve Assets (i.e. gold and foreign exchange and Australia's reserve position in the IMF), also sourced in the above releases. For the purpose of calculating a ratio, GDP at current prices is used, the source for which is National Accounts.

Graph 13 – Debt Servicing Burden

For the 1930's, data on interest payments on government securities domiciled overseas and export receipts in current prices are from Butlin (1977). He derives the data on interest payments from Bureau of Census and Statistics Finance Bulletins.

For the 1980's, data on interest on both public and private securities payable overseas are from balance of payments data available on request from Australian Bureau of Statistics. Exports in current prices are from the National Accounts.

Graph 14 – Exchange Rates

For the 1930's, the data are from Butlin (1977), originally from Butlin et. al. (1971), and are in most cases the end-December observation.

For the 1980's, the data are available in the Reserve Bank Bulletin, and are end-June observations.

Graph 15 – Indices of Competitiveness

For the 1920's and 1930's the series compares unit labour costs in Australia with those in the U.K. adjusted for exchange rates. The £stg/£A exchange rate is that shown in graph 14. The unit labour cost series for the U.K. is the wage share of GDP at factor cost taken from Table 18 of Feinstein (1972), multiplied by the implied GDP deflator from Tables 3 and 5. This source provides data on a calendar year basis so the data have been lagged six months, i.e. the 1922 observation is taken as the observation for 1921/22. The error so introduced is probably small given that the wage share is fairly stable, ranging from 60.3 in 1925 to 63.6 in 1932 and 1933.

The Australian series for unit labour costs over this period is that described in connection with graph 6, without deflation by the GDP deflator.

For the 1970's and the 1980's the seies shows Australia's unit labour costs relative to those of our six major trading partners – U.S.A., U.K., Japan, West Germany, Canada, France – adjusted for a weighted exchange rate. The unit labour cost measure for each country compares the compensation of employees with real GDP, data being taken from the OECD Quarterly National Accounts. The weights given to the six countries in the index were derived from their respective shares in Australian imports in 1979/80. These countries are the source of approximately 60 per cent of Australia's imports.

In both periods, the broken line is the relevant relative unit labour cost measure without the exchange rate adjustment.

Graph 16 – Indices of Tariff Rates

The data for the 1930's are taken from Carmody (1952). These earlier data do not take account of non-tariff barriers to trade which act to protect domestic industries, such as quotas, prohibitions, etc. Some of these were used heavily in 1930 to reduce Australia's imports bill. See Schedvin (1970) chapter VIII for a discussion.

For the 1980's, the nominal rate of assistance to manufacturing is from various Annual Reports of the Industries Assistance Commission.

Neither set of figures takes account of the indirect effects of protection on other industries. The effective rate of protection', which measures not only the protection of an industry's output but also the increased cost of protected inputs, tended to be higher than the nominal rate in the 1970's. No comparable data are available for the 1930's. For further discussion of the concept of the effective rate of protection, see the IAC's 1980/81 Annual Report, pp 57–58.

Graph 17 – Public Sector Outlays

Data for the 1930's are all from Butlin (1977); for the 1980's data are available in Australian National Accounts (Annual), ABS Cat. No. 5204.

Graph 18 – Budget Deficits

In the early part of the century, there were three funds on which the accounts of the Commonwealth and State governments were mainly based: The Consolidated Revenue Funds, the Trust Funds and the Loan Funds. The slashed line in this graph shows the combined Commonweatlh and States deficit on Consolidated Revenue Funds. The solid line shows the sum of this deficit and expenditure from Loan Funds, which, as the name implies, was expenditure for capital works purposes financed entirely by debt.

Funds held by governments in the Trust Funds, which were not for the purpose of meeting expenditure commitments, are ignored in this graph.

For the 1980's, the dashed line is the Commonwealth budget deficit, which is available from Budget Statement No. 6. The solid line is the consolidated deficit of all Commonwealth, State and Local Government Authorities. These are from Table 7 of Government Financial Estimates (ABS Cat. No. 5501); this figure is conceptually similar to the solid line in the left hand panel. The estimate for this line for 1982/83 is the estimate presented by the Statistician in the 1982/83 edition of Government Financial Estimates, adjusted for the extent to which the Commonwealth's deficit exceeded the original estimate.

Graph 19 – Interest Rates

For the 1930's, the interest rate is the yield on long term government securities from Butlin (1977). For the 1980's, the yield used is for 10 year non-rebate bonds, at end-June.

The real rate of interest is defined as:

where r and p are the proportionate rates of interest and inflation respectively.

The price deflator for the 1930's is the linked C-series reported by Butlin (1977) (page 48), and originally from Bambrick (1970). The GDP deflator is used for the 1980's. The rate of inflation is measured year-on-year (which is all there is available) for the 1930's but June quarter on June quarter for the 1980's, since the interest rates are at end-June.

Graph 20 – Growth in M3

Butlin (1977) reports data for M3, compiled from Butlin et.al. (1971), for the 1930's. For the 1980's, the source is the Reserve Bank Bulletin. To arrive at growth rates for real M3, a method of deflation similar to that in Graph 19 is used, using the linked C-series for the 1930's and the GDP deflator for the 1980's.

Table 1

The data are average annual growth rates calculated from Butlin's (1977) estimates of national accounts aggregates for the 1930's and from the National Accounts for the 1980's.

Table 2

For the 1930's, population data are from the Commonwealth Labour Report (various issues) and workforce and employment data are from Keating (1973). For the 1980's population is from Australian Demographic Statistics (ABS Cat. No. 3101). Workforce and employment data are from The Labour Force.

Table 3

For both the 1930's and the 1980's, data on interest on overseas debt is sourced as in Graph 10. Export receipts in current prices for the two periods are sourced in Butlin (1977) and National Accounts respectively.

Footnote

Butlin makes minor adjustments to Keating's data for overseas residents. [1]