This Selected Issues paper presents an empirical comparison of New Zealand’s growth performance with that of Australia during the post-reform period. The paper shows that most of the divergence in income per capita between the two countries has been the result of lower accumulation of capital per hour worked, and to a lesser extent, lower efficiency in utilizing resources in New Zealand. The paper also examines how migration has affected the income and welfare of New Zealand nationals.


This Selected Issues paper presents an empirical comparison of New Zealand’s growth performance with that of Australia during the post-reform period. The paper shows that most of the divergence in income per capita between the two countries has been the result of lower accumulation of capital per hour worked, and to a lesser extent, lower efficiency in utilizing resources in New Zealand. The paper also examines how migration has affected the income and welfare of New Zealand nationals.

I. An Exploration into the Income Divergence Between New Zealand and Australia1

A. Introduction

5. Despite the extensive program of reforms started in the second half of the 1980s, New Zealand has not performed as well as other industrialized countries. Since 1985, New Zealand’s real GDP growth has averaged around 2 percent per annum, compared with an average of 3 percent for OECD countries. As a result, New Zealand has slipped further from sixteenth to twentieth place in the OECD ranking of GDP per capita.

6. Meanwhile Australia, which has followed a broadly similar program of economic liberalization, has fared much better than New Zealand in the last 15 years in terms of growth of output per capita (Figure I.1). While New Zealand started from a higher level in the 1960s, Australia caught up in the 1970s, and since the start of the 1980s, the gap between the two economies has widened. With an average growth of around 3.7 per annum in the last 15 years, Australia has been able to significantly reduce the gap with the OECD average, and its GDP per capita now stands at about 30 percent above that of New Zealand.

Figure I.1.
Figure I.1.

Real GDP Per Capita

(In thousands of US$ at 1995 PPP)

Citation: IMF Staff Country Reports 2002, 072; 10.5089/9781451830231.002.A001

Source: OECD.

7. The increase in income per capita dispersion could be transitory (due to temporary shocks along the transition to steady-states), or it could be signaling that New Zealand and Australia are converging to two different long-run equilibria, where a high degree of inequality could persist. It could also be that the differences in the two country’s growth “fundamentals” are increasing and that inequality will increase further, before the relative position of the countries stabilizes. In each of these cases, it is the long-run difference in the levels of output between the two countries that is the most interesting one to explain.

8. The objective of this chapter is to assess in a coherent, quantitative framework some possible explanations for the increasing difference in output per capita between Australia and New Zealand in the post-reform period (1988–2000). In particular, it aims at quantifying the role played by differences in the accumulation of physical and human capital, and in changes of factor productivity. Special focus is placed on sectoral performance to examine whether aggregate divergence masks important differences in sectoral productivity.

9. A traditional Solow-type growth accounting framework is applied, both over time and across the two countries. Despite its limitations, this methodology remains an important first step in productivity analysis, as it permits disentangling the relative contribution to output from the accumulation of factors of production and the efficiency in their utilization (total factor productivity).2 This chapter differs from other recent studies that have utilized a growth accounting approach to evaluate New Zealand’s productivity performance (Diewert and Lawrence, 1999) in two respects: first, this chapter focuses on productivity levels; and second, it uses new chain linked estimates of output and productive capital stock from Statistics New Zealand and the Australian Bureau of Statistics.

10. The main results of the chapter are:

  • the 20 percent average difference in market sector’s GDP per capita between Australia and New Zealand over the last decade is entirely explained by a difference in labor productivity between the two countries.

  • Around ¾ of the gap in labor productivity is accounted for by the relatively lower capital deepening in New Zealand.

  • The difference in total factor productivity accounts for the other quarter.

  • Human capital accumulation contributed more to labor productivity in New Zealand than in Australia, owing to the significant upgrading of skills of New Zealand’s labor force.

  • In contrast to Australia, New Zealand’s TFP growth has benefited less from a reallocation of resources towards its most efficient sectors over the last decade.

11. Both the lower capital intensity and the absence of allocative gains since the end of the 1990s are in principle consistent with the existence of structural disadvantages that may limit New Zealand’s growth prospects, compared to those for Australia. In particular, the smaller domestic market may have prevented New Zealand from successful diversification away from primary production (where New Zealand has a strong comparative advantage) and towards higher-growth manufacturing and service sectors. Trade barriers that limit scale and investment opportunities may also have prevented full realization of New Zealand’s comparative advantage in primary production.

12. However, other factors may be at play in determining the divergence in GDP per capita between the two countries. In particular, the relatively higher share of New Zealand’s net capital invested in residential housing may have acted as a constraint on capital accumulation, especially given the relatively low household saving rate in New Zealand. Further, the relatively low capital intensity in agriculture, albeit partly explained by the omission of livestock from the stock of capital, seems to suggest that New Zealand has not fully exploited the potential for growth in the sector where it has a comparative advantage.

13. The chapter is structured as follows: Section B explains the approach followed to evaluate productivity levels in Australia and New Zealand and compare quality adjusted human and physical capital accumulation in the two countries. Section C presents the main results of the comparison, while Section D analyzes whether the two countries differed in their ability to benefit from a better allocation of resources (labor and capital) across sectors in the last decade. Section E concludes.

B. Comparing Productivity Levels

14. Table I.1 shows that on average, in the period 1988–1999, New Zealand’s market sector output per capita has been 20 percent smaller than that of Australia. A first step into the investigation of this difference is to break down the output per capita into its two components: output per hour worked (labor productivity) and hours worked per person (labor utilization). Table I.1 shows that New Zealand’s market sector utilized more labor resources relative to Australia.3

Table I.1.

Accounting for Differences in GDP Per Capita Between NZ and AUS, Market Sector, 1988-1999

article image
Output (Y): GDP at factor cost. Market Sector, in 1996 NZ dollars (Australian dollars converted using OECD PPP exchange rate). Source: Staff estimate based on data from SNZ and ABS.Hours worked (H): Total actual hours worked. Source: for New Zealand, the Household Labor Force Survey, for Australia, the Labor Force Survey (see Annex I.1).Employed (E): Employment of the business sector, source: OECD.Labor Force (L) and Persons (P) source OECD.

15. As New Zealand actually led Australia in terms of labor utilization, the difference in output per capita must be due to a gap in labor productivity. The next step is to decompose that gap into differences in factor utilization and in total factor productivity (TFP). This is done through a cross-sectional application of the Solow approach to growth accounting, which exploits the symmetry between time and space and approximates the relative TFP with the following expression:


where αH represent the income (y) share of labor input (h) and αK the income share of capital input (k). Equation [1] derives the relative TFP between New Zealand and Australia by subtracting the weighted average of the log difference in human and physical capital inputs from the log difference in output. A number of observations should be made regarding Equation [1]:

  • Comparing TFP levels of two countries amounts to asking the question: how much output could country X produce using country Y inputs or vice versa? As the choice of the base country affects the answer, this question implies an index number problem that must be solved with some form of weighting. Equation [1] is consistent with the index number theory, as the utilization of the arithmetic averages of the two factor income shares makes the relative TFP of Equation [1] a Tornqvist index, known to be the best discrete approximation of Divisia indexes.4

  • Equation [1] is based on the typical assumptions underlying the Solow approach to TFP, namely that factors of production are paid their marginal product (perfect competition in factor markets), and that technical change is not specific to any factor of production but improves the efficiency of all factors equally (it is Hicks-neutral). Constant returns to scale are assumed only when the income share of capital is estimated from the National Accounts as a residual (output less compensation of employees), but in principle are not required as the return to capital (and thus its income share) could be estimated directly.

  • The output concept is value added, not gross output. While the choice of the value added concept is justified by the focus on aggregate value added, the exclusion of intermediate inputs from the analysis may lead to a biased estimate of technical changes at a sectoral level (Basu and Fernald, 1995).

16. In light of these observations, great care should be exercised in taking the Solow residual as a measure of the technological endowment of a country. Other factors are included in the residual, such as those associated with changes in the level of organizational and managerial practices, model misspecification and measurement problems. The focus of this chapter however, is not as much in “explaining” TFP or in measuring differences in technological endowments of Australia and New Zealand, as it is in estimating the relative role played by capital accumulation and TFP as proximate causes of the large and increasing differences in output per capita between the two countries.

17. In this context, it is especially important to recognize that different types of labor and different capital assets contribute differently to output in Equation [1]. Treating them as if they were homogenous would amount to treating as productivity enhancements the investments that substitute among different types of inputs. A more accurate allocation of the sources of economic growth between investments and productivity would need to be based on constant quality indices of labor and capital inputs. Following the consolidated literature on TFP estimation (started by Jorgenson and Griliches, 1967), this study attempts to capture cross-country differences in the “quality” of human and physical capital accumulation by considering different types of labor and physical capital and by proxying their marginal productivities with market remuneration.

Labor Input

18. As for labor, three types of workers are identified for New Zealand and Australia: those with tertiary education, those with secondary education, and those without secondary school certification. For simplicity, they are all assumed to work the same number of hours, and the labor income share of each type of labor in Equation [1] is estimated by using the relative wages for the three groups.5

19. The difference between the weighted sum of hours worked by the three types of labor and the unweighted sum is an index of the compositional changes in labor inputs, or its quality improvement. Table I.2 shows that, according to this index, New Zealand has had an advantage over Australia in terms of human capital accumulation. The main force behind this result is the larger increase in workers with tertiary education, and the reduction in workers without secondary qualification in New Zealand compared to Australia.

Table I.2

A. Index of Labour Quality

Average growth rate of aggregate hours worked in market sector.

Weighted average of rates of growth of hours worked by different types of labor, with relative labor income as weights.

Difference between (2) and (1).

B. Employment by Educational Attainment, Market Sector 1/

Does not include finance and insurance.

Source: Census 1981 and 1996 (intermediate years obtained by interpolation)

Source: data for 1984 adopted from “Productivity and the Structure of Employment,” 1999, Productivity Commission; data for 1996 from Education and Training, ABS.

For New Zealand, includes school certificate and 6th form certificate (Year 12 or university entrance); for Australia, those who completed highest level of secondary school available (Year 12 or equivalent).

New Zealand 1981 includes non-university post-school qualification and undergraduate diploma; New Zealand 1996 includes basic, skilled, intermediate and vocational qualifications; Australia 1996 includes skilled and basic vocational qualification, associare and undegraduate diploma.

20. A limitation of this indicator is that it does not take into account differences in the “quality” of tertiary education. For example, it could be argued that, in terms of their contribution to output growth, it is more critical to have a larger pool of workers with tertiary degrees in scientific disciplines than in the arts or humanities, for example.6 Needless to say, any adjustment for the quality of tertiary education needs to be country-specific and, therefore, extremely difficult to pursue on a cross-country basis.7

Capital Input

21. The concept of capital input used in Equation [1] is the flow of “productive” services from a given stock of capital. Productive capital is obtained by adjusting each vintage in use for the loss of efficiency caused by physical decay (“wear and tear”) and retirement. This allows the expression of the stock of capital in standard “efficiency units,” and its use as a measure of the productive services from that asset (assumed to be proportional to the productive stock) (OECD, 2001).

22. This concept is different from net capital, as the latter measures the market value of the asset and thus captures its depreciation (loss of market value) rather than its deterioration (loss of productive capabilities) associated with aging. Productive capital is also different from gross capital as the latter only takes into account the withdrawal of the assets, and ignores the loss of productive capacity of those still in operation. Hence, gross capital stock can be interpreted as a special case of productive capital, where the physical efficiency of the asset remains intact over time and suddenly drops to zero when it is retired. While this is reasonable for certain types of assets, such as computers, for other types of capital “wear and tear” is a fact, and gross stock tends to overestimate the contribution of capital to production.

23. In Equation [1] capital input is represented by the productive services of five different assets: Plant and Machinery, Transport equipment, Building and Construction, Intangible Assets and Land.8 One important difference from previous estimates of capital stock for New Zealand is that capital input used in this study is a chain-linked measure of capital in constant prices. Chain-linking allows the removal of biases associated with fixed-weight volume measures, in particular the excessively large weights given to those IT capital assets whose relative prices have declined rapidly over the years.9

24. Each asset’s share of capital income in Equation [1] is obtained by estimating the assets rental prices, taken as a measure of their marginal productivity. Abstracting from tax considerations, the rental price for the asset j at time t is estimated as:


According to Equation [2], in an efficient capital market the return from renting one dollar of the asset ‘j’ must be enough to cover the opportunity cost of capital (r), the loss in the asset’s market value as it ages (δj, the depreciation rate for asset j), and the capital losses associated with a fall in the price of the asset (πj denotes the rate of change of the asset’s price pj).

25. The gross return on using an asset (r+δ-π) is higher for assets with relatively faster depreciation and larger negative capital losses, such as IT capital goods (computers within the category Plant and Machinery and software within the category Intangible Assets). The combination of high returns and falling prices should induce firms to substitute towards this type of capital. Weighting different assets through their rental prices aims at capturing this substitution process within the estimate of capital stock in Equation [1], as assets with higher rental prices receive a larger weight in Equation [1] compared to the case in which a homogeneous stock of capital is considered.

26. As for the labor input, an indicator of the compositional shift in capital (towards assets with higher productivity) is the difference between the weighted average of the assets rates of growth with weights given by the capital income shares and the weighted average of the assets rates of growth with weights given by the asset share of aggregate capital stock:


27. The growth of capital services (the first term) is higher than the growth of capital stock (the second term), if the accumulation of capital is biased in favor of assets with higher productivity. In the literature on productivity and the “new economy” (Jorgenson and Stiroh, 2000) the main objective of Equation [3] is to evaluate the compositional shift in capital accumulation that is associated with investment in IT assets. Unfortunately, Statistics New Zealand does not publish separate estimates of productive capital stock for IT hardware within Plant and Machinery, and this limits the extent to which the comparison of Equation [3] for New Zealand and Australia can indicate a different pace in IT capital accumulation.

28. That said, Table I.3 shows that New Zealand’s capital accumulation has suffered from a “quality gap” compared to Australia, especially in the second half of the 1990’s. Some features of the capital accumulation process in the two countries are also shown by Table I.3. The pace of capital investment in Australia has been higher than in New Zealand in the period considered, especially in Intangible Assets (mostly expenditure on mineral exploration, due to Australia’s large mining sector). Moreover, Australia has a larger share of net capital stock in Plant and Machinery and (especially) Building and Construction, while New Zealand has a much larger share of net capital represented by Residential Building.

Table I.3

A. Index of Capital Quality

article image
Source: Statistics New Zealand and Australian Bureau of Statistics.

Currency Conversion

29. A key difficulty in comparing productivity levels between New Zealand and Australia is translating real output and capital expressed in different currencies into common currency units. Market exchange rates are inappropriate for this conversion, as they fluctuate widely and in general do not reflect differences in countries’ real prices. In principle, local currency producer prices for specific goods should be compared (unit value ratios), and aggregated to build a sector specific conversion factor.10

30. In the absence of unit value ratios, the approach followed in this chapter has been to use the sector-specific expenditure Purchasing Power Parities produced by the OECD (1996). As these prices derive from retail surveys, they are adjusted for cross-country differences in trade margins and net indirect taxes, which should not affect the comparison (see Hooper, 2000).Table I.1.1 in Annex I.1 shows the sector-specific relative prices associated with the conversion factors used in this chapter.

C. Results

31. Once comparable estimates of output, labor, and capital (and their factor shares) have been obtained, it is possible to use Equation [1] to assess the relative contributions from capital accumulation and TFP in determining relative labor productivity in New Zealand and Australia. Table I.4 shows that:

Table I.4.

Levels Accounting: New Zealand’s Ratios with Australia, (Australia=1), by Sectors

article image
  • the contribution from human capital has been about the same in the two countries (on average in the period considered);

  • around ¾ of the 20 percent average New Zealand gap in labor productivity in the period 1988–1999 comes from differences in capital accumulation;

  • the other quarter of the gap is due to differences in TFP;

  • had New Zealand had the same capital intensity as in Australia, the labor productivity gap would have been reduced to around 7 percent on average in the period (point C in Figure I.2).

Figure I.2
Figure I.2

Labor Productivity, Capital Intensity and TFP in Australia and New Zealand (AUS=1), Averages 1988–1999.

Citation: IMF Staff Country Reports 2002, 072; 10.5089/9781451830231.002.A001

32. The existence of different capital intensities between the two countries reflects different relative factor costs. Figure I.3 shows that the estimate of the cost of labor relative to capital has been higher in Australia than in New Zealand in the period considered. While the two countries were at a relatively similar position at the start of the period, the cost of labor relative to capital has eventually become larger in Australia, especially over the second half of the 90’s. The higher relative wage growth in Australia was mainly responsible for the different pattern in the relative cost of factors.11

Figure I.3.
Figure I.3.

Cost of Labor Relative to Capital

Citation: IMF Staff Country Reports 2002, 072; 10.5089/9781451830231.002.A001

Cost of labor: Compensation of employees (from National Accounts, in 1996 NZ dollars) divided by hours worked.Cost of capital: Weighted average of the assets rental prices, (see Annex I.1), with weights given by the assets shares of capital income.

33. This difference could be due to the different timing of labor market reforms. While New Zealand started these reforms at the start of the 1990s by making welfare benefits less generous and deregulating the labor market through the 1991 Employment Relations Act, Australia’s labor market reforms started only in the second half of the 1990s. Given its relatively more flexible and cheaper labor force, New Zealand may have been induced to move towards relatively labor-intensive production technologies, compared to Australia (see also OECD, 2001).12

34. At the same time, however, other forces may be at play that explain the relatively faster capital accumulation in Australia. Table I.3 shows that a larger share of net capital is invested in residential buildings in New Zealand than in Australia. The share of New Zealand’s household portfolio invested in housing is one of the largest among OECD countries (OECD, 2001). Several factors may explain this bias, such as cultural attitudes and economic convenience (large capital gains from housing investment in the past). However, it is likely that the taxation regime, which exempts housing from any form of taxation, while taxing income and capital gains from other types of investments, has also played a role.13

35. Australian household portfolios are also mostly dominated by housing (Ellis and Andrews, 2001). However, the concentration of wealth in housing may have a larger impact on growth potential in New Zealand, because of the relative scarcity of savings available to finance more productive forms of investments in this country. The low household saving rate and the small size of private pension funds (whose size has decreased after the removal of tax concessions in the 1990s) in New Zealand are likely to reduce the proportion of housing assets that would be consistent with a given rate of output growth.14

36. A different explanation for the relatively lower capital deepening experienced by New Zealand focuses on the limited investment opportunities that would be open to countries with a small potential market (Skilling 2001). Skilling speculates that the small size of domestic market combined with the distance from foreign markets would make the “effective size” of New Zealand smaller than some critical mass required to have high growth performance. Nonlinearities in the production function associated with internal and external economies of scale would make it much more difficult for New Zealand’s small and geographically isolated firms to invest, export and grow. Accordingly, it is Australia’s relatively larger domestic market that ultimately explains the differences in both capital deepening and TFP growth.

37. This hypothesis is not entirely consistent with the sectoral analysis of productivity levels. For example, one sector in which New Zealand should be less affected by market size is the primary sector, which exports almost all of its product and in which New Zealand has a distinct comparative advantage. Table I.4 shows that in the agriculture fishing and forestry sector, New Zealand’s average gap in terms of labor productivity is associated with a much lower capital intensity than in Australia.15 This could reflect the difference in the structure of the two countries’ agricultural sectors, and in particular the combination of a cooperative structure and a single-buyer desk for export in New Zealand, which may have restricted capital investment in the sector.16

38. In contrast, almost all of the New Zealand’s relative gap in labor productivity in the manufacturing sector (where effective scale is more likely to play a role) is due to a lower level of relative TFP, while capital intensity has been broadly similar to Australia. This result contrasts with a recent paper by Fare, Grosskopf and Margaritis (2001), showing that New Zealand’s TFP record in manufacturing sector has been on average slightly better than Australia’s. The reason for this difference could be two-fold; first, this study uses the new chain linked series of capital stock; second, the final period of the analysis in Fare et al is 1996, and misses the period during which Australia’s IT capital accumulation accelerated.

39. Figure I.4 plots the time profile of relative TFPs of the different ANZSIC sectors and for the market sector as a whole. Eyeballing these charts shows that New Zealand has generally been catching up in sectors where the gap with Australia was larger at the start of the period (in particular Communication), and has instead lost ground in sectors where it had a relative advantage to start with. The only exceptions are Manufacturing and Construction, where New Zealand has failed to catch up with Australia. The divergence of aggregate TFP at a market sector level over the period considered in this study is thus primarily explained by the poorer performance of these two sectors.

Figure I.4.
Figure I.4.

Relative TFP Levels

Citation: IMF Staff Country Reports 2002, 072; 10.5089/9781451830231.002.A001

D. Resource Reallocation Across Sectors

40. The existence of significant differences in relative productivity levels across ANZSIC sectors suggests that there is scope for aggregate productivity gains through a reallocation of resources from low to high productivity sectors. Since one of the objectives of the microeconomic reforms implemented in the two countries in the last 15 years was to improve the economy’s allocative efficiency, it is interesting to assess to what extent Australia and New Zealand have differed in this respect.

41. As showed by Annex I.2, an indicator of an economy’s success in moving resources into sectors with higher returns is provided by the difference between the growth rate of aggregate TFP and the weighted average of sectoral TFP growth rates (with the weights given by the sector shares of aggregate output). This difference measures the contribution to aggregate TFP growth from a shift of labor and capital inputs to sectors where they earn a higher than average remuneration (that is, sectors with higher than average TFP levels). Alternatively stated, the market sector’s TFP growth rate can be expressed as the sum of two terms; one which reflects productivity growth within each sector, and one (RF) that depends on the reallocation of factors across sectors:


42. Table I.5 shows that in the period considered by this study, the size of the reallocation factors has been on average quite small in both countries. This may be due to the fact that the bulk of the reallocation of capital and labor had taken place in the 1980s (and thus before the period considered).17 Moreover, it is consistent with evidence that static allocative efficiency gains are generally low (Timmer and Szirmai, 2000). More significant, though, is the difference in the sign of the RF terms in the two countries, as the average annual RF is slightly negative in New Zealand but positive in Australia. Interestingly, the positive reallocation factor in Australia is entirely due to the reallocation of capital towards its most productive uses, which has more than offset the negative effect of the reallocation of labor.

Table I.5.

Reallocation Effect and Aggregate Market Sector TFP Growth

article image

43. Since New Zealand has undergone larger structural change than Australia since the mid 1980s,18 this difference could indicate that allocative gains in New Zealand may have been prevented by the limited access of resources to sectors with comparatively high relative productive levels. This hypothesis finds support in the decline of the share of capital invested in agriculture (one of the high-productivity sectors) in the period considered.19

44. Alternatively, these results may be suggesting the existence of structural factors that limit New Zealand’s ability to gain from policies aimed at improving allocative efficiency compared to Australia. Again, effective size may be playing an important role in this respect. As pointed out by Skilling (2001b), even if the New Zealand economy has been reasonably successful in diversifying away from agriculture, structural factors related to size and distance made a profitable shift from the primary sector into the manufacturing and services sector problematic. In these sectors, New Zealand lacks a comparative advantage and New Zealand firms face real difficulties in establishing a competitive advantage.

45. According to this hypothesis, New Zealand’s relative low growth performance is due to its having a comparative advantage in the “wrong” sector of the economy, namely, primary production. “Wrong” here means that this sector is characterized by lower growth prospects than manufacturing and services, where much of the technological progress in the 1990s has taken place and for which world demand has grown at a faster rate.20 Australia’s advantage over New Zealand is thus the relatively larger domestic market, which has made relatively easier for this country to move away from commodity production and to establish and sustain a competitive advantage in other areas of the economy.21

E. Conclusions

46. Using new and comparable data on output and capital stock in New Zealand and Australia this study shows that New Zealand per capita income divergence in the post-reform period has been primarily associated with a slower, quality-adjusted, physical capital accumulation process and, to a lesser extent, with lower TFP growth. The relative gap in the capital/labor ratio reflects differences in relative factor costs and is consistent with a smaller range of investment opportunities in New Zealand. New Zealand’s relatively lower capital deepening may also reflect the larger share of its net capital represented by residential buildings.

47. Although cultural attitudes are probably a large factor behind the large investment in housing, a role has also been played by economic factors, namely the significant (in the high inflation environment of the 1990s) and untaxed capital gains and the absence of taxation on the imputed rental income derived from this type of investment. In terms of policy implications, the analysis above suggests that some consideration could be given to reforms of the tax system that would “level the playing field” for different vehicles of savings.

48. The study also shows that New Zealand’s relative efficiency in the utilization of primary inputs has been declining in the period after the reforms. The divergence of aggregate markets sector TFP is ultimately caused by divergence in manufacturing, while New Zealand has been catching up Australia in some of the service sectors, where it had a relative productivity gap at the end of the 1980s, e.g., communication.

49. The analysis also cast doubts on the notion that allocative efficiency gains from the reforms have significantly boosted New Zealand’s TFP growth. While structural factors, in particular, the small size of domestic market and protective trade barriers limiting scale and investment opportunities, may be playing a role, there are also “distortions” that have reduced the effectiveness of structural changes in enhancing productivity growth. In particular, the cooperative structure and the monopsonistic nature of a significant part of the New Zealand’s most efficient sector (agriculture), and one which suffers least from scale disadvantages, may have prevented New Zealand from fully benefiting from its main comparative advantage.

ANNEX I.1: Methodological and Measurement Issues

Productivity level comparisons of the type presented in the chapter require developing comparable measures of output and input levels. As results can be quite sensitive to differences in estimation procedures, extreme care should be devoted to obtaining comparable indicators for the two countries, even if they are not necessarily the best measure for each individual country. This Annex briefly outlines some of the problems encountered in this process and outlines the methodology used to get around them.


Sectoral output is defined as gross value added adjusted to a factor-cost basis. The valuation at factor costs amounts to exclude all indirect taxes on products (for example, GST, excise duties, import duties) and on production (for example, levies), and all subsidies. Hence, the value added concept considered in this study is the sum of compensation of employees and gross operating surplus from the National Accounts. Since sectoral value added is available in chain volume terms at basic prices for Australia and at producer prices for New Zealand, the annual rates of change from these series have been applied to GDP at factor costs in 1995/96 prices.

A difference between the two countries involves bank service charges, as Statistics New Zealand does not deduct these charges from the output of individual sectors. The 1995/96 value added for New Zealand’s ANZSIC sectors is thus adjusted by using the 1995/96 New Zealand input-output table, which shows each sector’s intermediate consumption of bank service charges.


An important step in comparing productivity levels is the conversion of the two countries’ output and capital into a common currency. The approach in this chapter is to use PPP exchange rates. Using a single, economy-wide PPP exchange rate, however, would amount to ignoring variations in relative price levels across sectors. Hence, whenever possible this study applies industry-specific PPP exchange rates calculated by the OECD (see OECD, 1996 and 1993). These rates are obtained as weighted averages of national relative prices evaluated at a retail level and aggregated using expenditure shares.

The OECD 1996 price comparison covered around 2500 goods and services that have been allocated into 65 expenditure categories. While in some cases the match between these expenditure categories and the sectors considered in this study is straightforward (as for transport services, accommodation cafes and restaurant, and cultural services), for the other sectors the OECD PPP rates were used as a proxy. This was the case of the PPP rate for fuel and power, used for the electricity gas and water sector, while for manufacturing this study uses the weighted average of PPPs rates for the following categories: food, beverage and tobacco; clothing and footwear; household equipment; recreational goods; books, newspapers, and other printed matters; transport equipment; miscellaneous goods, and machinery and equipment (with weights given by the goods’ shares of total expenditure). For retail and wholesale trade sector, the PPP rate for total consumption was used. For all other sectors, in the absence of a better conversion factor, the PPP rate for the whole GDP was used. The only exception is for communication, for which the relative price level is obtained from the Australian Productivity Commission (1999). The conversion rates for capital assets were also derived from the OECD expenditure-based PPP rates.

The conversion rate for sector i (to be applied to magnitudes in Australian dollar) are obtained as ratios between the expenditure prices PE:


As these prices include distribution margins, the PPPs conversion rates used in this study are adjusted for the cross country-differences in these margins:


where δ are the distribution margins obtained as the ratio between distribution margins and total supply purchaser prices (from the Australian and New Zealand 1996 input-output tables). A further correction is made to take into account cross country-differences in net indirect taxes:


where t are the ratios between net taxes on products and total supply purchaser prices. Table I.1.1 shows the price level indices used in this study (defined as the ratio between EPPP* and the 1996 market exchange rate).

Table I.1.1.

New Zealand/Australia: Relative Price Levels 1/


article image

Estimated from Purchasing Power Parities and Real Expenditure, OECD 1993 and 1996.

Relative prices of communication services from International Benchmarking of Telecommunication Services, New Zealand Summary, April 1999, Productivity Commission. Canberra, Australia.

Labor Input

The comparability of hours worked is hindered by the adoption of different data collection and processing procedures across the two countries. The Australian Bureau of Statistics publishes a measure of hours worked by ANZSIC sector based on the quarterly Labor Force Survey (LFS). Hours worked are measured as total hours actually worked during a reference week of the quarter by all those employed (employees, employers and self-employed). Hence, hours paid for but not worked (because of annual leave, sickness, holidays) are excluded from the estimate.

In New Zealand, there are two main sources of hours worked by sector. The first one is the Quarterly Employment Survey (QES), which records the hours paid for, not the actual hours of work. As this is a business survey, the statistic units are business locations with at least one full-time equivalent employee (one full time or two part-timers). This means that those self-employed who do not employ staff (own-account workers, in the definition of the ABS) are outside the scope of the survey. Finally, the survey does not cover Agriculture and Mining.

The other source of hours worked by sector is the quarterly Household Labor Force Survey (HLFS), which provides data on actual hours worked and covers all sectors. Even if the QES has a greater level of sectoral detail (based on an ANZSIC classification), for the sake of comparability this study adopts the estimates of hours worked from the HLFS, as reported in the Diewert and Lawrence database, and uses data on hours paid from the QES to match the ANZSIC classification.

Hours Worked, New Zealand

  • Agriculture, fishing and forestry: total hours worked (sum of 4 quarters) from the HLFS, as in the Diewert and Lawrence (DL) database, Table C7;

  • Mining and quarrying: as above;

  • Manufacturing: as above;

  • Electricity, gas and water supply: as above,

  • Construction: as above;

  • Wholesale trade: obtained applying the share of wholesale trade of total hours worked in the sectors wholesale trade, retail trade, and accommodation cafes and restaurant as in the QES to the aggregate “Trade, Restaurant and Hotels” from the HLFS as in the DL database;

  • Retail trade: as above;

  • Accommodation, cafes and restaurants: as above;

  • Transport and storage: obtained applying the transport share of the total hours worked in the sectors transport and communication in 1996 from HLFS to the aggregate “Transport and Communication” sector in the HLFS as in the DL database; (this amounts to assume that before 1996, around 70 percent of the total hours worked in the sector “Transport and Communication” was attributable to the transport sector, as in 1996);

  • Communication services: as above;

  • Finance and insurance: obtained applying the finance and insurance share of the total hours worked in finance and insurance and property & business services from the QES to the aggregate “Financial Services” in the HLFS as in the DL database; and

  • Cultural and recreational services: obtained applying the cultural and recreational services share of the total hours in the sectors cultural and recreational services, health & community services, personal & other services and education from the QES to the “Community Services” aggregate in the HLFS as in the DL database.

The major difference between the two surveys is their timing, as the Australian Labor Force Survey refers to a reference week of the quarter, while the HLFS carries out the interviews each week and thus refers to a weekly average for the quarter. Hence, the Australian series is more exposed than the New Zealand’s one to the noise arising from public holidays and other days lost during the reporting period.

An adjustment for different quality of labor input is made by using data on employment by educational attainment in different sectors of the economy. For New Zealand, the data on educational attainment by sector have been derived from the Census 1981, 1986 and 1996. For Australia, the data on educational attainment by sector for the year 1984 and 1997 are those published in Barnes et al (1999). Data for intermediate years are obtained by cubic interpolation.

The wage shares of the three types of labor considered (with tertiary qualification, with secondary qualification, and without secondary school qualification) have been estimated by using the relative wages of these three groups as reported in the OECD publication “Education at a glance.”

Capital Input

Productive Capital Stocks

Equation [1] in the text includes the productive stock of five different capital assets: Transport equipment, Plant and Machinery, Non Residential Building and Constructions, Intangible assets and Land. These stocks are obtained from the official statistical offices, with the only exception of the stock of Land used for productive purposes for New Zealand that is taken from Diewert and Lawrence (1999). All series are in chain volume terms and are available for each ANZSIC sector.

Using capital inputs produced by the official statistical offices may be a source of biases in international productivity comparison if the methodologies followed by the national offices are very different. This does not seem to be the case for SNZ and ABS, however, because they follow a reasonable similar methodology in estimating productive capital stocks. In particular, both statistical offices use the same hedonic prices to deflate investments in computers and the same age efficiency reduction parameters utilized by the U.S. Bureau of Labor Statistics. The assets’ average service lives used in the perpetual inventory model are also quite similar, with the only exception of Transport Equipment, whose average life is shorter for New Zealand than for Australia.22

The different number of assets for which official estimates of the productive capital stock are available in the two countries, raises a series of aggregation issues. For instance, while Statistics New Zealand publishes the productivity capital stock for Plant and Machinery as a whole, the Australian Bureau of Statistics publishes estimates of four different capital assets within this group (Computers and peripherals, Industrial machinery and equipment, Electrical and electronic equipment, Other plant and equipment). As the SNZ’s estimate of Plant and Machinery is the sum of these distinct assets, for the sake of comparability the stock of Plant and Machinery for Australia is obtained by adding the stocks of its four components.23 Summing the productive capital stock of different assets, however, amounts to ignoring the differences in the proportional factor that links the stocks of the asset to its productive services.

Rental Prices

The rental price of asset j in sector i at time t is estimated according to the equation:



pi,j,t is measured as the investment deflator from Statistics New Zealand and the Australian Bureau of Statistics;

πi,j,t is the rate of change of pi,j,t;

δi,j,t is the depreciation rate of the asset i in sector j calculated as follows:


where NKS refers to the chain volume net capital stock and I to chain volume investments (both obtained from the two national statistical offices);

τt is the corporate tax rate;

zj,t is the present value of depreciation allowances for asset j, calculated as follows (see Moes, 1999):


where fj,t is the depreciation tax loading factor (taken from Moes (1999) for New Zealand and the Australian Taxation Office for Australia) and ρt is the discount rate (10 years government bond rate).

Finally, the opportunity cost of capital r can be estimated in two ways. According to the first one, a sector-specific rate ri,t can be derived as the rate that makes the aggregate income from capital services (rental price times capital stock) equal to the aggregate capital income estimated from the National Accounts (gross operating surplus):


where GOSj,t is the current price gross operating surplus for sector j at time t obtained from the NA, Ki,j,t the productive capital stock for the asset i in sector j at time t, and all other variables are obtained as above. As GOS is obtained residually (as the difference between value added and compensation of employees), adopting this methods implies to assume constant returns to scale.

The other method consists in estimating an economy-wide nominal cost of capital from market rates. A vast choice of market rates is available, and possible options are the use of an average of borrowing and lending rates, government bond rates, corporate bonds rates, or E/P ratios (see Hsieh, 2000).

In this chapter the first method (constant return to scale) is applied, but the second one is also explored (using the interest rate on 10 years government bonds) as a way to implicitly assess the distortions associated with imposing constant returns to scale. Alternatively stated, the absence of a significant differences between the dynamic of factor returns from the National Accounts and those estimated from market data is taken as an implicit “dual” confirmation of the TFP estimates obtained under constant returns to scale.

Table I.1.2.

Estimated Rented Prices of Capital

article image

Factor Shares

The labor income shares αL for the three types of labor considered in Equation [1] in the text are obtained as follows:


where the first ratio on the right hand side is the sector-i compensation of employees as share of sector-i value added, and the second term is the share of sector-i compensation of employees that is attributed to the type of labor denoted by j, e.g., tertiary education. wi,j,t is the relative wage of workers j (normalized to the wage of those with secondary education), as derived from OECD’s Education at Glance.

The labor income share for sector-i has been adjusted to take into account that the income of self-employed, part of which is remuneration for labor, is included in the gross operating surplus. Following OECD (2000), part of the gross operating surplus is thus allocated to labor assuming that self-employed have the same average remuneration than employees (the share of self-employed over total employed by ANZSIC sectors is obtained from the labor force surveys).

As for the capital income shares αK in Equation [1] in the text, they are obtained as follows:


where the asset rental prices μi,j,t are obtained as described above, assuming constant returns to scale.

ANNEX I.2: Reallocation Factor

The growth rate of market sector’s TFP is:


where VA is the market sector’s value added, L is the market sector’s hours worked and K the market’s sector capital stock.

Aggregate TFP can be obtained also as the weighted average of sector i TFPs (i=l…S), with weights given by the sector i share of value added:


To explain the meaning of the difference between Equation [1] and Equation [2], Equation [1] must be expressed in a different way. Using the equivalencies:


Equation [1] can be written as:


Using the following equivalencies:


Equation [3] can be expressed as:


The difference between Equation [2] and Equation [4] can thus be expressed as follows:


The difference between the aggregate TFP growth rate and the weighted average of sectoral TFPs is positive if labor and capital inputs increase in those sectors with a higher than average return (and, thus, with higher TFP). Denoting the term on the right hand side with RF (which stands for reallocation factor), it is possible to express the rate of growth of aggregate TFP as the sum of two terms, one that reflects the growth of TFP within each sector, and one that reflects the ability of the economy to move resources towards sectors with higher productivity:



  • Barro, Robert J., 1999, “Notes on Growth Accounting,NBER Working Paper, No. 6654.

  • Barnes Paula, Rick Johnson, Anthony Kulys and Scott Hook, 1999, “Productivity and the Structure of Employment,Staff Research Paper, Productivity Commission, Canberra.

    • Search Google Scholar
    • Export Citation
  • Basu Susanto and Fernald John, 1995, “Returns to Scale in U.S. Production: Estimates and Implication,International Finance Discussion Paper, Board of Governors of the Federal Reserve System, No. 546.

    • Search Google Scholar
    • Export Citation
  • Briggs, Phil, Phil Bishop and Emma Fan, 2001, “New Zealand Economic Growth: Why has it been so low?,NZIER Working Paper, 01/02.

  • Buckle, Bob, David Haugh and Peter Thompson, 2001, “Calm After Storm? Supply Side Contributions to New Zealand’s GDP Volatility Decline,NZ Treasury Working Paper, 01/33.

    • Search Google Scholar
    • Export Citation
  • Claus Iris, David Haugh, Grant Scobie and Jonas Tornquist, 2001, “Saving and Growth in an Open Economy,NZ Treasury Working Paper, No. 01/32.

    • Search Google Scholar
    • Export Citation
  • Cummings, Diane, Lauritis Christensen and Dale Jorgenson, 1981, “Relative Productivity Levels, 1947-1973: An International Comparison,European Economic Review, Vol. 92.

    • Search Google Scholar
    • Export Citation
  • Diewert, Denis and Erwin Lawrence, 1999, “Measuring New Zealand’s Productivity,NZ Treasury Working Papers, 1999/5.

  • Ellis Luci and Dan Andrews, 2001, “City Sizes, Housing Costs, and Wealth,Reserve Bank of Australia Discussion Paper, 2001-08.

  • Fare, Rolf, Shawna Grosskopf and Dimitri Margaritis, 2001, “Productivity Trends in Australian and New Zealand Manufacturing,The Australian Economic Review, Vol. 34, No. 2.

    • Search Google Scholar
    • Export Citation
  • Hall, Viv, 1996, “Economic Growth,” in A Study of Economic Reform: The Case of New Zealand, ed. by B. Silverstone, A. Bollard and R. Lattimore, Elsevier.

    • Search Google Scholar
    • Export Citation
  • Hall, Robert E. and Charles I. Jones, 1996, “The Productivity of Nations,NBER Working Paper, No. 5812.

  • Harrigan, James, 1997, “Cross-Country Comparisons of Industry Total Factor Productivity: Theory and Evidence,Federal Reserve Bank of New York Research Paper, No. 9734.

    • Search Google Scholar
    • Export Citation
  • Hooper, Peter, 1996, “Comparing Manufacturing Output Levels Among the Major Industrial Countries,” in Industry Productivity: International Comparison and Measurement Issues, OECD.

    • Search Google Scholar
    • Export Citation
  • Hsieh, Chang-Tai, 2001, “What Explains the Industrial Revolution in East Asia? Evidence from the Factor Markets,forthcoming, American Economic Review.

    • Search Google Scholar
    • Export Citation
  • Hulten, Charles R., 2000, “Total Factor Productivity: A Short Biography,NBER Working Paper, No. 7471.

  • Jorgenson Dale and Zvi Griliches, 1967The Explanation of Productivity Change,The Review of Economic Studies, Vol. 34, No. 99.

  • Jorgenson, Dale and Kevin Stiroh, 2000, “Raising the Speed Limit: U.S. Economic Growth in the Information Age,Brookings Papers on Economic Activity, No. 1.

    • Search Google Scholar
    • Export Citation
  • Lally, Martin, 2000, “The Real Cost of Capital in New Zealand: Is It Too High?,New Zealand Business Roundtable.

  • Lewis Evans and Neil Quigley, 2001, “Watershed for New Zealand Dairy Industry,NZ Institute for the Study of Competition and Regulation (ISCR), Monograph Series, No. 1.

    • Search Google Scholar
    • Export Citation
  • Little, Struan, 2001, “Growth and Policy in Other Countries: Lessons from the Losers,NZ Treasury.

  • McLean Ian and Alan Taylor, 2001, “Australian Growth: A California Perspective,NBER Working Paper, NO. 8408.

  • Moes, Alowin, 1999, “Effective Tax Rates on Capital in New Zealand—Changes 1972-1998,NZ Treasury Working paper, No. 99/12.

  • Organization for Economic Co-Operation and Development, 1996, “Purchasing Power Parities and Real Expenditures,Paris.

  • Organization for Economic Co-Operation and Development, 2000, “OECD Productivity Manual: A Guide to the Measurement of Industry Level and Aggregate Productivity Growth.

    • Search Google Scholar
    • Export Citation
  • Organization for Economic Co-Operation and Development, 2001, “2000/01 Annual Review-New Zealand,Paris.

  • Productivity Commission, 1998, “Aspects of Structural Change in Australia,Research Paper, Productivity Commission, Canberra.

  • Purdue, Donna, 1999, “Productivity Growth and Industrial Structure: The Case of New Zealand,NZIER, mimeo.

  • Scarpetta Stefano, Andrea Bassanini, Dirk Pilat and Paul Schreyer, 2000, “Economic Growth in the OECD Area: Recent Trends at the Aggregate and Sectoral Level,ECO/WKP 21, OECD.

    • Search Google Scholar
    • Export Citation
  • Skilling, David, 2001a, “The Importance of Being Enormous—Towards An Understanding of the New Zealand Economy,NZ Treasury.

  • Skilling, David, 2001b, “Comparative Disadvantage: New Zealand’s Economic Growth in Historical Perspective,NZ Treasury.

  • Sinclair, Geoff, 1999, “Costs and Benefits of Producer Board Deregulation,NZ Treasury Working Paper, No. 99/4.

  • Statistics New Zealand, 1998, “Chain Volume Measures in the New Zealand National Accounts.

  • Timmer Marcel and Adam Szirmai, 2000, “Productivity Growth in Asian Manufacturing: the Structural Bonus Hypothesis Examined,Structural Change and Economic Dynamics, Vol. 11

    • Search Google Scholar
    • Export Citation
  • Van Ark, Bart, 1996, “Issues in Measurement and International Comparison Issues of Productivity—An Overview,” in Industry Productivity: International Comparison and Measurement Issues, OECD.

    • Search Google Scholar
    • Export Citation

Prepared by Roberto Cardarelli (x38059), who is available to answer questions.


See Hulten (2000) and Barro (1999) for an extensive discussion of TFP and Solow’s residual.


As the main focus of the chapter is on average magnitudes over the period considered, no adjustment is made to filter out business cycles effects. Moreover, the period 1988–1999 almost fully coincides with a peak-to-peak cycle for both countries.


Other alternatives have been used in the productivity literature in the context of a multi country framework, as the one that compares each country with the average of all countries (Cummings, Christensen and Jorgenson, 1981). Studies that have compared these two methods arrived at the conclusion that they produce similar results (see Harrigan, 1997, and Jones and Hall, 1996).


These assumptions are similar to those used in Scarpetta, Bassanini, Pilat and Schreyer (2000).


Little (2001) refers to the low share of engineers in graduates in New Zealand (6 percent of total college graduates, against around 20 percent in Finland). From this result, he notes that, while aggregate data could lead to a relatively sanguine view around human capital, at a more disaggregated level there are more worrying questions on whether the structure of the New Zealand education system is sufficiently directed to growth-oriented activities.


It could, of course, be argued that the larger proportion of workers with vocational secondary education in Australia signals a stronger connection between the education system and the specific skill requirements of the production process.


These assets capture around 90 percent of the total productive capital stock for the market sector in Australia (including livestock and inventories). Using the estimates of livestock and inventories as in Diewert and Lawrence, a similar percentage is found for New Zealand.


Over 1988–1998, chain-linked gross capital formation in New Zealand grew on average 0.5 percent less per year than the previously published, fixed weighted figures (see Statistics New Zealand, 1998).


Unit-value ratios are the ratios between values and volumes of different goods as reported in production statistics. However, using these ratios also has its limits, as they do not reflect differences in product quality across countries, and usually allow coverage of only a very limited sample of goods (Van Ark, 1999).


The higher estimated rental price of capital in New Zealand is a consequence of its smaller stock of capital (and, thus, larger marginal productivity). However, the same result holds if adopting market-based measures of the rental price of capital, for example, by estimating the opportunity costs of capital r in Equation [2] with the return on long-term government bonds (see Annex I.1). Adopting a CAPM approach, Lally (2000) also confirms that the real cost of capital for a typical firm in New Zealand has been modestly larger than in Australia, and considerably higher than in the United States. A critical factor behind this result is the allowance for currency risk, and a relatively higher market risk premium in New Zealand’s market.


Labor market reforms implemented in New Zealand had obvious beneficial effects in terms of inducing higher employment and of enhancing the resilience of New Zealand’s economy in face of adverse shocks. Moreover, the potential negative impact of labor market reforms on capital deepening may be a temporary phenomenon that does not reflect long-run trends.


New Zealand does not have a comprehensive capital gain tax, but the income from any asset held with the purpose of resale is taxable. For example, equities transactions involving managed funds are taxed, while individuals can hold and trade (within certain limits) equities without being taxed.


It is widely perceived that the relative low level of domestic savings has not been a constraint on New Zealand’s economic growth, as New Zealand has been able to access foreign savings to meet investment demand (Claus and others, 2001). It may be argued, though, that the heavy dependence on foreign capital is one of the factors that contributed to a relatively higher capital costs in New Zealand compared to Australia, as mentioned in Footnote 11.


It should be stressed, however, that the sectoral results shown by Table I.4 should be taken with some caution, for at least two reasons. First, as noted above, this study does not consider livestock as a capital assets, as such data do not exist for New Zealand. Because the relatively larger diary sector in New Zealand implies that this country is likely to use relatively more livestock than Australia, the capital deepening gap in the agricultural sector may be overstated. Second, the estimates of TFP at a sectoral level suffers from some potentially important sources of measurement errors, such as the one associated with the use of expenditure PPPs and that arises from the monopsonistic nature of the diary sector in New Zealand. Under the latter, the allocation of total value added in the diary sector across raw milk production (agriculture), processing (manufacturing) and marketing (wholesale trade) is likely to have been distorted relative to outcomes in a competitive market structure.


The fact that equity capital can only be raised by suppliers limits the source of capital for cooperatives relative to traded corporations. As stated by Lewis and Quigley (2001), this is an important issue where profitable opportunities for expansion exists. See also Sinclair (1999).


Buckle and others (2001) show that substantial changes in output sector shares had ceased by the beginning of the 1990s. Clearly, most of the rationalization process induced in New Zealand by the elimination of assistance and protection polices took place at firms levels within each sectors (Savage and Bollard, 1990, as quoted in Skilling, 2001). This suggests that the sectoral level of the analysis above does not capture adequately the static efficiency gains involved in the adjustment process that followed the reform program.


This results from extending to the 1990s the measurement of structural changes in Australia and New Zealand based on a structural change index as in Productivity Commission (1998).


Hall (1996) also finds that New Zealand employment shift in the post reform period (1985–1993) have been into sectors with lower rather than higher output and labor productivity growth. Purdue (1999) finds that during the 1990s the distribution of output by sectors changed so as to produce only a small (but positive) effect on aggregate productivity.


Briggs, Bishop and Fan (2001) find that New Zealand’s relative low export growth in 1985–1998 is due to its comparative advantage in primary sectors, which have shown relatively weaker growth over this period (partly because of still high degree of protectionism).


Mc Lean and Taylor (2001) emphasize another structural advantage of Australia, namely, its larger endowment of mineral resources (that accounts for almost 40 percent of its exports). Absent this advantage, they speculate that the postwar growth trajectory of Australia may have looked similar to that of New Zealand.


According to SNZ, such diversity reflects the fact that New Zealand imports a significant larger amount of second-hand cars (with a shorter mean life) than Australia (where import protections are more stringent).


As a general rule, this chapter ignores the bias arising from the non-additivity of chain volume figures, which is likely to be of a second order magnitude.

New Zealand: Selected Issues
Author: International Monetary Fund