Dr Glen Otto
Economics, Commerce and Industrial Relations Group
30 June 1997
Dr Glenn Otto is a Senior Lecturer in the School of Economics at
the University of New South Wales. He has degrees from the
University of Queensland, the Australian National University and
Queen's University in Canada. His research interests and
publications range across areas such as business cycles, the
current account, productivity, the economic role of public capital,
and applied econometrics.
Major Issues Summary
Issues and Problems Arising in the Measurement
of Productivity Growth
Standard Measures of Productivity
Productivity and Living Standards
Australian Productivity Growth
Productivity in Australia Relative to the
The Determinants of Productivity
Investment in Physical Capital
Investment in Research And Development
Investment in Human Capital
Microeconomic Reform and Productivity
Improvements in productivity are widely agreed to be the
fundamental cause of long-term improvements in a country's standard
The two most widely used measures of productivity are labour
productivity and multifactor productivity. The latter refers to a
measure of productivity that attempts to account for all inputs
into the production process, not just labour. While both have their
uses, multifactor productivity is typically a better measure of an
economy's level of technology. Thus, the growth rate of multifactor
productivity is an indicator of the underlying technological
progress in an economy.
Since multifactor productivity is obtained as a residual, i.e.
it is the change in output that remains after accounting for the
contribution of all the factors of production, it can be seriously
affected by measurement error. In particular, short-term movements
in multifactor productivity are often related to the state of the
business cycle, rather than reflecting changes in technology. As a
consequence, longer term trends in multifactor productivity are
generally a better guide to underlying technical progress.
One prominent line of economic modelling (the Swan-Solow model
of economic growth) suggests that productivity growth is the
fundamental source of long term growth in per capita incomes. In
the absence of productivity growth the long-run growth rate of
income per capita will fall to zero. In this analysis an increase
in a country's saving (and/ or investment) rate will only produce a
temporary increase in its growth rate.
From a policy perspective, an important limitation of such
analysis is that productivity growth is 'exogenous', i.e. the
economic forces that determine the growth rate of productivity are
left unexplained. Recent developments in growth theory have
attempted to identify the specific factors which determine the rate
of productivity growth in the economy.
Estimates of productivity growth for Australia over the last
thirty years suggest two broad trends:
- The trend rate of productivity growth was lower in the 1980s
than in the 1960s. This is particularly evident in measures of
- The average rate of productivity growth for Australia has
tended to be slower than for other OECD countries, especially
during the 1980s.
Identifying the main determinants of productivity growth is not
straightforward. While a large number of possible influences have
been suggested, the precise economic mechanisms by which many of
the candidates actually affect the rate of productivity growth are
not clearly specified. Essentially, what is required is an
understanding of the incentives which exist for the production of
knowledge and ideas (of both a theoretical and an applied
Given the nature of knowledge, the benefits to society from its
production, e.g. through research and development (R&D) or
education, may exceed the benefits obtained by the firm or
individual who actually produces it. As a consequence, the
production of knowledge may be less than optimal if it is left
entirely to market forces. Thus, there may be a role for
governments to subsidise investment in R&D, education and
training. To the extent that the growth rate of technology is
affected by the amount of resources devoted to knowledge
production, government policy may be able to influence the
economy's long-run rate of growth.
Some of the main factors that seem likely to affect productivity
growth include the extent to which there is scope for a country to
adopt best-practice technology from other countries, the level of
investment in certain types of physical capital (e.g. plant and
equipment and core public infrastructure), the quantity of
resources devoted to research and development, the level of human
capital of the labour force, the macroeconomic environment and the
nature of the social and economic institutions that a country
Over the last three decades Australia's investment in physical
capital has tended to exceed the OECD average. However, this partly
reflects our higher-than-average population growth rate and our
relatively low level of capital productivity. It terms of its
effect on productivity, the form of investment expenditure appears
to be of as much importance as the aggregate level. Some recent
empirical research suggests that investment in plant and equipment
and in certain types of public infrastructure (e.g. transportation
and communication systems) can have very high returns relative to
other types of investment.
Australia's total expenditure on R&D has tended to be below
the OECD average. This partially reflects relatively low levels of
private sector R&D investment in Australia. In particular,
there is a view that the performance of Australian business in the
commercial application of basic research has been poor. This has
been attributed to a number of factors, including the role of
protection in reducing competitive pressures on domestic industry,
the lack of adequate managerial skills, an inadequate exchange of
information between basic researchers and industry and the lack of
venture capital to fund innovative ideas and companies.
Microeconomic reform is one means by which economic institutions
in Australia are being changed. The aim of microeconomic reform is
to reduce unnecessary restrictions on trade (both domestic and
international) and to allow the allocation of resources in the
Australian economy to better reflect the outcomes of markets.
There have been a number of studies that suggest microeconomic
reform will produce substantial benefits for the Australian
economy. One of the beneficial effects that is expected to flow
from the microeconomic reform process is an improvement in
Australia's productivity performance.
While it is reasonable on the basis of economic theory to expect
microeconomic reform to have some positive effect on the
level of productivity in Australia (e.g. due to the
effects of increased competitive pressures), whether the changes
associated with microeconomic reform can lead to a permanent
increase in the growth rate of productivity is much more
uncertain. Certainly, it is difficult to identify the exact
economic mechanisms through which a permanent increase in
productivity growth will occur. This reflects economists'
uncertainty about exactly what are the fundamental causes of
Productivity is defined as the ratio of an index of outputs to
an index of inputs. In the simplest case where there is only one
output (e.g. number of cars produced) and only one input (e.g.
hours of labour) productivity would be measured as the number of
cars produced per hour of labour. In this case we would have a
measure of the productivity of labour. When two or more inputs (or
outputs) are involved in the production process it is necessary to
construct an index of all the outputs and an index of all the
inputs involved in the production process.(1)
An improvement in productivity occurs if either:
- an increased level of output can be produced for a given level
of inputs, or
- a decreased level of inputs are necessary to produce a given
amount of output.(2)
Ideally, the level of productivity is a measure of the current
state of technology and changes in productivity then reflect
underlying technological change in an economy. (3) The level or
state of technology is the currently known (best) methods of
converting a country's resources (i.e. labour, capital, land and
minerals) into those goods and services desired by the economy. (4)
In practice, existing measures of productivity can be affected by
many factors other than changes in technology (see below) and this
can make it difficult to interpret short-run fluctuations in these
It seems reasonably self-evident that improvements in
productivity which reflect underlying technological progress are
desirable phenomena, since they allow more goods and services to be
produced (and consumed) from a given level of scare resources.
While technological change can have adverse effects on the economic
welfare of some individuals and groups in the short term, over the
longer term growth in technology is the fundamental source of
improvements in a country's standard of living. (5)
It is useful to make a distinction between the absolute
level of productivity at a point in time and the
growth rate of productivity over time. For example,
Country A may have a high level of productivity and be able to make
20 cars per hour compared with Country B where only 10 cars per
hour can be produced. Clearly, Country A is more productive than
Country B. However, if Country B has a higher growth rate of
productivity that Country A, then over time Country B's level of
productivity will tend to approach (or catch up) with Country A's
productivity. The greater the difference between the growth rate of
A and B's productivity, the more rapid the rate of convergence.
Because of compounding, small differences in growth rates of
productivity can have significant effects. For example, with a
productivity growth rate of 1.5 per cent per year Australia's
absolute level of productivity (and correspondingly our standard of
living) takes about 50 years to double. If average productivity
growth could be raised by just half a per cent (to 2.0 per cent per
year) then living standards would double in just 36 years. Such
calculations make very clear the importance of being able to
identify the key factors that influence the rate of productivity
There are a number of widely used measures of productivity. The
simplest approach is to compute a single factor measure of
productivity. However, a conceptually better approach is to compute
a multifactor measure of productivity.
Single factor measures of productivity focus on the productivity
of a single input into production. It is common to assume that
output produced (by a firm or industry or economy) depends on two
inputs, labour and capital. (Here, 'capital' refers to physical
capital inputs into production such as factories, offices, tools
and equipment, etc.) In this case, two single factor measures can
be defined: labour productivity and capital productivity.
Labour productivity is defined as output (Y) per worker or
per worker-hour (L). It is calculated as (Y/L). Growth in
labour productivity is calculated as the percentage change in (Y/L)
Capital productivity is defined as output (Y) per unit of
capital (K). It is calculated as (Y/K). Growth in capital
productivity is calculated as the percentage change in (Y/K) over
When labour and capital are treated as the only factors of
production then a standard value-added measure of output such as
constant-price Gross Domestic Product (GDP) is the appropriate
measure of (Y) to use in calculating labour or capital
productivity. However, if there is interest in calculating the
productivity of an intermediate input such as raw materials or
energy, then output should be measured as gross output.(6)
One advantage of single factor measures of productivity is that
they are relatively straightforward to calculate and data
requirements are not particularly heavy, particularly for labour
productivity. However, as measures of technological progress they
will generally be deficient. By focusing on a single input to the
production process they ignore the possibility of substitution of
this input for others in response to relative price changes. For
example, one problem with labour productivity is that it does not
account for the fact that an increase in (Y/L) could result from a
move towards a more capital intensive production process or
increased use of energy, neither of which is necessarily consistent
with better economic performance or reflective of an improvement in
One means of controlling for the effects of substitution among
inputs is to compute multifactor (or total factor) productivity.
The principle underlying multifactor productivity (MFP)
calculations is to divide output by a combination of all the
relevant inputs into the production process. By accounting for the
contribution of all the factors of production on output we are, in
theory, left with contribution due to the current state of
In principle, improvements in the quality of labour and capital
inputs should be accounted for in measures of (L) and (K). However,
quality-adjusted measures of labour and capital are not generally
available from official data sources, e.g. the Australian Bureau of
Statistics (ABS), so measures of MFP based on official data will
tend to capture improvements in the quality of inputs. Work by
Denison (1985) for the United States (US) represents one attempt to
account for improvements in the quality of various inputs. While
accurate measurement of many economic variables is difficult, this
is particularly the case for the capital stock, which has both
conceptual and practical difficulties.(8)
Measurement of output (Y) is difficult for the service sector
since output in many of the service industries either is not well
defined or is difficult to measure accurately. Thus, in the
finance, community service, public administration and defence
sectors output is generally derived from employment data. For some
countries like Australia and the US, the estimate of output is
based on an assumption of zero productivity growth in labour. Other
countries, such as Sweden and the United Kingdom (UK), make
different assumptions about labour productivity growth in the
service sector. Thus, for many calculations of aggregate
productivity the service sector has to be excluded. Given the
tendency for the service sector to account for an increasing share
of GDP, this poses a potentially serious problem for obtaining
accurate economy-wide productivity measures.
Finally, all measures of productivity will be affected in the
short run by changes in the rate of utilisation of the factor
inputs associated with the business cycle. For example, labour
productivity tends to decline as the economy goes into recession.
The usual explanation for the fall is that even though demand for
their product has fallen, firms hold onto (or hoard) skilled
workers who may be expensive to fire and re-hire during the
recovery stage of the business cycle. Analogously, when demand
rises the under-utilised workers will tend to work harder so
measured labour productivity tends to rise sharply as the economy
recovers. The main implication of this is that year-to-year
movements in productivity can be difficult to interpret and do not
necessarily reflect underlying changes in technology.
According to the Australian Bureau of Statistics (1996,
Catalogue No. 5234.0, Page 6), 'In practice ...; (MFP) is largely a
measure of the effect of improvements in the quality of ...; inputs
and how they are used. It includes technical progress, improvements
in the work force, improvements in management practices, economies
of scale, etc.' In addition, one might also include the quality of
infrastructure, strength of property rights and cultural
A widely used measure of the standard of living in a country is
income (or output) per capita. While there are certainly debatable
issues about whether per capita output is a sufficiently
comprehensive indicator of average living standards (or
welfare),(10) they are not pursued in this paper. Instead we will
simply take income per capita as a convenient proxy for the
considerably more difficult-to-measure concept of economic
The fundamental source of sustained improvements in living
standards over time is productivity growth. The standard model of
economic growth due to Swan (1956) and Solow (1956) proposes
productivity growth (or technological change) as the primary cause
of growth in output per capita over the long term. According to
this model, increases in a country's savings rate or population
growth rate can temporarily raise the growth rate of output per
capita, but they do not permanently raise a country's growth rate
of output per capita. (This is because diminishing returns to
investment are assumed to come into play.) Thus, increasing the
national savings rate will (other things held equal) raise our
level of output per capita, but the long term growth
rate of output per capita will not be affected.(11) Only
persistent growth in productivity leads to continuing improvements
in living standards. In the absence of productivity growth there
would be no growth in output per capita over the longer term.
Although the Swan-Solow model provides a direct link from
productivity growth to improvements in living standards, it does
not provide any explanation of the source of growth in
productivity. In the Swan-Solow model productivity growth is
'exogenous' (i.e. determined outside the model).
The reasons why productivity (or technological efficiency) tends
to increase over time are left unexplained. One implication of this
is that the model does not give rise to any policy proposals as to
how productivity, and therefore a country's long-run growth rate,
can be increased.
Recent developments in the theory of economic growth have begun
to address this limitation of the Swan-Solow model. These new
models of economic growth are often classified under the heading of
'endogenous growth models'.(13) The new growth models typically
proceed in one of two ways. One approach is to allow the growth
rate of technology to be affected by some other variables, e.g. by
the amount of resources devoted to research and development
(R&D) or by the existing stock of knowledge.(14) The
alternative approach is to broaden the concept of capital to
include human capital.(15)
New growth theories have a number of interesting features.
First, they allow for the possibility of self-sustaining growth in
output per capita, i.e. the long run growth rate of output per
capita is determined within the model rather than by an exogenous
rate of productivity growth. A second feature is the importance of
spillover benefits (or positive external effects) which can occur
when a firm or an individual makes an investment decision. As
Dowrick (1995) points out, investment in R&D or in human
capital by workers can generate new knowledge, the benefits of
which are not necessarily completely captured by the person making
the investment decision.(16) This implies that the production of
knowledge is likely to be sub-optimal if it is entirely determined
by market forces. As a result, there may be a role for governments
to subsidise investment in R&D, education and training. Thus, a
third feature of the new growth theories is that government policy
choices may be able to influence an economy's growth rate of output
per capita in the long run. These issues are discussed further in a
later section of this paper.
Data on productivity growth in the Australian economy can be
obtained from a number of sources. The Australian Bureau of
Statistics (ABS) currently publishes annual estimates of single and
multifactor productivity growth. However, there are a number of
other studies which present estimates of productivity growth for
Australia using a variety of methodologies, e.g. see EPAC (1989)
for a survey of studies before 1989, Dao, Ross and Campbell (1993),
Howe (1993) and Fox and Kohli (1996). Broadly speaking there are
two main empirical regularities that have emerged from the various
attempts to measure productivity growth in Australia:
- There was a slowdown in the trend rate of productivity growth
during the 1980s. However, it is difficult both to identify exactly
when this slowdown began and, as a consequence, to associate it
with a particular event.
- The slowdown in productivity growth is more pronounced in
labour productivity than in multifactor productivity.
The single and multifactor estimates of productivity growth
published by the ABS are for the market sector of the
Australian economy. The market sector excludes a number of sectors
involved in the production of services in the economy: government
administration and defence, finance and insurance, property and
business services, education, health and community services and
part of personal and other services. These sectors are excluded
because the measures of output for these industries are either
based on the assumption of no change in labour productivity or do
not measure it adequately. (17)
Figure 1 shows the ABS indexes (base year 1989-90 = 100) of
labour, capital and multifactor productivity for the last thirty
years. Clearly, both labour productivity and multifactor
productivity exhibit positive growth over the period. In contrast,
the level of capital productivity has been roughly constant over
the thirty year period.
Figure 2 shows the annual growth rates of the three measures of
productivity. There are two notable features in the graph. First,
there is a tendency for all three growth rates to move together
over time, although this is particularly true for labour
productivity and multifactor productivity.(18) Second, it is
apparent that all three measures of productivity are affected by
business-cycle fluctuations. Productivity (however measured) tends
to fall during recessions and increase in booms. This
characteristic is particularly evident in the data after 1980.
Since productivity measures seem to be affected by fluctuations in
the business cycle it is difficult to interpret year-to-year
changes in such measures. However, looking at average growth rates
over longer periods of time is one method for identifying the
longer term trends in productivity.(19)
Source: ABS 5234.0
Source: ABS 5234.0
Table 1 shows estimates of productivity growth over the period
1965-66 to 1994-95 and for five and ten year sub-intervals. Over
the entire period labour productivity grew at an average rate of
2.2 per cent per year while multifactor productivity grew at an
average rate of 1.5 per cent per year. The data suggest that labour
and multifactor productivity growth has been slower in the decade
from 1985-86 to 1994-95 then in the previous twenty years. However,
this largely reflects the relatively low rates of labour and
multifactor productivity growth in the second half of the 1980s. In
the five years since 1990-91 productivity growth rates have been
slightly above the average for the full period. Over the thirty
year period growth in capital productivity has been essentially
zero. This reflects negative growth rates in the first twenty years
and positive (above average) capital productivity growth rates over
the most recent decade.(20)
Table 1: Productivity Growth Rates-Aggregate Economy
Labour Capital Multifactor
Productivity Productivity Productivity
1965-66 - 1969-70 2.46 -0.10 1.57
1970-71 - 1974-75 3.14 -0.25 2.10
1975-76 - 1979-80 2.47 -0.30 1.70
1980-81 - 1984-85 2.45 -0.88 1.40
1985-86 - 1989-90 0.47 0.84 0.61
1990-91 - 1994-95 2.23 0.63 1.66
1965-66 - 1974-75 2.80 -0.18 1.84
1975-76 - 1984-85 2.46 -0.59 1.55
1985-86 - 1994-95 1.35 0.74 1.14
1965-66 - 1994-95 2.20 -0.01 1.51
Source: ABS 5234.0.
A number of other studies have also provided estimates of
productivity growth in Australia. Typically, one of the main issues
addressed by these studies is whether there is evidence that
productivity growth, either labour or multifactor productivity, has
slowed-down since the 1960s. Has there been a trend decline in
productivity growth ? Dowrick (1995, 1990) argues that the trend
rate of labour productivity growth in Australia did decline (as
happened in other OECD countries) following the first OPEC-induced
oil price shock in 1973 and then again in the 1980s. Dowrick (1995)
characterises the trend behaviour of labour productivity in
Australia in the following manner:
In the late 1960s and early 1970s, annual growth in labour
productivity averaged nearly 3 per cent. This rate of growth
declined to 2 per cent in the ten years after the watershed of the
1973 oil crisis, and has declined even further to just 1 per cent
in the most recent 10 year period. (page 29)
He suggests that the slowdown in the mid-1980s can be accounted
for by faster employment growth and slower growth in the
With respect to MFP, Dowrick argues that while MFP growth did
slow down in Australia after 1973, there is less evidence of a
slowdown in the 1980s. Consistent with this view, EPAC (1989)
surveyed a number of Australian studies and concluded that the
slowdown in productivity in the 1980s was more pronounced for
labour productivity than for multifactor productivity. The ABS data
presented in Table 1 is also broadly consistent with these
The ABS does not publish official estimates of multifactor
productivity growth for different sectors of the economy. However,
indexes of labour productivity are published in the National
Accounts for a number of sectors in the economy. Table 2 presents
estimates of labour productivity growth for eleven sectors of the
Australian economy over the last decade. Consistent with the data
in Table 1, the labour productivity index measures constant-price
gross product per hour worked.
Over the last decade, the sectors with the strongest growth in
labour productivity were Electricity, Gas and Water and
Communications. Mining, Manufacturing and Transport and Storage had
labour productivity growth rates that were above the average for
the market sector as a whole. The remaining sectors were all
relatively poor performers. The majority of the sectors showed
higher productivity in the first half of the 1990s compared to the
second half of the 1980s.
Table 2: Labour Productivity Growth Rates-Sectoral
Sector Average for the Period
1985-86 1990-91 1985-86
to to to
1989-90 1994-95 1994-95
Agr, forest, fish 0.64 -0.01 0.32
Mining 1.33 5.43 3.38
Manufacturing 2.46 4.24 3.35
Elec, gas, water 8.50 6.03 7.27
Construction -1.37 0.47 -0.45
Wholesale trade -0.42 1.82 0.70
Retail trade -1.60 0.99 -0.31
Accom, cafes, rest -0.58 -0.92 -0.75
Trans, storage 1.67 3.89 2.78
Communication 8.17 7.84 8.01
Culture, recre -3.41 1.75 -0.83
Market Sector 0.47 2.23 1.35
Source: ABS 5204.0
The Australian Bureau of Agricultural and Resource Economics has
produced estimates of MFP (based on gross-output measures) for both
the rural sector and the resource sector.(21) They find that for
the period 1971-72 to 1988-89 productivity growth in the rural
sector was about 2 per cent per year, and for the resources sector
it averaged about 1.4 per cent per year over the period 1971-72 to
1985-85. By way of comparison, productivity growth in the
manufacturing sector averaged 1.3 per cent per year over the same
There is a general perception that in recent years productivity
growth in Australia has been slower than what has occurred in other
OECD countries. Studies such as Baumol, Blinder, Gunther and Hicks
(1992) present evidence which indicates that the average growth of
labour productivity for Australia over the period 1961-92 was lower
than for Japan, Germany, France, the UK and Italy, although it was
better than for the US and Canada. Similar claims are made by Dao,
Ross and Campbell (1993) who present empirical estimates indicating
that for 13 OECD countries only the US experienced a lower growth
rate of MFP than Australia over the period 1970 to 1987. The
average growth over the period for all countries was 1.3 per cent
per year while Australia achieved a rate of 0.8 per cent.
Table 3: Comparison of Productivity Growth in Australia and the
Australia 2.6 1.4
OECD 3.5 1.8
Australia 1.4 0.3
OECD 1.4 0.4
Australia 0.6 0.4
OECD 1.8 1.2
Source: Howe (1993, Table 2, page 22)
Howe (1993) presents some additional evidence on Australia's
relative productivity performance, which is summarised in Table 3.
Note that Howe's primary data source is the OECD Outlook database,
so the numbers in Table 3 are not strictly comparable with those in
Table 1. From Table 3 we can see that while Australia's
productivity growth was roughly comparable with the OECD average in
the late 1970s and early 1980s, our performance was relatively
weaker in the period from 1983 to 1991.
Given the widely held view that Australia's productivity growth
rate over the last two decades has been low relative to other OECD
countries, we now examine the various arguments that have been
proposed to explain our poor performance. Recall that in the
Swan-Solow growth model productivity growth is exogenously
determined. Thus, it is necessary to look outside that model to
find the fundamental causes of technological improvements. Some of
the main economic determinants of productivity growth are discussed
One interpretation of Australia's relatively low productivity
growth rate compared to other countries is that it simply reflects
productivity 'catch-up' by other countries. The argument is that
countries with low initial (or starting) levels of
productivity will tend to experience relatively faster
growth in productivity than countries which already have
high levels of productivity. The idea is that economies with low
levels of productivity are able to adopt best-practice techniques
from leader economies. (International trade would be one means by
which new technology could be transferred between countries, or by
copying the best-practice technology.) If Australia had a high
level of productivity in 1970 relative to other OECD countries it
would be reasonable to expect the other OECD countries to
experience relatively faster productivity growth as they caught up
Helliwell and Chung (1992), Dowrick and Nguyen (1989) and
Dowrick (1992) present results for OECD countries which suggest
that there has been evidence of catch-up and convergence in
productivity levels. The poorer OECD countries have, on the whole,
experienced faster growth in multifactor productivity.(22) Dao,
Ross and Campbell (1993) also examine whether there is any
empirical support for the catch-up argument as an explanation for
the differences between productivity growth in Australia and the
OECD. They compare the level of productivity in Australian industry
to that in other OECD countries for the periods 1970-73 and
1984-87. They argue that Australia's labour and capital
productivity levels were typically below the OECD average
in both sample periods. This suggests that in fact there may have
been some opportunities for Australian industries to catch up with
productivity levels in other OECD countries. It is evident from the
data that such catch-up did not occur.
While catch-up cannot explain Australia's relatively poor
productivity performance compared to other industrialised nations,
it may be a more reasonable argument as to why growth rates of
output per capita in the newly industrialising countries
(particularly in East Asia) have exceeded our own. Consider the
data in Table 4, which is taken from Dowrick (1995, Table 2.5).
Notice that in 1970 Australia had a high level of per capita GDP
relative to other countries in the East Asian region. However, over
the next twenty years the average growth rates of the other
economies in the region have exceeded Australia's, sometimes by a
significant margin. Dowrick argues that this cross-country pattern
of growth rates is broadly consistent with the catch-up phenomena
and that the growth rates for the newly industrialising East Asian
economies will slow down as their level of GDP per capita
approaches those of developed countries.(23)
Table 4: Level and Growth Rate of Output per Capita for
Australia and Selected East Asian Economies
Level of GDP Average annual growth
per capita rate of GDP per capita,
1970 1970-80 1980-90
Australia 10917 1.5 1.3
Japan 7500 3.2 3.7
Hong Kong 4456 6.8 4.9
Singapore 3155 7.9 4.5
Malaysia 2117 5.8 2.6
Taiwan 2387 7.0 5.7
Thailand 1508 3.5 5.0
South Korea 1688 6.2 7.6
China 825 4.1 6.3
Indonesia 700 5.8 4.4
Source: Dowrick (1995, Table 2.5, page 32)
Investment in Physical Capital
Growth of labour productivity (Y/L) is affected by growth in the
capital to labour (K/L) ratio and the growth rate of MFP. Increases
in (K/L), or 'capital deepening', tend to improve labour
productivity since capital productivity is relatively slow to
change (see Table 1). This provides a direct link between physical
investment and labour productivity. If investment in physical
capital is low this will eventually produce a decline in the growth
rate of the capital stock and (K/L), particularly if the labour
input (L) is growing strongly. However, Edey and Britten-Jones
(1990) report that over the three decades 1960-90, Australia's
gross investment to GDP ratio has tended to exceed that OECD
average by about 2 to 3 percentage points of GDP. This would seem
to be inconsistent with a relatively slow growth of labour
productivity (other things held equal).(24) However, our relatively
higher growth rate of population and lower level of capital
productivity than for the OECD means that Australia has needed to
maintain an investment to GDP ratio above the OECD average.(25)
In principle, investment in physical capital should not have any
direct effect on the rate of growth of MFP, nor in the Solow-Swan
framework does it have any long term affect on the rate of economic
growth (due to diminishing returns). However, a number of possible
exceptions have been suggested.
First, to the extent that technological improvements are
embodied in the capital stock and these quality improvements are
not appropriately captured in measures of the capital stock, then
capital deepening may lead to higher levels of measured MFP.
Second, it has been argued that investments in some types of
physical capital appear to have very high returns (particularly at
the aggregate level) which may reflect the existence of large
external benefits. The two most widely discussed examples are
investment in equipment and in public infrastructure.
De Long and Summers (1991, 1992) find, on the basis of
cross-country evidence, that real returns to investment in
equipment are very high relative to other types of investment, e.g.
in dwellings and structures. They claim that spillover benefits are
an important component of this higher return. For example, when a
worker learns to use a new machine the acquired knowledge is a
public good that can be transferred to other industries.
Using both time series and cross section data, a number of
empirical studies have shown that certain types of public
infrastructure appear to be very productive (i.e. they have high
marginal products).(26) In general, this fact will not be
adequately captured by a simple aggregate measure of the capital
stock which includes both private and public capital. Measures of
MFP for the private sector in the US and Australia are
significantly correlated with public investment expenditure.(27)
Aschauer (1989) interpreted his results as indicating that the
decline in the ratio of public investment to GDP in the US since
the mid-1970s was an important causal factor in the observed
decline in US productivity growth over that period.(28)
Investment in Research And Development
A number of authors have developed models in which the growth
rate of technological progress (roughly speaking, the growth rate
of MFP) depends on the existing level of technology and on the
quantity of resources (the labour and capital) devoted to the
production of new ideas and inventions-i.e. on the share of the
economy's resources devoted to R&D.(29) Under certain
conditions these models imply that an increase in the fraction of a
country's resources devoted to R&D will raise the growth rate
of MFP and, as a consequence, the growth rate of output per capita.
What the models do not directly address is what determines the
amount of resources devoted to R&D. However, important
influences are likely to include:
- public (or private) funding of pure (or fundamental) scientific
- the private incentives that exist for undertaking R&D, e.g.
the patent system, and
- the types of opportunities that exist for talented
R&D expenditure in Australia is low relative to other OECD
countries.(30) In 1989-90 Australia's gross expenditure on R&D
was 1.23 per cent of GDP compared to the OECD average of 1.87 per
cent of GDP. (31) Part of the explanation for this fact is that
private sector R&D investment in Australia has been low by
world standards. For example, in 1989 only about 40 per cent of
total R&D expenditure was due to the business sector. In most
OECD countries business sector R&D spending exceeds that by the
public sector. According to EPAC (1987), while Australia's
expenditure on basic and applied research compares favourably to
other advanced economies, our performance in the commercial
application of basic research to product design and development has
been poor. The poor performance by business is attributed to a
number of factors including:
- the role of tariffs in reducing competitive pressures on
- lack of managerial skills,
- an inadequate exchange of information between basic researchers
and industry, and
- a lack of venture capital to fund innovative ideas and
Of course, there are other ways to obtain the benefits of
R&D. For example, businesses can buy the technology directly or
indirectly (though a licence or by purchasing the product that
embodies the desired technology). Thus, a low level of R&D does
not necessarily imply technological inferiority. As Lattimore
(1991) points out, it is important to measure not just expenditure
on R&D but also the extent of joint ventures, licensing, and
borrowed (or copied) technology.
Given the external effects that are generally associated with
the production of knowledge, firms that undertake R&D may not
be able to appropriate all the benefits from such expenditures and
this can provide a basis for government intervention, e.g.
governments typically provide subsidies or tax breaks for firms
undertaking R&D. For a small, open economy like Australia,
another means of encouraging innovation is through increased
international trade. This is one way of benefiting from R&D
Murphy, Schleifer and Vishny (1991) note that innovations and
advances in knowledge are often the product of talented
individuals. However, such individuals have other opportunities. In
particular they can also pursue other activities which are forms of
rent-seeking, i.e. they can seek to capture existing wealth rather
than produce new wealth. If talented members of society are
encouraged to pursue research then economic growth will be
higher.(32) Factors that might work to encourage individuals to
pursue such socially productive activities include:
- the existence of a large market from which rewards can be
- a well-functioning capital market that allows firms to expand
- well-defined property rights which ensure people can keep the
returns from their activities.
Investment in Human Capital
Another factor which seems to be an important determinant of
economic growth rates is human capital accumulation (i.e.
increasing labour skills). Mankiw, Romer and Weil (1992) present
evidence that variations in human capital are an important
determinant of cross-country differences in income per capita.
There is also some evidence to suggest that the very high rates of
growth in some East Asian countries (see Table 4) are primarily due
to human capital accumulation.(33)
Individuals can acquire human capital by two basic means:
through formal schooling and by on-the-job training e.g. learning
by doing.(34) While both are likely to be important,(35) it is the
latter which appears to provide the most likely explanation for
persistently high rates of growth in income and productivity.
The effects of human capital accumulation through learning by
doing has been examined in the new growth literature. The basic
idea is that as individuals produce goods they tend to think of
ways of improving the production process. Thus, productivity rises
without any evident changes in the production process.(36) With
learning by doing, knowledge accumulates, not by deliberate effort,
but as a by-product of economic activity. If learning by doing is
an empirically important phenomena, then the rate of MFP growth
depends on the amount of new knowledge that is generated by
What actually determines the amount of learning by doing in a
particular industry or economy is an unresolved question. Lucas
(1993) argues that if learning by doing is itself subject to
diminishing returns to scale, a sustained rise in the growth rate
of productivity will require the continued introduction of new
goods, not just continued learning on a given set of goods. One
means by which a small economy can expand the range of goods it can
(potentially) produce is selling on the world market. Thus, a
relatively open economy seems to be an important precondition for
A wide variety of other factors have been proposed as affecting
productivity levels and growth rates. These include:
- government regulation, e.g. environmental standards or labour
- the stance of macroeconomic policy, e.g. Fischer (1993)
presents evidence that productivity growth is negatively related to
- cultural factors, and
- the type of social and economic institutions that a country
While the effect of many of these variables on productivity is
difficult to quantify, it is difficult to disagree with the
following statement from Hansen and Prescott (1993: 281):
Every nation has a set of rules and regulations that govern the
conduct of business. These have consequences for the incentives to
adopt more advanced technologies and for the resources required to
operate an existing one. Bureaucracies that assist in the adoption
of new technologies … foster technological growth. Systems
that divert entrepreneurial talent from improving technologies to
rent-seeking activities have an adverse effect on growth. The
reason for the huge difference between the United States and India
must be that India has been less successful than the United States
in setting up institutions conducive to economic development.
Microeconomic reform is one process by which existing economic
and social institutions in Australia are being changed. In the
following section we consider the possible impact of these
institutional changes on productivity.
According to Quiggin (1996), microeconomic reform began in
Australia in 1973 with the decision by the Whitlam Government to
reduce tariffs across-the-board by 25 per cent. Since then, state
and federal governments of all political persuasions have pursued
microeconomic reform with varying degrees of intensity.
Microeconomic reform is the generic term for 'government policies
designed to deregulate or re-regulate product, service and factor
markets in such a way as to promote competition and efficiency in
relation to both domestic and international markets' (Robertson,
Quayle and McEachern, 1994: 384). Economic policies that tend to be
encompassed by the microeconomic reform heading include:
- deregulation of the labour market,
- reducing tariffs and other forms of protection for domestic
industry against foreign competition,
- deregulation of the financial markets,
- privatisation and corporatisation of public enterprises,
- reducing distortionary taxes.
At a general level, microeconomic reform can be seen as
'removing or reducing restrictions on trade', not just
international trade but also trade among individuals and firms
within a country. The basic objective is to allow the allocation of
resources in Australia to better reflect market (as opposed to
non-market) outcomes. Where there are clear examples of market
failure (e.g. monopoly power or externalities) then 'microeconomic
reform may involve redesigning and improving regulations rather
than deregulation or smaller government'.(38)
There is a widely held view that microeconomic reform will
produce considerable benefits for the Australian economy. For
example, see EPAC (1990), Filmer and Dao (1994) and Clark (1995).
It is generally perceived that one of the beneficial outcomes of
microeconomic reform will be an improvement in Australia's
productivity performance. According to Filmer and Dao (1994: 1)
'microeconomic reform aims to boost productivity growth by creating
an environment in which resources are allocated to their most
productive uses and firms use the most efficient methods of
On the basis of standard economic theory it is easy enough to
argue that microeconomic reform, by eliminating restrictions on
trade and increasing competitive pressures (both domestic and
external), will tend to improve both productive efficiency (i.e.
each firm adopts the technique which, given the existing state of
knowledge, maximises its output from employed resources) and also
allocative efficiency (i.e. resources are allocated across firms
and industries so that the set of goods and services being produced
are consistent with preferences of domestic and foreign
While such changes are likely to have important effects on the
level of productivity (and, as a consequence, on the level
of output), what is less obvious is whether they will lead to a
permanent improvement in the growth rate of productivity.
It seems important to distinguish between a once-and-for-all
increase in the level of productivity and an increase in the growth
rate. In the context of the Swan-Solow model of economic growth it
is the latter which is the key to longer term improvements in
While this is not necessarily the case in endogenous growth
models, it is typically the case that the exact mechanisms by which
microeconomic reform will raise the rate of productivity growth are
not clearly specified.
A reading of the literature suggests two ways in which
microeconomic reform may have an impact on productivity growth, at
least in the medium term, if not permanently. The first of these is
through the exposure of Australian firms to increased competition,
both internationally via the reduction of protection and
domestically via competition policy.(39) According to Dao, Ross and
Campbell (1993: 22) 'competitive pressure that makes effective use
of high quality productive resources is a strong formula for
improving productivity'. In a similar vein Filmer and Dao (1994:
43) argue that 'a more market-friendly and pro-competitive economy
in product markets and enhanced cooperation in labour markets
provides a better basis for productivity growth overall'.
A second aspect of the productivity growth effect from
microeconomic reform is based on the belief the Australian firms
will have the incentive and the ability to exploit 'catch-up'
opportunities. The basic idea is that for many Australian firms and
industries their methods of production (e.g. management practices,
capital equipment etc.) are below current best practice in other
countries. If Australian firms were given the necessary incentives
(and opportunities) to adopt world best practice this would have a
significant effect on domestic productivity growth, at least during
the catching-up phase. In addition, if there are continuing
improvements in world best practice this would further stimulate
domestic productivity growth.
There have been a number of studies that have attempted to
quantify the expected gains from microeconomic reform. For example,
see the Bureau of Industry Economics (1990), Business Council of
Australia (1994), Filmer and Dao (1994), Industry Commission
(1995). The estimated benefits range between 5 to 20 per cent of
GDP, with the increase occurring within 5 to 10 years. In contrast
to these studies, Quiggin (1996) presents a critical analysis of
the microeconomic reform process in Australia and is clearly
sceptical about whether it can deliver the output and productivity
growth effects that have been claimed. He argues that the
microeconomic reforms undertaken to date have not had any
significant impact on the aggregate rate of productivity growth.
According to Quiggin, this reflects the fact that the net gains
from microeconomic reforms already undertaken have been small,
certainly less than 1 per cent of GDP. He also argues that 'large
estimates of the benefits of reform have been obtained primarily by
the invocation of supposed dynamic benefits unknown to mainstream
neoclassical microeconomics' (page 222).
It has to be admitted that there is considerable uncertainty
about the exact magnitude of the output and productivity benefits
of microeconomic reform. Standard economic theory predicts that
microeconomic reform will have a positive effect on the level of
Australian output. The economic mechanisms by which this will occur
are well understood. However, these are essentially
once-and-for-all gains and will have, at best, a temporary effect
on the growth rate of productivity and output. Whether
microeconomic reform will lead to any permanent increase in
productivity growth is much more uncertain. Certainly, it is
difficult to confidently identify the mechanisms by which this will
occur. This largely reflects our continuing uncertainty about
exactly what drives technological progress.
This paper has stressed the importance of productivity growth as
a fundamental source of long-term improvements in Australia's
living standards. However, standard measures of productivity
suggest that the rate of productivity growth in Australia slowed
down in the 1980s, relative to the 1960s. While Australia is not
unique amongst OECD countries in experiencing a slowdown in
productivity growth, cross-country comparisons of productivity
growth indicate that Australia's performance has frequently been
below the OECD average. Raising the long-term productivity growth
rate in Australia, even by a fraction of a percent, is likely to
yield significant benefits in terms of higher average living
Identifying the factors which determine the rate of productivity
growth has proven a difficult task. A large number of possible
influences have been suggested, such as spending on research and
development, education and human capital accumulation, and the
degree of competition. However, as yet no robust consensus has
emerged as to what are the most important factors. This uncertainty
places limits on the ability of policymakers to design and
implement economic policies that will raise productivity
In Australia, a currently preferred mechanism for raising
productivity is microeconomic reform. Exactly what effect various
microeconomic reforms will have on productivity is difficult to
estimate with any precision. While one of the likely effects of the
increased competition associated with microeconomic reform is a
one-off improvement in the level of productivity, it seems less
certain that these reforms will produce a permanent increase in the
growth rate of productivity.
- See Karmel and Polasek (1978) for an introduction to the
construction of index numbers.
- This framework is sufficiently general to incorporate external
effects such as pollution. Suppose a country produces less
pollution for a given level of input and given level of output,
then this amounts to a productivity increase.
- The economic concept of a production function is a useful
device for thinking about technology and technological change. A
production function summarises the current state of technology by
indicating the maximum amount of output that it is technically
possible to produce from a given quantity of inputs. Technical
progress (or productivity growth) arises from a shift in the
production function such that more output can be produced from the
existing amount of inputs.
- Hansen and Prescott (1993) use a broader definition of the
state of technology to include the set of rules and regulations
that control the conduct of business and trade.
- It seems reasonable to suppose that most people in Australia
would prefer the current state of technology to that which existed
one hundred or even fifty years ago.
- See Morrison (1993).
- Another name for MFP is the Solow residual: see Solow
- See Harcourt (1972) and Lattimore (1991). On a related issue,
it may be difficult to accurately measure all of the inputs to the
production process, e.g. the stock of research and development
capital. The end result is that these omitted inputs will show up
in the productivity measure.
- See Romer (1996).
- See ABS (Australian National Accounts, March 1990) and Nordhaus
and Tobin (1972).
- Essentially, while higher saving implies higher investment and
capital accumulation, as the level of capital per head rises there
are diminishing returns, i.e. doubling capital per head leads
output per head to increase by a factor which is less than
- To see the relatively close match between productivity growth
and growth in output per capita, observe that over the period
1965-66 to 1994-95 real output per capita in Australia grew at an
average rate of about 2 percent per annum while over the same
period MFP growth averaged about 1.5 percent per annum.
- See Romer (1996) and Barro and Sala-I-Martin (1995).
- See Romer (1986, 1990).
- See Lucas (1988).
- Knowledge is nonrival when the use of an idea or theory by one
individual does not preclude its simultaneous use by someone else.
Whether or not others can be prevented (or excluded) from using the
knowledge depends on the nature of the knowledge and the system of
- Essentially, the market sector consists of those industries
where a direct measure of output is available.
- This reflects, in part, the fact that labour productivity
depends on both MFP and the capital to labour ratio which is
relatively slow to change.
- The ABS accounts for business-cycle effects by computing growth
cycles and comparing productivity growth from the peak of one
growth cycle to the peak of the next.
- As Dao, Ross and Campbell (1993) note, low or declining capital
productivity is not necessarily a problem for economic growth
provided it is offset by gains in labour productivity associated
with an increase in the capital to labour ratio.
- See EPAC (1990).
- See Dowrick (1992).
- What the primary source of the rapid economic growth in some of
the East Asian economies is remains an open question.
- Strictly speaking, it is not gross investment but net
investment (i.e. gross investment less depreciation) that
contributes to the stock of capital and obtaining accurate
estimates of economic depreciation is a difficult problem.
- See EPAC (1995).
- This is particularly the case for core infrastructure like
roads and highways, water and sewer systems, ports, airports,
- See Aschauer (1989) and Otto and Voss (1994).
- One unresolved difficulty that confronts all of these studies
is the question of causality, i.e. does investment cause growth or
vice versa ?
- See Romer (1990) and Grossman and Helpman (1991).
- See Lattimore (1991).
- Australian Science and Innovation Resources Brief,
AGPS, Canberra, 1992.
- Murphy, Schleifer and Vishny (1991) find a positive correlation
between economic growth and the fraction of engineers in a country
but a negative correlation between growth and the fraction of
- Lucas (1993).
- See Arrow (1961).
- See EPAC (1986) and Clare and Johnston (1993).
- Lucas (1993) discusses the example of the Liberty Ship-a type
of cargo vessel built in US shipyards during World War Two-where
there were reductions in man-hours per ship of the order of 12 to
24 percent with each doubling of cumulative output.
- See Sloan (1992).
- EPAC (1990: 4).
- See Hilmer, Rayner and Taperell (1993).
Australian Bureau of Statistics, Australia's Environment:
Issues and Facts, Catalogue No. 4140.0, 1992.
Australian Bureau of Statistics, Australian National
Accounts: Multifactor Productivity, 1994-95, Catalogue No.
Australian Bureau of Statistics, Australian National
Accounts: Annual 1995-96, Catalogue No. 5204.0.
Aschauer, D., 'Is Public Expenditure Productive?', Journal
of Monetary Economics, 24 (2), 1989.
Arrow, K. 'The Economic Implications of Learning by Doing',
The Review of Economic Studies, 29, 1961.
Barro, R., and X. Sala-I-Martin, Economic Growth,
McGraw-Hill, New York, 1995.
Baumol, W., A. Blinder, A. Gunther and J. Hicks,
Economics, 2nd ed., Harcourt Brace Jovanovich, Sydney,
Bureau of Industry Economics, Microeconomic Reform, BIE
Discussion Paper 9, AGPS, Canberra, 1990.
Business Council of Australia, Business Council
Bulletin, Melbourne, 1994.
Clare, R. and K. Johnston, 'Education and Training in the
1990s', EPAC Background Paper, No. 31, July 1993.
Clark, D., 'Microeconomic Reform', in The Australian
Economy, ed. Peter Kriesler, Allen & Unwin, Sydney,
Dao, Dan, Steve Ross and Robert Campbell, Structural Change
and Economic Growth, EPAC Background Paper No. 28, Office of
EPAC, June 1993.
De Long, J., and L. Summers, 'Equipment Investment and Economic
Growth', Quarterly Journal of Economics, 106, May
De Long, J., and L. Summers, 'Equipment Investment and Economic
Growth: How Strong is the Nexus?' Brookings Papers on Economic
Activity, 2, 1992.
Denison, E., Trends in American Economic Growth
1929-1982, Washington D.C., Brookings Institution, 1985.
Dowrick, S., and T. Nguyen, 'OECD Comparative Economic Growth
1950-85: Catch Up and Convergence', American Economic
Review, 79, 1989.
Dowrick, S., 'Explaining the Productivity Slow-down of the
1980s', Australian Bulletin of Labour, 16 (3), September
Dowrick, S., 'Technological Catch Up and Diverging Incomes:
Patterns of Economic Growth 1960-88', The Economic
Journal, 102 (412), May 1992.
Dowrick, S., 'Economic Growth: Recent History and Prospects', in
The Australian Economy, Peter Kriesler (ed.), Allen &
Unwin, Sydney, 1995.
Edey, M., and M. Britten-Jones, 'Saving and Investment', in
The Australian Macroeconomy in the 1980s, S. Grenville
(ed.), RBA, Sydney, 1990.
Economic Planning Advisory Council (EPAC), Human Capital and
Productivity Growth, Council Paper, February 1986.
EPAC, High Technology Industry in Australia, Council
Paper No. 25, January 1987.
EPAC, Productivity in Australia: Results of Recent Studies,
Council Paper No. 39, April 1989.
EPAC, Microeconomic Reform, Council Paper No.42, May
EPAC, Productivity Growth and Developments in Australia's
Primary Industries, Discussion Paper 90/06, September 1990.
EPAC, Investment and Economic Growth, Commission Paper
No. 9, 1995.
Filmer R. and D. Dao, Economic Effects of Microeconomic
Reform, EPAC Background Paper, No. 38, February 1994.
Fischer, S., 'The Role of Macroeconomic Factors in Growth',
Journal of Monetary Economics, 32, 1993.
Fox, K., and U. Kohli, 'GDP Growth, Terms of Trade Effects and
Total Factor Productivity', Mimeo, August 1996.
Grossman, G., and E. Helpman, Innovation and Growth in the
Global Economy, MIT Press, Cambridge, 1991.
Hansen, G. and E. Prescott, 'What Caused the Last Recession ?',
American Economic Review, Papers and Proceedings, 83 (2),
Harcourt, G., Some Cambridge Controversies in the Theory of
Capital, Cambridge University Press, Cambridge, 1972.
Helliwell, J., and A. Chung, 'Aggregate Productivity and Growth
in an International Comparative Setting', in International
Productivity and Competitiveness, B. Hickman, (ed.), Oxford
University Press, Oxford, 1992.
Hilmer, F., M. Rayner and G. Taperell, National Competition
Policy, Report by the Independent Committee of Inquiry,
Howe, John, Australia's Growth Potential in the 1990s,
EPAC Background Paper No. 35, Office of EPAC, December 1993.
Industry Commission, The Growth and Revenue Implications of
Hilmer and Related Reforms, AGPS, Canberra, 1995.
Karmel, P.H. and M. Polasek, Applied Statistics for
Economists, 4th ed, Pitman, Melbourne, 1978.
Lattimore, R., 'Investment and the Capital Stock: What do the
Figures Really Mean?', in The Economic Dynamics of Australian
Industry, Clive Hamilton, (ed.), Allen & Unwin, Sydney,
Lattimore, R., 'Research and Development: Hidden Investment in
Australian Industry', in The Economic Dynamics of Australian
Industry, Clive Hamilton, (ed.), Allen & Unwin, Sydney,
Lucas, R., 'On the Mechanics of Economic Development',
Journal of Monetary Economics, 22 (1), July, 1988.
Lucas, R., 'Making a Miracle', Econometrica, 61 (2)
Mankiw, G., D. Romer and D. Weil, 'A Contribution to the
Empirics of Economic Growth', Quarterly Journal of
Economics, 107, May 1992.
Morrison, C., A Microeconomic Approach to the Measurement of
Economic Performance, Springer-Verlag, New York, 1993.
Murphy, K., A. Shleifer and R. Vishny, 'The Allocation of
Talent: Implications for Growth,' Quarterly Journal of
Economics, 104, August, 1991.
Nordhaus, W. and J. Tobin, 'Is Growth Obsolete?' Fiftieth
Anniversary Colloquium, V, National Bureau of Economic
Research, Columbia University Press, New York, 1972.
Otto, G. and G. Voss, 'Public Capital and Private Sector
Productivity', Economic Record, 70, 1994.
Quiggin, J. Great Expectations: Microeconomic Reform in
Australia, Allen & Unwin, Sydney, 1996.
Robertson, P., 'A Productivity Manifesto: Concepts and
Definitions' Mimeo, School of Economics, University of New South
Robertson, T., M. Quayle, and W. McEachern, Economics: A
Contemporary Introduction, Nelson, Melbourne, 1994.
Romer, D., Advanced Macroeconomics, McGraw-Hill, New
Romer, P., 'Increasing Returns and Long-Run Growth', Journal
of Political Economy, 94 (5), October, 1986.
Romer, P., 'Endogenous Technical Change', Journal of
Political Economy, 98 (5), October, Part 2, 1990.
Sloan, J. 'Until the End of Time: Labour Market Reform in
Australia', Australian Economic Review, 4th Quarter,
Solow, R., 'A Contribution to the Theory of Economic Growth',
Quarterly Journal of Economics, 70, 1956.
Solow, R., 'Technical Change and the Aggregate Production
Function', The Review of Economics and Statistics, 39 (5),
Swan, T., 'Economic Growth and Capital Accumulation',
Economic Record, 32, 1956.