2 Australia’s Long-Term Economic Performance
Author: International Monetary Fund
Language:
English
Keywords:
BOOK; government; current account; production function; inflation rate; OECD TFP; OECD estimate; CES production function; employment OECD; Total factor productivity; Productivity; Inflation
Download PDF (2.2 MB)

Abstract

For much of the postwar era, Australia’s economic performance has been relatively disappointing. From the 1960s through the 1980s, per capita growth rates were significantly lower than those of other industrial countries, causing Australia’s income rank to slip steadily from one of the highest in the OECD to merely average levels.1 Based on GDP levels per capita, measured in purchasing power exchange rates, Australia was the third-richest country in the OECD in 1960, after the United States and Switzerland, and its per capita income level was more than 50 percent higher than that of the European OECD average. But, because of its persistently lower growth performance, Australia fell to fifteenth place within the OECD in 1992, surpassed by most industrial countries in Europe. At the same time, unemployment gradually increased, from low levels in the 1960s to OECD average levels in the 1980s.

For much of the postwar era, Australia’s economic performance has been relatively disappointing. From the 1960s through the 1980s, per capita growth rates were significantly lower than those of other industrial countries, causing Australia’s income rank to slip steadily from one of the highest in the OECD to merely average levels.1 Based on GDP levels per capita, measured in purchasing power exchange rates, Australia was the third-richest country in the OECD in 1960, after the United States and Switzerland, and its per capita income level was more than 50 percent higher than that of the European OECD average. But, because of its persistently lower growth performance, Australia fell to fifteenth place within the OECD in 1992, surpassed by most industrial countries in Europe. At the same time, unemployment gradually increased, from low levels in the 1960s to OECD average levels in the 1980s.

This chapter discusses the reasons for the disappointing performance in the past, beginning with an analysis of why economic growth slowed down, especially during the 1970s and 1980s. A production function is estimated, which allows the roles played by the underlying factors—capital, labor, and productivity growth—to be separated and quantified. In each case, it turns out that factor growth slowed from the 1960s through the early 1980s. Capital accumulation decelerated, especially in plant and equipment, as national saving declined and inflation helped channel investment into other areas. Growth in effective labor inputs also slowed, partly because rigid labor market regulations came into increasing conflict with changing economic circumstances, generating higher rates of structural unemployment. Moreover, productivity growth fell sharply, as high tariff levels and extensive product market regulations protected large sectors of the economy from the need to innovate and improve efficiency. According to some estimates, productivity growth was, for a long time, among the lowest in the OECD (Figure 2.1).

Figure 2.1.
Figure 2.1.

Growth, Productivity, and Capital

(Annual average, in percent)

Sources: Australian Bureau of Statistics; IMF, World Economic Outlook; and OECD Historical Statistics, 1960-1989.

As these problems became clear, the Australian authorities responded vigorously. Over the past decade and a half, they have implemented reforms across a wide range of sectors, with one overarching objective in mind: to transform Australia into a more dynamic, export-oriented economy, with both its goods and financial markets integrated into the world economy. Financial deregulation began in 1983, and a comprehensive tariff reduction was introduced in 1988. Starting in the late 1980s and early 1990s, public enterprises were corporatized or privatized, moves were made to decentralize the labor market, and product markets were deregulated. More recently, the product market reforms have been extended under the National Competition Policy, which has thrown previously sheltered sectors open to competition.

This bold response is discussed in subsequent chapters, which focus on efforts to improve national saving and the current account performance (Chapters 3 and 4), maintain low inflation (Chapter 5), and reform the labor and product markets (Chapters 6 and 7). Dividends from these reform efforts, which are already apparent in the impressive economic performance in the 1990s, are discussed in Chapter 8.

Why Has Australia’s Growth Performance Been Disappointing?

To attribute Australia’s relatively poor growth performance to the underlying factors—labor, capital, and productivity—it is useful to adopt a production function framework. Most commonly in this context, a Cobb-Douglas framework is used, with factor shares either estimated empirically or derived from the national accounts shares of capital and labor income. The Cobb-Douglas production function, however, implicitly assumes that the elasticity of substitution between capital and labor is one, and the estimate of factor shares in production is either assumed to be constant (if estimated over the entire time horizon) or taken from national income accounts and assumed to be equal to the share in income.2 These assumptions may not always be realistic. When estimated freely, the substitution elasticity has for many countries been found to be about one-half, and factor shares in national income tend to fluctuate owing to changes in the tax structure or regulated interest rates, rather than changes in marginal returns of factors.3 Therefore, a constant-elasticity-of-substitution (CES) production function is used where the elasticity of substitution between capital and labor can be estimated freely from the data and where income shares are allowed to vary endogenously.

The CES production function used is specified as follows:

Y(t)=A(t)γ[δL(t)−ρ+(1−δ)K(t)−ρ]−γρ(1)

where: Y(t) = output in quarter t

L(t) = labor inputs, given by aggregate hours worked per quarter

K(t) = capital inputs, given by the aggregate net capital stock

A(t) = a term reflecting technological progress

In this framework, γ determines the growth rate of technological progress; ρ is the substitution parameter; δ and (1–δ) are the distribution parameters for labor and capital, respectively; and r determines the degree of returns to scale. The elasticity of substitution of capital and labor, σ, is given by σ = 1/(1 + ρ); ρ = 0 would yield the Cobb-Douglas case of a substitution elasticity of unity; and ρ = ∞ (ρ = 0) would yield the Leontief case where capital and labor must be used in fixed proportions. The distribution parameter measures the degree to which technology is labor intensive and also serves as an indicator of labor’s share of total income; if σ = 1 and factors are paid their marginal products, then this share is equal to δ. The returns to scale are constant if r = 1, in which case the production function is linearly homogenous. Returns to scale are increasing if r > 1 and decreasing if r < 1.4 Technological progress A(t) is specified as a linear trend term l+(t/T), where t = 1..T, and T is equal to the number of observations.5 This formulation allows for a gradually rising level of productivity, with its average growth rate over the entire estimation period given by γ Productivity growth over individual cycles or subperiods is derived as the rate of output growth minus the growth rate of effective factor inputs (the expression in square brackets in equation (1)). Alternatively, the entire production function could be estimated over individual subperiods, but this would unduly lower the degrees of freedom.

For estimation, equation (1) is expressed in logarithms to yield:

In(Y(t))=β1In(1+tT)+β4In[β2L(t)β3+(1-β2)K(t)β3]+ut(2)

The coefficients reflect the underlying parameters as follows. β1 represents the growth rate of technological progress (β1 = γ); β2, the distribution parameter (β2 = δ); β3, the substitution parameter (β3 = –ρ), with the elasticity of substitution given by σ = 1/(1 – β3). The returns to scale can be derived from r = β3β4.

The estimation period covers the past 30 years (1967:Q1–1996:Q4) and is restricted to the market sector. It thereby excludes those sectors of the economy for which productivity is measured inadequately because output is given by labor inputs.6

The results for the coefficient estimates and associated standard errors are shown in Table 2.1. All coefficient estimates are statistically significant at the 5 percent level. The estimate of β1 reveals an average growth rate of technological progress of roughly 0.3 percent per quarter or about 1.1 percent per annum.7 The estimate of β3 implies an elasticity of substitution of about 0.5, indicating that a CES specification is indeed more appropriate than a Cobb-Douglas production function. The returns to scale are given by r = β3·β4 = 0.94, implying that the production function displays roughly constant returns to scale.

Table 2.1.

Estimates of CES Production Function: 1967:Q1–1996:Q4

(Estimation by nonlinear least squares)1

article image
Source: IMF staff estimates.

Since the equation is nonlinear in L and K, it cannot be estimated by ordinary least squares. Instead, a procedure for estimating nonlinear econometric models with deterministically trended variables was used. This estimation technique is based on Andrews and McDermott (1995).

The numeric gradient was used to compute the standard errors. Standard error of the model is 0.08.

As far as the growth pattern over time is concerned, the output path shows a sharp decline in growth rates from more than 5½ percent per annum in the late 1960s and early 1970s to about 2½ percent in the late 1970s and 1980s (Table 2.2). At the same time, total factor productivity (TFP) growth declined from more than 2½ percent in the late 1960s and early 1970s to less than 1 percent on average for the 1980s—one of the poorest performances in the OECD (Box 2.1).8 In fact, the results show that more than half of the slowdown in output growth from the late 1960s and early 1970s through the 1980s was attributable to lower TFP growth.9

Table 2.2.

Growth Rates of Potential Output, Inputs, and Total Factor Productivity (TFP) for the Market Sector: Estimates from a CES Production Function

(Annual averages of selected time periods, in percent)

article image
Sources: Australian Bureau of Statistics, Australian National Accounts Multifactor Productivity 1995/96, July 1997; and IMF staff estimates.

The time periods approximate the growth cycles, from peak to peak, identified in Australian Bureau of Statistics (1997). The analysis above is based on calendar years, except for the data from the Australian Bureau of Statistics, which are for the year ending June 30. The estimation framework is explained in the text. The Australian Bureau of Statistics estimates are based on factor shares in the national accounts. Figures may not always add up due to rounding.

International Comparisons of Productivity Growth and Levels

A number of studies have shown that Australia’s productivity performance was for many years significantly below that of other countries. According to a report by the Australian Economic Planning Advisory Council (EPAC, 1996), total factor productivity growth in Australia was the lowest in a sample of 14 industrial countries during 1970–89. OECD estimates rate Australia the fourth lowest over a similar time period (see table below). For the period of 1970 to 1989, average total factor productivity growth in Australia was estimated at about ½–¾ percent per annum in these studies, compared with the OECD average of 1½ percent.

Economic convergence may have played some role in these results. Given that Australia had the third highest level of per capita GDP in the OECD in 1960, technological catch-up may partly explain why other OECD countries recorded higher total factor productivity and per capita GDP growth rates in subsequent years. Nevertheless, many of the poorer OECD countries not only caught up to Australia’s per capita GDP level but surpassed it, as Australia’s ranking fell to fifteenth place in the OECD by 1992.

Australia’s relatively low total factor productivity growth can be traced largely to its poor performance in labor productivity. The EPAC study found that from 1970–89, labor productivity growth in Australia was only 1½ percent compared with 2¾ percent on average in the OECD—the second lowest in a sample of 14 countries. As a result, Australia’s labor productivity level slipped from fourth to twelfth in a sample of 13 industrial countries, according to a recent OECD study of manufacturing labor productivity covering 1960 to 1995 (Pilat, 1996).

Capital productivity growth has also been somewhat lower in Australia than in other OECD countries. This partly reflects the structure of Australia’s economy: since capital intensive sectors, particularly mining, play a relatively large role, the capital/output ratio is high and capital productivity levels tend to be lower.

Growth of Labor, Capital, and Total Factor Productivity (TFP) in Selected OECD Countries

(In percent per year, averages of 1970–89, OECD: 1979–94)

article image
Sources: EPAC (1996); and OECD (1996).

The estimates also show that the decline of economic performance has first been halted and now reversed. After a period of volatility in growth during the 1980s, potential output has accelerated more steadily in the 1990s (Figure 2.2). Furthermore, productivity growth has accelerated in the latest cycle (1989–96) with average total factor productivity growth reaching 1.3 percent per annum, ½ percentage point above the level during 1982–88.10

Figure 2.2.
Figure 2.2.

Potential Output and Total Factor Productivity: Estimates from a CES Production Function

(In percent per annum)

Source: Australian Bureau of Statistics; and IMF staff estimates.

Although the production function suggests that the slowdown of output growth was chiefly attributable to lower productivity growth, slow labor and capital input growth also played important roles.

Growth in effective labor inputs stalled during 1974–81 after having grown by an average of more than 1½ percent per annum from 1967–73, mainly because growth in full-time employment virtually ceased (Table 2.3). Employment growth accelerated in 1982–88, expanding by more than 2 percent per annum on average, largely in the form of part-time employment. In the latest economic cycle, however, overall employment growth has again been subdued, with effective labor inputs in the market sector growing by only ¾ percent per annum in 1989–96. Meanwhile, over the entire period, the number of hours worked per employee has declined steadily, caused by a rising share of part-time employment, which offset a small increase in average hours worked by full-time employees.

Table 2.3.

Developments in Main Labor Market Indicators

(Annual average growth rates of selected time periods, in percent, unless otherwise indicated)

article image
Sources: Australian Bureau of Statistics, Australian National Accounts Multifactor Productivity, 1995/96, July 1997; and OECD, Historical Statistics.

Percentage share of part-time employment in total employment.

Average weekly hours of all persons employed.

The growth in labor input refers to the market sector only and is therefore not strictly comparable to the other figures in the table, which refer to the overall economy.

Growth in capital inputs also slowed over the past three decades, falling from 4½ percent per annum in 1967–73 to about 2¼ percent in 1989–96. This slowdown was roughly in line with the slowdown in output growth, leaving the capital intensity broadly unchanged.11 But because of rapid population growth, the growth of real investment per capita declined significantly from the late 1960s through the 1980s and 1990s, to levels significantly below the OECD average (Table 2.4).12

Table 2.4.

Growth Rates of GDP and Investment Per Capita

(Annual averages of selected time periods, in percent)

article image
Sources: Australian Bureau of Statistics, National Accounts; IMF, World Economic Outlook database; and OECD, Historical Statistics.

Why Has Capital and Labor Input Growth and Productivity Performance Been Poor?

Several interrelated factors contributed to slowing the growth of capital and labor inputs and total factor productivity. Prominent among these were the decline in national saving; the increase in inflation; and the persistence of the centralized labor market institutions, high tariff levels, and extensive product market regulation.

Reasons for the Slowdown in Capital Accumulation

For most of the postwar period, the high level of investment required to develop Australia’s natural resources and meet the physical infrastructure needs of its increasing population was financed primarily from domestic savings. Recourse to foreign savings was relatively small, with the current account deficit averaging about 2½ percent of GDP in the 1960s and 1970s. In the early 1980s, national saving fell sharply, and the current account deficit widened to about 4½ percent of GDP. Since then, the deficit has remained around this level, fluctuating between 3 and 6 percent of GDP, in line with the economic cycle.

In past episodes where the current account deficit has risen rapidly, investors have tended to impose a risk premium on the returns they require on their Australian investments, resulting in higher domestic interest rates, a lower exchange rate, or some combination of the two, coupled with increased volatility. This pattern was most apparent in 1985 and 1986 as the financial markets responded to concerns about the sustainability of macroeconomic policy settings in an environment of a rapid widening of the current account deficit to more than 6 percent of GDP. The exchange rate fell sharply (by almost 35 percent on a trade-weighted basis), although this was also partly due to a significant fall in Australia’s terms of trade. In response, short-term interest rates increased from about 13 percent at the end of 1984 to almost 20 percent by the end of 1985, putting downward pressure on investment and growth. This experience prompted the authorities to take measures to boost national savings and address factors that may have been distorting investment, so as to reduce the current account deficit and facilitate growth.

Investment was also adversely affected by inflation. In the 1970s, Australia’s inflation rate rose more sharply than in other OECD countries, and in the 1980s it came down more slowly. The interaction of high inflation with the tax system reduced the real return on capital, particularly plant and equipment, and created an incentive for investment in property (see Box 2.2). This was most apparent in the late 1980s when plant and equipment investment growth slowed markedly while investment in property increased sharply, financed by readily available credit following financial sector deregulation in the early 1980s.

Reasons for the Slowdown in Labor Input Growth

The causes of the slowdown in labor input growth arose primarily from the centralized industrial relations system and structural change. In the 1970s and early 1980s, the industrial relations system gave rise to an aggregate wage problem. Wage agreements in booming sectors were spread to the wider economy through application of the principle of comparative wage justice. This problem was addressed in the mid-1980s when the Australian Council of Trade Unions and the ruling Australian Labor Party reached an accord that moderated wage growth and contributed to strong growth in employment.

In the late 1980s, a relative wage problem emerged. As tariff barriers fell, the pattern of Australian production began to shift from manufacturing to services, and from certain manufacturing sectors, such as textiles, clothing, and footwear, to others, such as machinery and equipment. These shifts, together with changing technology, demanded changes in wages and work practices. This need for change, however, conflicted with the rigid system of centralized industrial agreements, and as a result, both growth in labor inputs and productivity suffered.

Reasons for the Slowdown in Total Factor Productivity Growth

Some of the most important factors behind Australia’s poor productivity performance have been its limited degree of openness, because of high tariff barriers and tight domestic regulation of labor and product markets. New growth theories also provide a possible further explanation for the slowdown in productivity growth.

Protected Markets

Tariff barriers have been detrimental to productivity by sheltering the economy from pressure to adapt to world-best practices and most efficient means of production. Domestic regulation has limited competition in labor and product markets and contributed to a suboptimal allocation of resources in the economy, thereby lowering productivity and growth. Since these trends have persisted over long time periods, Australia has fallen behind other industrial countries in terms of productivity levels and GDP per capita. The role of labor market regulation in the trend decline in productivity growth is discussed further in Chapter 6, while the role of tariff barriers and product market regulation in the productivity slowdown is discussed in Chapter 7.

The Impact of Inflation and the Tax System on Investment in Australia

Some of the most widely recognized costs of inflation in Australia are the distortions arising from the interaction of inflation with the tax and accounting systems. These distortions changed the relative rates of return on different types of capital and changed the incentives for different forms of financing of capital. (For a fuller discussion of inflation and corporate taxation in Australia, see Ryan, 1990.)

First, until 1986, the absence of a capital gains tax in Australia created a bias toward investments in projects, particularly property, that yielded returns as capital gains rather than income. While this distortion has since been removed (with the introduction of a real capital gains tax in 1986) a bias remains in favor of investment in owner-occupied housing, as this activity is still not subject to capital gains or imputed-rental income taxes.

Second, during high inflation periods, the first-in-first-out accounting method used by Australian companies increases the effective tax on profits of firms with significant investment in inventories. Under this method, whenever inflation causes the sale price of goods to exceed the book value, firms are assessed for tax purposes as making increased profits even though this is purely due to inflation.

Third, companies’ ability to deduct nominal interest payments (the inflation premium component that is akin to the repayment of real capital) from their taxes lowers the cost of debt financing compared with equity financing. Debt financing was also favored by the double taxation of dividends before the introduction in 1987 of dividend tax imputation. (The double taxation of dividends arose because dividends were paid from after-tax corporate income and the investor was subject to tax on dividends received. The imputation of tax already paid by the corporate removed this distortion.)

These distortions, combined with the increased availability of credit after financial sector deregulation in the early 1980s, created an incentive for highly leveraged investment in assets such as property and stocks rather than in plant and equipment (Macfarlane, 1989 and 1990). This contributed to the asset price boom in the late 1980s.

Inflation also affected investment by distorting relative price signals. Parkinson (1991) finds a significant positive relationship between inflation and uncertainty in Australian data (where uncertainty is proxied by the variance of unanticipated inflation). This supported a similar finding reached by Pagan, Hall, and Trivedi (1983).

Only limited research work has been done to quantify the costs of these distortions, largely because of the difficulty inherent in such research. Nevertheless, some tentative evidence suggests that inflation hurt business investment. McTaggart (1992) shows that if inflation had been constant at zero, investment in plant and equipment would have been, on average, around 14 percent higher from 1963–91. These results need to be interpreted with caution, however, because they were obtained from a partial analysis of investment, not from a general equilibrium model.

The Implications of New Growth Theories for Australia

New growth theories can also provide some insight into the reasons behind the slowdown in productivity growth. Two studies that have received much attention argue that there are strong links between growth and two specific types of investment, both of which declined markedly in Australia in the 1970s and 1980s: plant and equipment investment (De Long and Summers, 1992), and public infrastructure investment (Aschauer, 1989).

De Long and Summers (1992) hypothesize that plant and equipment investment creates important learning externalities because operating personnel learn to improve production equipment and trigger a self-propelled increase in the efficiency of production. Consequently, the low level of plant and equipment investment in Australia, particularly in the 1980s, may have had a disproportionately large effect on the slowdown in productivity and growth.

Private sector growth may also have been strongly affected by a decline of investment in public infrastructure. Using United States data, Aschauer (1989) finds a complementarity between public infrastructure investment and private sector growth, and attributes three-quarters of the decline in private sector growth in the United States since 1970 to the decline in the rate of growth of public infrastructure.13 In Australia, the average growth rate of the public capital stock halved from 4 percent in the 1970s to 2 percent in the 1980s,14 and two Australian studies confirm the link of this investment with slower private sector growth (see Otto and Voss, 1994; and Nhu, 1993).

A further argument at the core of the endogenous growth theory is that investment in research and development (R&D) has a strong impact on private sector growth. In Australia, expenditure on R&D has traditionally been relatively low. With R&D spending of only 1¼ percent of GDP, Australia ranks sixteenth in the OECD, placing it below the average of almost 2 percent of GDP for the OECD and well below that of the United States, which spends 2¾ percent of GDP on R&D (EPAC, 1994, p. 59).

References

  • Andrews, D., and C. J. McDermott, 1995, “Nonlinear Econometric Models with Deterministically Trending Variables,” Review of Economic Studies, Vol. 62, pp. 343–60.

    • Search Google Scholar
    • Export Citation

  • Aschauer, David Alan, 1989, “Is Public Expenditure Productive?” Journal of Monetary Economics (Netherlands), Vol. 23, pp. 177–200.

    • Search Google Scholar
    • Export Citation

  • Australian Bureau of Statistics, 1997, Australian National Accounts Multifactor Productivity, 1995–96 (Canberra).

  • Commonwealth Treasury of Australia, 1996, Economic Roundup, Autumn (Canberra).

  • De Long, J., and L. Summers, 1992, “Equipment Investment and Economic Growth: How Strong Is the Nexus?” Brookings Paper on Economic Activity: 2, Brookings Institution.

    • Search Google Scholar
    • Export Citation

  • Economic Planning Advisory Council (EPAC), 1994, “Investment for Growth,” EPAC Background Paper No. 39 (Canberra).

  • Economic Planning Advisory Council (EPAC), 1996, “Tariff Reform and Economic Growth,” Economic Planning Advisory Commission Paper No. 10 (Canberra).

    • Search Google Scholar
    • Export Citation

  • Macfarlane, I. J., 1989, “Money, Credit and the Demand for Debt,” Reserve Bank of Australia Bulletin (May).

  • Macfarlane, I. J., 1990, “Credit and Debt: Part II,” Reserve Bank of Australia Bulletin (May).

  • McTaggart, D., 1992, “Cost of Inflation in Australia,” in proceedings of Reserve Bank of Australia conference on Inflation, Disinflation, and Monetary Policy (Sydney: Reserve Bank of Australia).

    • Search Google Scholar
    • Export Citation

  • Nhu, C., 1993, “Impact of Government Expenditures on Economic Performance and Growth in Australia” (unpublished; Canberra: Australian National University).

    • Search Google Scholar
    • Export Citation

  • OECD, Historical Statistics (Paris: OECD, various issues).

  • OECD, 1996, Economic Outlook (Paris: OECD, June).

  • Otto, G., and G. Voss, 1994, “Public Capital and Private Sector Productivity,” Economic Record, pp.121–32.

  • Pagan, A. R., A. D. Hall, and P. K. Trivedi, 1983, “The Effects of Inflation: A Review with Special Reference to Australia,” in The Effects of Inflation: Theoretical Issues and Australian Evidence, ed. by Pagan and Trivedi (Canberra: Centre for Economic Policy Research, Australian National University).

    • Search Google Scholar
    • Export Citation

  • Parkinson, M., 1991, “Inflation and Uncertainty,” Treasury Research Paper No. 3 (Canberra: Commonwealth Treasury of Australia).

  • Pilat, D., 1996, “Labor Productivity Levels in OECD Countries: Estimates for Manufacturing and Selected Service Sectors,” OECD Economics Department Working Papers, No. 169, pp. 1–33.

    • Search Google Scholar
    • Export Citation

  • Ryan, C. J., 1990, “Inflation and Corporate Taxation in Australia,” Reserve Bank of Australia Discussion Paper 9011 (Sydney).

1

In the 1960s, Australian per capita GDP expanded by 3 percent per annum, compared with almost 4 percent on average in the OECD. In the 1970s and 1980s, average per capita growth rates declined to 2 percent per annum in Australia, compared with about 2½ percent on average in the OECD. Australia’s strong overall growth rates have been largely achieved through rapid population growth, with the population rising by about 60 percent from 1960 to 1990, an average growth rate of 1.6 percent per annum. This was the highest population growth rate of any single industrial country in the OECD and well above the OECD average of 0.9 percent.

2

The elasticity of substitution is the percentage change in the capital/labor ratio, in response to a 1 percent change in the factor price (interest/wage rate) ratio.

3

In Australia, factor income shares have fluctuated significantly in the past. In the market sector, the share of labor income, which averaged 67 percent from 1967–68 to 1995–96, reached a high of 73 percent in. 1975–76 and a low of 63 percent in 1995–96; see Australian Bureau of Statistics (1997).

4

This can be seen by multiplying both inputs by a constant, which yields Y(λK,λL)=λrY.

5

When logarithms are introduced, ln(A) = [0..ln(2)].

6

Six subsectors are excluded: government administration and defense; finance and insurance; property and business services; education; health and community services; and part of personal and other services. The data are taken from Australian Bureau of Statistics (ABS, 1997), with quarterly data interpolated from the annual data. As the objective is to estimate the underlying potential output that relates to the normal utilization of capital and labor, the factor inputs (K and L) have been smoothed by a moving average to reduce the short-term fluctuations that are likely to reflect cyclical variations.

7

This derivation follows from the fact that 1n(l+t/T) ≈ 0.7*t/T in equation (2); therefore the coefficient β1 is multiplied by 0.7/T (where T = 120) to yield a value of 0.27 percent per quarter or 1.1 percent per annum. The estimate of β2 (i.e., δ) is difficult to interpret. As noted above, if r = 1 (the Cobb-Douglas case), then β2 would reflect the share of wage income in national income. In the general CES case, however, the wage share also depends on ρ as well as the level of output and labor at a given point in time. If factors are paid their marginal products, the share of wage income wL/Y, for example, would be given by wL/Y = ∂Y/∂L·L/Y = δ(Y/L)ρ. Moreover, since the underlying data are index numbers rather than absolute values, the shares of capital and labor income cannot readily be obtained.

8

Total factor productivity growth, the difference between output and input growth, measures the amount of growth that can be attributed to more efficient use of capital and labor inputs rather than a higher level of inputs. Measuring productivity growth is difficult because of measurement problems surrounding inputs—especially capital inputs—in production. Furthermore, the choice of time periods influences the results of average productivity growth significantly, especially in international comparisons, as productivity growth is strongly procyclical and cycles do not match exactly across countries.

9

From 1967–73 to 1982–88, average output growth declined by more than 3 percentage points; of this decline, about 1.8 percentage points were attributable to lower productivity growth.

10

Australian Bureau of Statistics (ABS) figures differ somewhat from the estimates presented here because they are based on factor shares in the national income accounts. Still, the only significant difference occurs during 1974–81, where the ABS estimates of total factor productivity growth are almost ½ percentage point higher. This difference is caused by the extraordinarily high share of labor income during the 1970s, where a sharp acceleration of real wage growth pushed the share of labor income to above 70 percent (ABS, 1997, p. 10), a trend that was subsequently reversed by the accord between the Australian Labor Party and the Australian Council of Trade Unions (see Chapter 7), which contained real wage growth and caused the labor share to fall back to its traditional level of about 65 percent. Since labor inputs actually fell during the 1970s, the high labor weight in the national income accounts caused a low estimate of overall input growth and thus a higher estimate of total factor productivity growth by the ABS.

11

Real investment growth in the 1990s is likely to have been slower than shown by the Australian Bureau of Statistics data due to a measurement bias. The bias exists because of the growing importance of investment in computer equipment, the prices of which have fallen significantly relative to other goods. As a result, real investment in 1989/90 prices is likely to be overstated (Commonwealth Treasury of Australia, 1996, page 13). A forthcoming update of the base year for the measurement of real data by the Australian Bureau of Statistics may result in a significant downward revision in real investment (the previous movement from a 1984/85 to a 1989/90 base for constant price measurement resulted in a significant fall in real investment growth). Such a revision would lower capital inputs and imply somewhat faster total factor productivity growth than estimated in this paper.

12

The decline in the growth of investment per capita in the 1990s was largely due to a sharp fall in investment in nonresidential buildings following an investment boom in the 1980s that left a sizable overhang of such capital. This offset strong growth in investment in plant and machinery per capita in the 1990s.

13

However, the result is sensitive to the definition of public infrastructure investment, with a weaker result found if investment in national parks is excluded.

14

This occurred before the bulk of privatization shifted public investment to the private sector in the Australian national accounts.

Within Same Series

Other IMF Content

Other Publishers Content

Cover Australia
Australia
Benefiting from Economic Reforms
Author: International Monetary Fund
ISBN:
9781557757326
Publication Date:
20 Nov 1998
Publisher:
International Monetary Fund
DOI:
https://doi.org/10.5089/9781557757326.071
Pages:
171
  • View in gallery

    Growth, Productivity, and Capital

    (Annual average, in percent)

  • View in gallery

    Potential Output and Total Factor Productivity: Estimates from a CES Production Function

    (In percent per annum)

Growth, Productivity, and Capital

(Annual average, in percent)

Potential Output and Total Factor Productivity: Estimates from a CES Production Function

(In percent per annum)