Regional economic performance in New Zealand: how does Auckland compare?
Auckland is New Zealand's largest urban centre of population and economic activity--comprising around one third of the total on each measure. It is almost axiomatic therefore that for the New Zealand economy to perform well, Auckland needs to perform well. Over and above the fact that Auckland is a large fraction of New Zealand, there is also a substantial body of economic theory and evidence that finds that large cities can be dynamic sources of innovation and productivity growth. (1) To what extent is this true of Auckland? This paper reports new evidence on how well the Auckland economy has been performing relative to other cities and regions of New Zealand.
Any investigation of regional economic performance in New Zealand is hampered by the relative paucity of official statistics at the regional level. For example, there are no data on the levels and growth rates of regional gross domestic product, or regional labour productivity. (2) However some nongovernmental organisations have developed proxies for regional economic performance in New Zealand and report them quarterly or less frequently.
The National Bank of New Zealand produces quarterly estimates of the growth of regional economic activity. It uses a variety of official and unofficial sources to construct an index of overall economic activity for each of 14 local authority regions of New Zealand, but does not attempt to estimate labour productivity growth rates or levels by region. The general picture that emerges from its quarterly reports on changes in the indices and their components is that Auckland's growth performance at least over the last business cycle has been no better than average.
The New Zealand Institute of Economic Research (NZIER) has produced estimates of regional labour productivity (NZIER, 2004). Its method assumes that average labour productivity for each industry at the national level is constant across regions and therefore that differences in productivity performance across regions and over time simply reflect differences and changes in industry structure across regions. This is a restrictive assumption because it rules out, for example, that productivity in the business services sector could be growing faster (or slower) in Auckland than in Christchurch. This study finds Auckland's economic performance over the early 2000s is a mixed bag with most measures very much sitting in the middle part of their regional distributions.
The research reported in this paper attempts to overcome some of the problems with these existing proxy measures of regional economic performance. It examines wage and income data from the annual New Zealand Income Survey (IS). According to human capital theory, hourly earnings of workers are indicative of their labour productivity, while real earnings growth reflects labour productivity growth. We divide the country into five areas that are broadly homogenous in terms of where they sit on an urban-rural spectrum. These areas are the cities of Auckland, Wellington, and Christchurch, a composite of medium-sized cities and a composite of rural areas and small centres. Then, we examine wage and income levels and trends from 1997 to 2004 across these areas. In contrast to the National Bank and NZIER findings over a comparable period, our results suggest that productivity levels are generally higher in Auckland and have been growing faster than in other regions.
In section 2, we provide background and explain the motivation for the research. Section 3 describes our data, while section 4 presents our results. Section 5 compares our results with other studies of regional economic performance and tries to understand and explain the differences. Section 6 summarises and concludes.
2. Background and Motivation
New Zealand's Economic Performance
New Zealand's overall economic performance picked up in the early 1990s compared with previous decades. From 1994 to 2004 GDP grew at an annual average rate of 3.4 per cent. GDP per capita over a recent 6-year period (1997-2003) grew at an annual average rate of 2.56 per cent compared with an OECD average of 2.10 per cent (Ministry of Economic Development and Treasury, 2005). Nevertheless the level of New Zealand's GDP per capita remains low in the OECD rankings in twentieth place. Moreover, the relatively high GDP per capita growth has been driven more by fast growth in labour utilisation than by good labour productivity growth. New Zealand's level of labour productivity is low by OECD standards and its growth rate is also below average (Ministry of Economic Development and Treasury, 2005).
The scope for further per capita growth through achieving a higher fraction of the population in paid work is limited--New Zealand has one of the highest level of labour utilisation in the OECD. This has been achieved through one of the lowest unemployment rates in the OECD and above average levels of labour force participation. Therefore the main challenge for New Zealand in order to maintain and enhance its per capita growth rate and achieve higher relative levels of GDP per capita must lie largely in achieving higher labour productivity growth.
It is difficult to find clear answers to explain New Zealand's relatively slow labour productivity growth. On most policy settings and in its institutional architecture, New Zealand measures up well compared to other OECD economies. It may be that slow labour productivity growth is a by-product of fast employment growth--both because most growth in employment has been at the lower end of the skill distribution and because firms have expanded output more through hiring labour than by investing in plant and equipment and the modern technology embodied in them (McLellan, 2004; Hyslop and Yahanpath, 2005).
Another hypothesis is that New Zealand struggles to achieve higher productivity because of the small size of its domestic market and its distance from markets in other countries. These factors limit competition, create higher transport costs and other barriers for exporters, and slow down technology adoption. The economic geography literature that highlights size and distance as influencing factors in economic performance focuses mainly on so-called agglomeration economies. (3) These arise either from clusters of firms within an industry (the Silicon Valley syndrome) or the dense co-location of a variety of industries and skilled people in cities.
Some commentators point to cities that stand out as centres of creative innovation and high value-added industries, places such as Boston, San Francisco, London, Helsinki, Dublin, Amsterdam, Singapore, Sydney and Melbourne. In terms of size, Auckland is on the cusp of having the population that is typical of some of these successful smart cities. A question for policy analysts interested in New Zealand's economic performance is whether Auckland is playing the energising role of one of these dynamic cities within the national economy? (4)
The key motivation for the research reported in this paper is to throw more light on the question of how far Auckland is performing a dynamic role within the national economy. We do this by examining how well it has performed relative to other regions in New Zealand. Specifically we examine whether Auckland is performing above or below average in labour productivity levels and growth rates compared to smaller cities in New Zealand, and to small centres and rural areas. In asking this question, we need to bear in mind that New Zealand's economic structure is different from most other developed OECD economies--it has a relatively large, export-based primary sector. This characteristic could mean that it is New Zealand's rural areas rather than its cities that drive overall economic performance.
This study takes a previously unexplored route to measuring Auckland's relative economic performance. It focuses on productivity performance by analysing the wage rates for wage/salary and self-employed workers, and their growth across regions in New Zealand. We make the standard economic assumption that pay rates reflect marginal labour productivities of the relevant workers. We also look at employment rates, rates of benefit receipt and hours worked as indicators of relative labour utilisation performance in Auckland compared to other cities and regions.
There are perceptions that Auckland's economy is underperforming, and thus is not fulfilling the role of a dynamo for the rest of the New Zealand economy (LEK, 2001). It is possible that the National Bank's proxies for regional economic activity growth have contributed to these perceptions. For example, a recent leaflet put out by the Auckland Regional Council (ARC, 2005) used the National Bank's data to illustrate that Auckland's growth in 2004 was the 4th lowest among fourteen New Zealand regions. However, one needs to keep in mind that overall growth in average income in a region reflects both labour productivity and labour utilisation. The National Bank measures do not contain enough information to uncover Auckland's relative performance in either measure on its own.
While New Zealand's regions vary according to how urban or rural they are, most are a mixture. Even the Auckland region includes a significant hinterland of smaller centres and rural land beyond the city boundary. In order to capture whatever differences in labour productivity exist between densely populated urban areas and other parts of New Zealand, we divided households in the Income Survey into five areas: the Auckland urban area; the Wellington and Kapiti urban area, the Christchurch urban area; a composite of all other main urban areas and a composite of minor urban centres and rural areas. This is a significantly different geographical division compared to the local government regions of New Zealand, which are the focus of both the National Bank and NZIER measures.
As already mentioned, the NZIER study estimates regional labour productivity statistics by assuming that average labour productivity for each industry at the national level is constant across regions and that differences in productivity performance across regions and over time simply reflect differences in industry structure across regions and over time. If Auckland and other New Zealand cities generate the sort of productivity effects that evidence indicates occur in cities in other countries, then the NZIER approach will at best capture them only partially. For the first time, our study provides an insight into labour productivity levels and growth rates across New Zealand, where allowance is made that these may differ within an industry across the country.
It is possible that Auckland could be underperforming compared to other regions in GDP per head, but performing above average in terms of labour productivity. This could occur, for example, if Auckland has experienced a low level and low growth rate of labour utilisation compared to other regions. Thus, it is important to examine indicators of labour utilisation as well as proxies for labour productivity in order to understand the underlying differences across regions and over time. (5)
New Zealand Income Survey
Since 1997, the Income Survey (IS) has been carried out by Statistics New Zealand (SNZ) each June quarter as a supplement to the Household Labour Force Survey (HLFS). Taken together, the two surveys collect data on household structure, the socio-demographic characteristics of household members, and labour force activity in the reference week and recent incomes for individuals at least 15 years-old. The IS collects information on sample members' actual pre-tax weekly income in the reference week for the survey, and uses a series of detailed questions on each component of income in an effort to ensure that the final estimate of total income (from all sources) is as accurate as possible. (6) The HLFS has a sample size of approximately 15,000 households and 28,000 adults. (7) About 85% of these respondents also complete the IS. (8) Sampling weights are calculated by SNZ to increase the representativeness of the HLFS, and are used in all analyses in this paper unless noted.
The IS also records the geographical location of all interviewed households. This is coded in two separate variables: the first records the Local Government Region (LGR) of each household; (9) the second records whether the household is in a Main Urban Area, Secondary Urban Area, Minor Urban Area, or Rural Centre and other rural areas. (10) These variables can be used together to uniquely identify most main urban areas. For example, Auckland is the only main urban area in the Auckland LGR and thus all households located in the Auckland urban area are identified in the IS. In contrast, both the Wellington and Kapiti urban areas are in the Wellington LGR and thus it is not possible to differentiate in the IS between households located in these two urban areas. (11)
The sample for all analyses in the paper is adults aged 25-59 years (prime-age) with non-imputed income data in the 1997-2004 IS. This restricted age range drops most individuals who are studying or retired. As discussed in previous sections, this paper examines the economic performance of the Auckland urban area (population 1,074,507 in 2001 census) in comparison to other areas of New Zealand. Four comparison areas are focused on in all analyses: 1) the Wellington and Kapiti urban areas (population 373,416 in 2001 census), 2) the Christchurch urban area (population 334,104 in 2001 census), 3) a composite of all other main urban areas (population 872,823 in 2001 census), (12) and 4) a composite of all rural areas and minor urban areas (population 845,778 in 2001 census). Secondary urban areas are excluded from the comparison groups; these are large rural towns and it was unclear whether to include these in the other urban or rural comparison area. (13) Appendix Table 1 displays the sample size for each area in each IS year.
A subset of the analyses in this paper examines economic performance within the Auckland urban area. Auckland is divided by SNZ into four zones (see Appendix Figure 1), Northern (population 219,936 in 2001 census), Western (population 173,643 in 2001 census), Central (population 359,469 in 2001 census), and Southern (population 321,465 in 2001 census), which can be identified in the IS data. (14) Appendix Table 2 displays the sample size for each zone in each IS year.
Measures of Economic Performance
A range of measures of economic performance are examined in order to capture differences in both productivity and labour utilisation between Auckland and other areas in NZ (and within Auckland). The analyses in this paper focuses on six measures. (15)
The first three measure overall economic performance (income) and workplace productivity (wages) in each area:
(1) Real Annual Labour Income, which is the sum of actual income from wage/salary employment in the last week multiplied by 52 and actual self-employment income in the last year. (16)
(2) The Real Hourly Wage for Wage/Salary Workers, calculated by dividing actual earnings from wage/salary employment in the last week by actual hours worked in wage/salary employment in the last week. (17) The self-employed are excluded from this measure of labour productivity because hourly wages are notoriously hard to calculate for these workers due to the difficulty in separating investment returns from self-employment income and in reporting accurate measures of work hours. The next measure captures labour productivity for the self-employed subject to these caveats.
(3) Real Hourly Wage for All Workers, calculated by dividing annual labour income by actual total hours work in the last week multiplied by 52.
The second three measure labour supply utilisation in each area:
(4) Employment, which is a discrete variable measuring whether an individual worked any hours in the last week for pay, was away from work but receiving ACC, or worked any unpaid hours for a family business.
(5) Weekly Hours Worked by the Employed, which is directly reported in the HLFS.
(6) Benefit Receipt, which is a discrete variable measuring whether an individual received income from IRD or WINZ for any social benefit besides superannuation or student allowances in the previous fortnight.
4. Empirical Results
We first examine the 'raw' gap in economic performance and economic growth across the five 'regional' areas and four Auckland zones. These results are presented in a series of figures. These graphs indicate the relative level of economic performance in each area/year without controlling for any of the underlying differences in the population of these areas. They also display the mean of each measure of economic performance in each year indexed to the level of that measure in 1997 in each of the areas and four Auckland zones, which indicates the relative change in economic performance in each area/year, again without controlling for any of the underlying differences in the population of these areas.
We begin by discussing our results which compare the average level and average growth of the six measures of economic performance across the five 'regional' areas. Figure 1 presents the results for Real Annual Labour Income, the Real Hourly Wage for Wage/Salary Workers, and the Real Hourly Wage for All Workers. The left column presents the 'levels' results and the right column the 'growth' results. Auckland has the second highest average labour income behind Wellington, but ahead of all other areas. Wellington and Auckland have the highest average wages with the other areas clustered much lower. In general, wages are higher when the self-employed are included, but the differences between areas are similar. Labour income growth in Auckland is middle of the pack, faster only than in Wellington, but wage growth for wage and salary workers has been strongest in Auckland by a fair margin. Wage growth for all workers is also strongest in Auckland but much more similar to the other areas.
[FIGURE 1 OMITTED]
[FIGURE 2 OMITTED]
Figure 2 presents the results for Employment Rates, Weekly Hours Worked by the Employed, and Benefit Receipt Rates. This figure follows the same layout as the previous. Employment rates in Auckland are lower than all other areas besides Other Urban. Workers in Auckland work similar hours to those in Wellington and Rural, but more than those in Christchurch and Other Urban. Auckland has lower levels of benefit receipt than all comparison areas other than Wellington. Employment growth occurred in all areas other than Auckland (and Wellington) and by 2004 employment rates in Auckland were much lower than in the other areas. There has been a small increase in work hours in Christchurch relative to the other areas but no other changes during the sample period. Benefit receipt has declined in Auckland (and Christchurch) relative to the comparison areas.
We next compare the wage distribution across the five 'regional' areas. Here, graphs are used to display the overall distribution (ie. the 10th, 25th, 50th (median), 75th, and 90th percentile) of the two hourly wage measures in each year for the five 'regional' areas. We compare wages across areas at different points in the distribution to give further insight into relative productivity performance in each area. Wage distributions tend to be highly skewed, thus focusing only on means can conceal how circumstances are changing for typical workers, if changes are, in fact, mostly occurring at the extremes of the wage distribution. Percentile measures are also robust to outliers that are presumably caused by measurement error in either earnings or hours worked.
Figure 3 graphs the 10th, 25th, 50th (median), 75th, and 90th percentile and mean of the Real Hourly Wage Distribution for Wage/Salary Workers in each year for the five 'regional' areas, while Figure 4 graphs the same for the Real Hourly Wage Distribution for All Workers. Auckland (and Wellington) have higher wage/salary wages across the entire wage distribution. The magnitude of these differences are larger at the upper-end of the wage distribution. For example, differences are $0.70-$1.40 at the 25th percentile, $1.00-$2.25 at the median, and $1.70-$3.10 at the 75th percentile. Wages for the self-employed have greater variance leading to a reduction in percentiles at the median and below, but an increase in the 75th and 90th percentile in all areas. Wage differences between Auckland and other regions are generally larger at all points in the wage distribution when including the self-employed. The difference in wages between Auckland and Rural is much larger at the 10th and 25th percentile of the distribution but smaller at the 75th percentile.
[FIGURE 3 OMITTED]
[FIGURE 4 OMITTED]
Finally, we discuss our results comparing the average level and average growth of the six measures of economic performance across the four Auckland zones. Figure 5 presents the results for Real Annual Labour Income, the Real Hourly Wage for Wage/Salary Workers, and the Real Hourly Wage for All Workers. The left column presents the 'levels' results and the right column the 'growth' results. Labour income and average wages are higher in Northern and Central Auckland than in Southern and Western Auckland, regardless of whether the self-employed are included in the wage rates. Labour income growth has been similar in all Auckland zones besides the Northern zone which has had slower growth and wage growth has been similar in all Auckland zones.
Figure 6 presents the results for Employment Rates, Weekly Hours Worked by the Employed, and Benefit Receipt Rates. This figure follows the same layout as the previous. Employment rates are highest in Northern Auckland, then Western and Central, and lastly Southern Auckland. Conversely, employment growth has been slowest in Northern, then Western, Central, and fastest in Southern Auckland. Aucklanders in the Central and Southern zones work more hours than those in the Northern and Western zones and work hours have increased in the Southern, Central, and Western zones relative to the Northern zone. Benefit receipt rates are lowest in Northern Auckland, followed by Western and Central, and highest in Southern Auckland, but receipt rates have greatly declined in Southern Auckland relative to all other Auckland zones.
We next turn to a regression analysis to examine whether the graphical differences are statistically significant and whether controlling for differences in the attributes of individuals living in different areas 'explains' difference in regional economic performance. (18) We follow the approach taken in Glaeser and Mare (2001), which estimates differences in outcomes between big cities and other areas in the US, and use individual level data to estimate the following standard human capital regression model augmented with location control variables:
[Y.sub.it] = [X'.sub.it] [beta] + [L'.sub.it][GAMMA] + [e.sub.it], (1)
where [Y.sub.it] is one of the six measures of economic performance for individual i at time t, [X.sub.it] is a vector of individual characteristics, [beta] is the returns to those characteristics in the labour market, [L.sub.it] is a vector of dummy variables describing in which of the five 'regional' areas an individual lives, with Auckland the omitted area, the vector [GAMMA] indicates the differential economic performance in each 'regional' area relative to Auckland, and [e.sub.it] is a normal white-noise error-term.
[FIGURE 5 OMITTED]
[FIGURE 6 OMITTED]
Parallel to the graphical analysis, we use this regression model to examine the average relative difference in the mean level of economic performance over the entire sample period in each comparison area relative to Auckland (levels models) and to examine the change in the relative difference in mean economic performance between 1997 and 2004 in each comparison area relative to Auckland (growth models). We estimate three specifications of this regression model using Ordinary Least Squares (OLS). (19) In the first specification (no controls), the only variables included in [X.sub.it] are indicators for the survey year in the levels model and these variables plus interactions between the survey year indicator variables and the regional indicator variables in the growth models (this allows for differential growth patterns for each region). The results from this specification allow us to examine whether the observed graphical differences are statistically significant.
In the second specification (controlling for demographic covariates), we control for standard human capital variables, including gender, age, ethnicity, educational qualifications, and immigration status. (20) We use a series of indicator variables to control for each of these characteristics, thus these regression models are semi-parametric models, where the coefficient estimates are the weighted averages of mean outcomes in each cell defined by all of the indicator variables. This is an extremely flexible functional form that requires few assumptions about the relationship between the covariates and the outcomes. In the third specification (controlling for all covariates), we include additional controls for employment status, occupation, and industry, again using a series of indicator variables. (21) Unlike the mainly pre-determined human capital variables, these variables are clearly co-determined with the outcome variables and thus it is unclear whether they should be included as control variables. Thus, we examine both a specification excluding them and one including them to see whether the results are sensitive to this decision.
The regression models that control for covariates allow individuals with particular characteristics to have different outcomes than other individuals regardless of their location and reveal whether controlling for differences in the attributes of individuals living in different areas 'explains' differences in regional economic performance. (22) These models are purely descriptive models which partition the difference in economic performance across areas into components that are correlated with differences in attributes and those which are 'unexplained'. These attributes account statistically for the 'explained' variation in outcomes across areas, but do not necessarily have a casual impact on economic performance. Without having panel data, in which we can observe how outcomes change as individuals move between areas, it is not possible to identify casual effects because differences in the attributes of the population in each area are likely to be endogenously related to past and expected future economic performance in all areas.
Table 1 presents the 'regional' regression coefficients from the mean (OLS) regression for each of the six measures of economic performance. (23) In the left panel, the coefficients indicate the average relative difference in the mean level of economic performance over the entire sample period in each comparison area relative to Auckland. In the right panel, the coefficients indicate the change in the relative difference in mean economic performance between 1997 and 2004 in each comparison area relative to Auckland--these are the coefficients on the interactions between the 2004 survey year indicator variable and the regional indicator variables. The top panel presents regression coefficients from models that include no control variables besides survey year indicator variables. These are identical to the differences graphed in the previous figures averaged over entire sample period. The middle panel presents the regional regression coefficients from models that control for demographic covariates. The bottom panel presents the regional regression coefficients from models that control for demographic and employment covariates. All estimates are probability weighted using the sampling weights provided by Statistics NZ to create correct inferences based on the full prime-age population. Huber/White robust standard errors are presented which account for clustering at the Primary Sampling Unit (PSU) level. (24)
The difference in labour income between Wellington and Auckland is entirely explained by differences in demographic characteristics, but Auckland (and Wellington) has significantly higher labour income than all other areas even after controlling for all characteristics (with the addition of employment controls slightly reducing the differential). Auckland (and Wellington) also have significantly higher wages than other areas even after controlling for all characteristics; the results are almost the same whether examining just wage/salary workers or all workers (and again, differences are slightly smaller when adding employment controls). Once accounting for differences in attributes, Auckland had similar growth in labour income as all comparison areas besides Wellington (where it grew much faster than) and significantly faster wage growth for wage/salary workers than all areas besides Wellington. Auckland has also experienced faster overall wage growth than all other areas, but this is less so than for wage and salary wages alone and is not significantly different.
Controlling for demographic characteristics, Auckland has slightly lower employment rates than Wellington, higher than Other Urban, and the same as Christchurch and Rural, but the raw differences are mostly explained by differences in characteristics. Aucklanders (and Wellingtonians) work significantly more hours than workers in Rural, Christchurch, and Other Urban areas once controlling for differences in employment characteristics. Auckland and Wellington have significantly lower rates of benefits receipt than Other Urban and Rural. Once adding controls for demographic characteristics, these rates are also significantly lower than Christchurch's. Auckland has significantly slower employment growth than Christchurch and Other Urban once controlling for differences in demographic characteristics and insignificantly slower growth than Rural. Differences in growth in hours worked and benefit receipt rates are similar whether or not controlling for covariates and none of the differences are statistically significant.
In most cases, differences in demographic characteristics (in particular, educational qualifications and ethnic composition) are the main attributes that explain differences in outcomes between Auckland and other regions. In certain models, employment characteristics also explain regional differences. This paper only looks superficially at the role of different attributes in explaining differences and thus these findings may not be robust. For the remaining analyses, we only present the results from the models with no covariate and the results from the full regression model controlling for both demographic and employment characteristics to keep the discussion focused.
We next examine whether individual characteristics 'explain' regional differences in the wage distribution. Here, we estimate regression model (1), using Quantile regression models to examine hourly wage differences at various points in the distribution across the five 'regional' areas. For example, Median regression estimates the median of the dependent variable, conditional on the values of the independent variable. This is similar to an OLS regression, which estimates the mean of the dependent variable. In other words, median regression estimates coefficients that minimizes the sum of the absolute residuals rather than the sum of the squared residuals. Other conditional quantile functions are estimated by minimizing an asymmetrically weighted sum of absolute errors. (25) Sample weights cannot be incorporated into these analyses. Instead, bootstrapping (replication analysis) is used to estimate standard errors for these models. (26)
Table 2 presents the 'regional' regression coefficients from quantile regressions of the 10th, 25th, 50th (median), 75th, and 90th percentile of the Real Hourly Wage Distribution for Wage/Salary Workers. In the left panel, the coefficients indicate the average relative difference in Real Hourly Wages for Wage/Salary Workers at various points in the distribution over the entire sample period in each region relative to Auckland. In the right panel, the coefficients indicate the change in the relative difference in Real Hourly Wages for Wage/Salary Workers at various points in the distribution between 1997/98 and 2003/2004 in each comparison area relative to Auckland. The top panel presents the "raw" regression coefficients. These are identical to the differences graphed in the Figures 3 averaged over entire sample period. The bottom panel presents the regression coefficients controlling for all covariates. Again, these coefficients allow us to examine whether controlling for differences in the attributes of individuals living in different areas 'explains' any of the differences between these areas.
Controlling for characteristics explains all of the difference in wage/salary wages between Auckland and Wellington at all points in the distribution besides the median, but almost none of the difference in wages between Auckland and the other comparison areas at any point of the distribution below the 90th percentile. Controlling for characteristics does explain 16% of the difference in 90th percentile wages between Auckland and Christchurch, 30% between Auckland and Other Urban, and 47% between Auckland and Rural. While Auckland has faster wage/salary wage growth across the entire wage distribution then all comparison areas; the magnitude of these differences are fairly small and insignificant below the median, and differences in characteristics explain most of the (larger) differences between Auckland and the other areas above the median wage. Overall, controlling for covariates, wage/salary wages have grown in Auckland by an additional $0.25-$0.60 at the 25th percentile, $0.40-$0.90 at the median, $0.40-$0.70 at the 75th percentile, and $0.70-$1.55 at the 90th percentile compared to the other areas.
Table 3 presents the 'regional' regression coefficients from quantile regressions of the 10th, 25th, 50th (median), 75th, and 90th percentile of the Real Hourly Wage Distribution for All Workers. The layout of this table is identical to the previous table. Once again, the results are very similar to those found using only wage/salary wages, and controlling for characteristics explains almost all of this difference between Auckland and Rural at bottom and top of the all workers versus wage/salary worker wage distribution. Including the self-employed, Auckland has similar wage growth as the comparison areas at and below the median, and faster, but insignificantly different, wage growth at the upper-end of the distribution. Differences in characteristics also explains most of difference found between wage growth in Auckland and the comparison areas.
Finally, we examine differences in average outcomes within Auckland. Table 4 presents the 'Auckland zone' regression coefficients from the mean (OLS) regression for each of the six measures of economic performance. The layout is the same as in previous tables. In the left panel, the coefficients indicate the average relative difference in the mean level of economic performance over the entire sample period in each zone area relative to Northern Auckland. In the right panel, the coefficients indicate the change in the relative difference in mean economic performance between 1997 and 2004 in each zone relative to Northern Auckland. The top panel present the "raw" regression coefficients. The bottom panel present the regression coefficients controlling for all covariates.
Differences in labour income, employment, and benefit receipt rates within Auckland are entirely explained by differences in characteristics. All of the difference in average wages for wage/salary workers and for all workers between the Southern zone and the rest of Auckland and two-thirds of the difference between the Western zone and the rest of Auckland is explained by differences in characteristics. Aucklanders in Central and Southern zones work significantly more hours than those in Northern and Western zones; differences are smaller when controlling for covariates, but still significantly different between the Central and Northern zones. Controlling for covariates, work hours have also increased significantly in the Western zone relative to the Northern zone.
The results presented above tell a fairly consistent story. Furthermore, the main findings are unaffected by whether we examine total income or labour income, wages for wage/salary workers or wages for all workers, all individuals age 15 and above or just prime-age adults, entire regions or just the primary urban areas in each region. Auckland and Wellington have the highest levels of economic performance based on almost all measures. In particular, both have significantly higher average levels of labour income, hours worked, and wage rates than the three other comparison areas. These differences remain even after controlling for differences in population attributes across the areas, although "raw" differences between Auckland and Wellington, which generally show Wellington to have higher levels of economic performance, are 'explained' by differences in characteristics in the two areas. One exception is that Auckland has lower employment rates than Wellington, Christchurch, and Rural areas, but these differences are 'explained' by differences in characteristics in these areas. Auckland has also experienced stronger growth in wages, in particular for wage/salary workers, than other regions, but this has been accompanied by slower growth in employment rates (differences in attributes 'explain' a fairly limited fraction of these differences).
Wages are generally higher in Auckland (and Wellington) than in the other comparison areas at all points in the wage distribution and the self-employed in Auckland have relatively higher wages than the self-employed in all other areas across the wage distribution. These differences remain even after controlling for differences in population attributes across the areas and are of similar magnitude whether or not the self-employed are included. Auckland has experienced somewhat stronger growth in wages for wage/salary workers across the distribution, but much of this difference is explained by differences in regional attributes. Once self-employed workers are included, wage growth has been similar in all comparison areas (except Wellington where it has been slower).
Within Auckland, Northern Auckland has the highest and Southern Auckland the lowest level of economic performance based on most all measures. Western and Central Auckland are usually found between these two extremes on most measures. However, differences between Auckland zones are mostly 'explained' by differences in the characteristics of the people who live in these zones and economic performance appears to be converging within Auckland with slower growth found in the Northern zones in most measures of economic performance relative to the Central and Western zone and faster growth found in the Southern zone relative to the other zones. It is also worth noting that these differences reflect where people live within Auckland and may not accurately capture the economic performance of businesses in these zones if many individuals commute across zones.
3. Comparisons with Other Studies
NZIER Estimates of Regional Economic Performance
In September 2004, the New Zealand Institute of Economic Research (NZIER) produced a report for the Ministry of Economic Development, "New Zealand's Regional Economic Performance" (New Zealand Institute of Economic Research, 2004). In the absence of official measures, NZIER developed a set of regional GDP and GDP per capita estimates - both levels and growth rates. It used these and also its own estimates of regional labour productivity levels and growth rates as the basis for its report.
NZIER obtained its estimates of regional GDP (RGDP) by allocating official national GDP figures across regions. It did this largely by using regional employment data. (27) Each region's share of the national industry's employment (available from Statistics New Zealand) was used to determine its share of national real GDP in that industry. This approach assumes that the level of labour productivity within an industry is constant across New Zealand.
The NZIER derived its estimates of labour productivity in each region by once again assuming that labour productivity is constant within each industry across all regions. The estimate of a region's labour productivity is thus a weighted average of national labour productivity figures by industry where the weights are the share of the region's labour force that works in each industry. The effect of this approach is that regions with greater concentrations of industries that have high productivity (growth) at the national level (e.g. business services) come through as those with higher levels (growth rates) of labour productivity.
While satisfactory for some purposes, the assumption of constant labour productivity by industry across regions is not appropriate when investigating the sorts of differences in productivity that may arise in cities compared to less densely populated regions. The NZIER report recognises the limitation of its approach, but was constrained by lack of data. It notes the critical nature of the constant productivity assumption and concludes that it is difficult to draw any firm conclusions about regional productivity levels or growth. We argue that our approach using regional wage and income data from the Income Survey is thus likely to give more satisfactory estimates of regional labour productivity levels and trends.
Table 5 shows the NZIER's estimates of a number of indicators of regional economic performance. The NZIER's 12 regions are based on New Zealand's local authority regions (the sixteen regional councils) with some of the smaller ones combined. Note that these do not correspond with the areas in our study which we chose specifically to investigate differences in performance between large cities, medium sized urban areas and small centres/rural areas. Nevertheless, the Auckland Regional Council area corresponds broadly with "Auckland" in our study. On the other hand, the Canterbury Regional Council, for example, includes a considerable area of rural land and small centres. (28)
The NZIER results indicate that economic performance in most regions is a mixed bag viewed across the various indicators. Three regions stand out as consistently strong performers (Canterbury, Wellington and Southland) and Northland as a weak performer. Auckland on the other hand is very much a middle performer ranking fourth equal in productivity growth, 9th in GDP per head and 11th in growth in GDP per capita.
Our results give a different picture of Auckland's relative performance compared to the one that emerges from these NZIER figures. We find that Auckland has the highest level and growth (or equal first with Wellington) of real hourly earnings of workers, which are likely to reflect labour productivity levels and growth rates. (29) Moreover, the superior performance of Auckland is more marked at the high-wage, high-skill end of the distribution. We do find lower levels of employment and employment growth in Auckland than in some other areas. Thus, lower labour utilisation may help to explain the lower level of GDP per head in Auckland found in the NZIER study despite our finding of high wage levels.
National Bank Estimates of Regional Economic Performance
The National Bank of New Zealand produces quarterly reports on regional economic trends. (30) These are based on the Bank's index of regional economic activity. This is constructed from a mixture of official Statistics New Zealand data and other available regional data together with modelled components to fill in some gaps. (31) The result is a proxy for overall economic activity in each region rather than a true national-accounts type measure of regional GDP.
Each quarter the Bank reports the growth in this index for each region compared to the previous quarter and also the year-on-year growth rate. It also gives regional summaries of the most significant and interesting movements in the individual series that make up the overall activity index. The reports do not present estimates of GDP per capita or labour productivity growth rates or levels, as the data and the approach makes it difficult to derive such measures. (32)
[FIGURE 7 OMITTED]
Figure 7 shows the average growth performance of New Zealand's 14 regions as measured by the National Bank's regional activity index. From 1997 to 2004, Auckland's annual average growth rate of 2.7 per cent was fourth slowest and below the rate for New Zealand as a whole. If anything, the picture from the National Bank's measure shows Auckland performing even less well relative to the other regions than what is found in the NZIER study.
The reasons for the different findings between our results and the National Bank's index of regional activity are likely to be similar to those that account for the different results between our research and the NZIER study. First, the National Bank's index is an aggregate measure rather than either a per-person or per-worker (i.e. productivity) measure. Thus, a region could have slow growth in overall activity because of slow population growth, even though growth in activity per capita might be strong. Alternatively, and more likely in the case of Auckland, high growth in activity per worker (ie productivity) could be offset by low growth in the population of working age or in labour force participation, giving modest rises in overall activity or even in per capita activity. In fact, the growth of real annual labour income (a volume measure that reflects hours of work), not controlling for differences in characteristics, shows Auckland in the middle of the pack and is not inconsistent with the National Bank growth rate estimates over the same period (see Table 1).
4. Summary and Conclusions
In this study we set out to investigate Auckland's economic performance relative to other large cities in New Zealand, to medium-sized urban centres and to small towns and rural areas. We also examined relative performance within Auckland by dividing the region into four areas. A key motivation for the study was to investigate the hypothesis that Auckland is "underperforming". If true, this result would be concerning given Auckland's large size in the New Zealand economy and given a body of overseas research and evidence that large cities have the ability to generate innovation, learning and high productivity growth. Measures of regional economic performance are not well developed in New Zealand because of a relative lack of official data at the regional level. We expand on previous research in this area by using data from the annual New Zealand Income Survey, a sample of around 16,000 individuals with information on location, measures of income and hours worked, and a range of socio-economic variables, to measure regional differences in productivity levels and growth as proxied by labour earnings.
Our results tell a fairly consistent story. Auckland and Wellington have the highest levels of productivity performance based on almost all measures of earnings. In particular, both have significantly higher average levels of labour income, and wage rates than the three other comparison areas. These differences remain even after controlling for differences in population attributes across the areas. Auckland has also experienced stronger growth in wages, in particular for wage/salary workers, than other regions. When we examined wages across areas at different percentiles in the wage distribution, Auckland wages again tended to be at the top and grow fastest. Consistent with cities being places that foster and use skills, the magnitudes in favour of Auckland tend to be higher at the upper ends of the wage distributions. These differences remain for levels but not growth rates after controlling for differences in population attributes across the areas.
On measures of labour utilisation, we find that Auckland does not do as well. It has lower employment rates than Wellington, Christchurch, and rural areas, but these differences are 'explained' by differences in characteristics in these areas. Auckland has also experienced slower growth in employment rates (differences in attributes 'explain' a fairly limited fraction of these differences). A weaker performance in labour utilisation means that Auckland's performance in GDP per capita is "pulled down" compared to its good productivity performance. This is another potential reason for the weaker performance of Auckland that emerges from alternative measures of regional economic performance published by the National Bank and NZIER as compared to our results.
Our key findings cast doubt on the hypothesis that Auckland has been a productivity underperformer within New Zealand. In fact, Auckland appears to be a relatively good performer and this is consistent with agglomeration economies being at work in New Zealand's largest urban concentration. However, because we limited our investigations to within New Zealand we are not able to say how Auckland's productivity performance compares to innovative, high-skill cities in other countries. Given New Zealand's overall poorer performance in labour productivity and the rather modest wage rate growth that we find even for Auckland, it is unlikely to have been as good. From a policy perspective, our study should give pause for thought to those who view Auckland as underperforming and a drag on the rest of the country.
[FIGURE 1 OMITTED]
Appendix Table 1-Sample Size for Regional Analyses Survey Year Auckland Wellington Christchurch Other Urban 1997 3,598 2,016 1,504 5,462 1998 3,528 2,030 1,430 5,396 1999 3,234 1,760 1,465 5,227 2000 3,105 1,750 1,572 5,098 2001 3,306 1,606 1,591 5,099 2002 3,662 1,746 1,765 5,373 2003 3,357 1,685 1,711 5,073 2004 3,499 1,683 1,606 4,884 Survey Year Rural Total 1997 3,261 15,841 1998 3,283 15,667 1999 3,383 15,069 2000 3,734 15,259 2001 3,650 15,252 2002 3,771 16,317 2003 3,602 15,428 2004 3,437 15,109 Appendix Table 2-Sample Size for Within-Auckland Analyses Survey Year Northern Western Central Southern Total 1997 552 601 1,261 1,184 3,598 1998 527 639 1,302 1,060 3,528 1999 423 716 1,079 1,016 3,234 2000 468 615 1,084 938 3,105 2001 476 649 1,180 1,001 3,306 2002 546 721 1,221 1,174 3,662 2003 554 688 1,113 1,002 3,357 2004 504 616 1,156 1,223 3,499 Appendix Table 3-Average Regional Characteristics Akl Wlg ChCh Oth Urb Rural 25-29 15% 16% 16% 15% 11% 30-34 17% 17% 16% 16% 15% 35-39 17% 18% 15% 16% 17% 40-44 16% 16% 15% 16% 17% 45-49 14% 13% 14% 14% 15% 50-54 12% 12% 13% 12% 14% 55-59 10% 8% 10% 11% 12% Female 52% 51% 51% 52% 50% Pakeha 66% 76% 86% 82% 85% Maori 8% 8% 5% 12% 13% Pacific Is 11% 7% 2% 2% 1% Asian 9% 5% 2% 2% 1% Other 6% 3% 4% 2% 1% NZ Born 63% 76% 82% 86% 89% <=5 Yrs in NZ 12% 5% 6% 4% 2% 6-10 Yrs in NZ 6% 4% 3% 2% 1% 11-20 Yrs in NZ 8% 5% 3% 2% 2% 21-30 Yrs in NZ 6% 5% 4% 3% 3% 30+ Yrs in NZ 4% 4% 3% 3% 2% No Quals 17% 15% 19% 24% 27% Vocation Quals 6% 5% 10% 12% 12% Low School Only 11% 10% 10% 11% 12% Low Sch and Voc 12% 11% 12% 13% 14% High Sch Only 14% 13% 11% 10% 9% High Sch and Voc 21% 21% 21% 18% 17% University Quals 19% 25% 17% 12% 7% Wage/Salary 78% 83% 83% 83% 65% Employer 8% 6% 6% 7% 12% Self-Employed 13% 10% 10% 9% 21% Managers 19% 16% 15% 12% 11% Professionals 17% 22% 17% 16% 10% Technicians 14% 16% 14% 13% 9% Clerks 14% 16% 13% 13% 8% Service/Sales 11% 10% 12% 14% 10% Agric/Fishery 1% 1% 2% 4% 27% Trades 10% 8% 10% 10% 8% Plant/Machine Ops 7% 5% 10% 10% 11% Elementary Occ 6% 4% 6% 7% 5% Agric/Fishery 1% 1% 2% 5% 28% Manufacturing 19% 10% 18% 17% 14% Construction 7% 6% 7% 7% 6% Wholesale Trade 7% 5% 6% 5% 3% Retail Trade 11% 9% 10% 12% 9% Accom/Restaurant 3% 3% 4% 4% 3% Transport/Storage 5% 5% 6% 4% 3% Communication 2% 4% 2% 2% 1% Finance/insurance 4% 7% 3% 3% 2% Property/Business 15% 15% 11% 10% 6% Government 3% 12% 6% 5% 3% Education 7% 8% 9% 10% 8% Health/Community 8% 8% 10% 11% 7% Cultural/Recreation 3% 3% 2% 2% 2% Personal/Other Ser 4% 5% 4% 4% 3% Northn Westn Central Southn 25-29 12% 15% 18% 15% 30-34 14% 18% 18% 16% 35-39 18% 18% 17% 17% 40-44 17% 15% 15% 16% 45-49 15% 13% 13% 14% 50-54 14% 11% 11% 13% 55-59 10% 10% 8% 11% Female 53% 53% 51% 52% Pakeha 81% 67% 64% 57% Maori 4% 9% 6% 13% Pacific Is 3% 9% 10% 18% Asian 5% 9% 12% 9% Other 6% 6% 7% 4% NZ Born 67% 65% 61% 62% <=5 Yrs in NZ 13% 10% 14% 9% 6-10 Yrs in NZ 4% 6% 7% 7% 11-20 Yrs in NZ 6% 7% 7% 10% 21-30 Yrs in NZ 5% 7% 6% 7% 30+ Yrs in NZ 4% 4% 4% 4% No Quals 10% 23% 13% 24% Vocation Quals 5% 8% 4% 8% Low School Only 11% 12% 7% 14% Low Sch and Voc 13% 13% 9% 12% High Sch Only 15% 15% 14% 12% High Sch and Voc 26% 16% 22% 20% University Quals 20% 13% 31% 10% Wage/Salary 73% 78% 78% 83% Employer 9% 7% 8% 7% Self-Employed 18% 14% 13% 10% Managers 22% 17% 20% 17% Professionals 16% 12% 23% 13% Technicians 17% 12% 17% 12% Clerks 14% 16% 12% 15% Service/Sales 11% 12% 10% 11% Agric/Fishery 1% 2% 1% 2% Trades 10% 13% 7% 12% Plant/Machine Ops 4% 9% 5% 11% Elementary Occ 4% 6% 5% 9% Agric/Fishery 1% 1% 0% 1% Manufacturing 14% 20% 15% 25% Construction 8% 8% 5% 8% Wholesale Trade 7% 7% 6% 8% Retail Trade 11% 12% 9% 12% Accom/Restaurant 3% 2% 4% 3% Transport/Storage 4% 5% 5% 6% Communication 2% 3% 3% 2% Finance/insurance 6% 4% 5% 3% Property/Business 18% 12% 20% 10% Government 4% 4% 3% 2% Education 7% 6% 9% 7% Health/Community 8% 8% 9% 7% Cultural/Recreation 3% 2% 4% 2% Personal/Other Ser 4% 4% 3% 3% Note: Average characteristics are calculated for the prime-age (25-59) population.
Received on 24 April, 2006; received in revised form on January 25, 2007; Accepted on January 25, 2007.
ARC (2005) Business and Economy 2005, Auckland Regional Council, http://www.arc.govt.nz/shadomx/apps/fms/fmsdownload.cfm?fil e_uuid=0EFEE50A-BCD4-1A24-9E5E-3F729106CC3l&siteName=arc
Ciccone, Antonio (2002) "Agglomeration effects in Europe", European Economic Review 46(2), 213-27.
Crawford, Ron (2004) The effects of agglomeration on economic activity--the empirical evidence on mechanisms and magnitudes, mimeo, New Zealand Ministry of Economic Development, Wellington.
Dixon, Sylvia & David C. Mare (2007) "Understanding Changes in Maori Incomes and Income Inequality 1997-2003", Journal of Population Economics, forthcoming.
Glaeser, Edward L (1998) "Are cities dying?", Journal of Economic Perspectives 12(2), 139-160.
Glaeser, Edward L. & David C. Mare (2001) "Cities and Skills", Journal of Labor Economics 19(2), 316-342.
Hirsch, Barry T. & Edward J. Schumacher (2004) "Match Bias in Wage Gap Estimates Due to Earnings Imputation", Journal of Labor Economics 22(3), 689-722.
Hyslop, Dean & Suresh Yahanpath (2005) Income growth and earnings variation in New Zealand, 1998-2004, New Zealand Treasury Working Paper 05/11, Wellington.
Jacobs, Jane (1969) The economy of cities, New York: Vintage:.
Krugman, Paul (1991) Geography and trade, Cambridge: MIT Press.
Koenker, Roger & Kevin F. Hallock (2001) "Quantile Regression", Journal of Economic Perspectives 15(4), 143-156.
Lewis, Geoff & Steven Stillman (2005) Regional Economic Performance in New Zealand: How Does Auckland Compare?, New Zealand Treasury Working Paper 05/08, Wellington.
LEK Consulting (2001) New Zealand talent initiative." strategies for building a talented nation, http://www.executive.govt.nz/minister/clark/innovate/lek.pdf.
Maani, Sholeh & Tim Maloney (2004) Returns to post-school qualifications: New evidence based on the HLFS Income Supplement (1997-2002), Report to the New Zealand Department of Labour, Wellington, http://www.dol.govt.nz/PDFs/PostSchoolQuals.pdf.
McLellan, Nathan (2004) New Zealand's performance: context and challenges, paper presented at a workshop Productivity: performance, prospects and policies, Wellington, July. http://www.treasury.govt.nz/productivity/speakers.asp#mclellan.
Ministry of Economic Development and Treasury (2005) Economic Development Indicators.
National Bank (2004) "Convergence and Stability in the Regions", Economics Comment, November.
New Zealand Institute of Economic Research (2004) New Zealand's regional economic performance: report to MED, Wellington.
Rosenthal, Stuart S. & William C. Strange (2003) "Geography, industrial organization, and agglomeration", The Review of Economics and Statistics 85(2), 377-393.
(1) Some examples of this literature are Jacobs (1969), Glaeser (1998) and Ciccone (2002)
(2) Statistics New Zealand has recently completed a project to investigate the feasibility of producing regional GDP statistics on an ongoing basis. The project included estimating an experimental series of current price regional GDP for 2000-2003.
(3) For example Krugman (1991) and Rosenthal and Strange (2003)
(4) Lewis and Stillman (2005) includes a in-depth discussion of why larger cities might achieve the status of drivers of innovation and productivity growth within national economies and discusses some of the empirical evidence in this area. Crawford (2005) provides a comprehensive review of the literature on agglomeration mechanisms.
(5) We recognise that comparing regions' economic performance over time spans of a relatively few years runs the risk of confounding medium-term and cyclical factors--if the timing of the latter vary across regions. For example, (i) the timing of the impact of commodity price cycles may well vary between rural and urban regions, and (ii) the employment and earnings of lower-skilled and younger people is more volatile over the cycle and the proportions of these groups are likely to vary across regions. This caveat should be kept in mind for the comparisons in our study.
(6) June quarter data are collected in April, May, and June. Most income and wage data are measured for the reference week prior to the survey and are not seasonally adjusted in the analysis undertaken in this paper. In principal, seasonality may differ across regions, thus impacting our results. However, we believe that seasonal effects are likely to be small in the survey months (as opposed to in the summer and winter months) and that wage rates for adult wage/salary workers, one focus of our analysis, are unlikely to be significantly affected by seasonality. Fortunately, self-employment income, which is likely to have a large seasonal component, is measured for the entire previous year.
(7) Dwellings in the HLFS sample frame are surveyed for eight consecutive quarters with one eighth of the sample rotated out each quarter. Thus, most individuals are included in the IS sample in two consecutive years. Ideally, the standard errors in our regression analysis should be adjusted to account for this, but, unfortunately, the data provided to us by SNZ does not allow us to longitudinally identify individuals.
(8) Data are imputed for all HLFS sample members who fail to complete the IS. Individuals with imputed data are dropped from all analyses in this paper because, as discussed in Hirsch and Schumacher (2004), including imputed data leads to biased income gap estimates when the attribute being studied (here, geographical location) is not a criterion used in the imputation procedure. While non-income measures are unaffected by this issue, we decided to keep all measurements based on a common sample.
(9) Sixteen LGRs are recognised by SNZ: Northland, Auckland, Waikato, Bay of Plenty, Gisborne, Hawke's Bay, Taranaki, Manawatu-Wanganui, Wellington, Nelson, Tasman, Marlborough, West Coast, Canterbury, Otago, Southland. Because of confidentially restrictions, Hawke's Bay and Gisborne, as well as, Nelson, Tasman, Marlborough, and West Coast are aggregated together in the data made available to the authors.
(10) Urban Areas are statistically defined areas with no administrative or legal basis. Main urban areas are very large urban areas centred on a city or major urban centre and have a minimum population of 30,000. Secondary urban areas have a population between 10,000 and 29,999 and are centred on the larger regional centres. Minor urban areas are urbanised settlements (outside main and secondary urban areas), centred around smaller towns with a population between 1,000 and 9,999. The remaining population is in rural centres, which have a population between 300 and 999, and rural areas.
(11) Other main urban areas that are in the same LGR include Tauranga and Rotorua, Gisborne and Napier-Hastings, and Wanganui and Palmerston North. All other main urban areas (Whangarei, Hamilton, New Plymouth, Nelson, Christchurch, Dunedin, and Invercargill) can be uniquely identified in the IS.
(12) Whangarei, Hamilton, Tauranga, Rotorua, Gisborne, Napier-Hastings, New Plymouth, Wanganui, Palmerston North, Nelson, Dunedin, and Invercargill.
(13) These are Pukekohe, Tokoroa, Taupo, Whakatane, Hawera, Feilding, Levin, Masterton, Blenheim, Greymouth, Ashburton, Timaru, Oamaru, and Gore (population 235,686 in 2001 census).
(14) Urban areas in the main conurbations have been divided by SNZ into urban zones, with each urban zone defined as a separate urban area. The five criteria used for determining these zones were: 1) strong economic ties; 2) cultural and recreational interaction; 3) services for major business and professional activities; 4) an integrated public transport network; and 5) significant workplace commuting within the zone.
(15) All wage and income variables are converted into June 2004 real dollar values using the Consumers Price Index (CPI). Three additional measures of economic performance, the self-employment rate, the unemployment rate, and annual real non-labour income, are also examined. The results for these measures tell a similar story as those for the included variables, and thus are not presented to save space, but are available from the authors by request.
(16) Individuals reporting real annual labour income less than -$100,000 or greater than $250,000 are drop from all analyses for this measure. This is done because the income data has a small number of very large outliers that have a considerable impact on mean comparisons. It is assumed that these outliers are reporting errors. Large outliers in the wage data are also dropped for this reason.
(17) Individuals reporting real wages less than $4 or greater than $150 are drop from all analyses for this measure as well as the next.
(18) Statistical significance is an important criterion for judging whether observed differences in the data are 'true' but caution should be used in relying solely on this criteria because with the smaller sample sizes used for this regional analysis 'true' difference may not be statistically significant. The overall magnitude of the differences and of trends over time are also useful criteria for judging the 'economic' significance of different results.
(19) Thus, we estimate linear probability models for the two discrete outcomes: "employed" and "receiving benefits". We felt that the advantage of having comparable models across all outcomes outweighed the disadvantages of estimating linear probability models for these outcomes instead of either logit or probit models.
(20) Age is controlled for using seven indicator variables for five-year age groups. Ethnicity is controlled for using ten indicator variables for complex combinations of multiple ethnicities (i.e. instead of using prioritised ethnicity, we account explicitly for multiple response). Immigration status is controlled for using seven indicator variables capturing whether an individual is born in New Zealand, and if not, how long they have been in New Zealand. The categories are: 5 years or less, 6 to 10 years, 11 to 20 years, 21 to 30 years, more than 30 years, and foreign born but unknown number of years in New Zealand. Educational qualifications are controlled for using seven indicator variables combining information on highest school qualifications, vocational qualifications, and university qualifications to create a more complete picture of an individual's educational background. The categories are: no qualifications, only vocational qualifications, only school certificate (or other low school qualification), school certificate and vocational qualifications, only six-form (or other higher school qualification), six-form and vocational qualifications, university qualifications. Additional indicator variables are included to control for missing values in any of the characteristics. Appendix Table 3 display the percentage of the working-age population in each region/zone with certain characteristics. This is displayed for each of the covariates used in the regression models but some are simplified to more aggregate categories.
(21) Occupation is controlled for using twenty-six indicator variables for each of the two-digit NZSOC90 classification groups. Industry is controlled for using fifty-four indicator variables for each of the two-digit ANZSIC classification groups. Employment type is controlled using four indicator variables for: wage/salary employment, employer of others, self-employed, unpaid family worker. Additional indicator variables are included to control for missing values in any of the above characteristics. These variables cannot be included as controls the examined outcome is employment.
(22) We do not estimate any models where the 'returns' to characteristics are allowed to vary across areas, thus we are assuming that regional differences in economic performance are homogenous across the population.
(23) The regression models with control variables include 115 covariates and thus we do not present the results for any of the variables besides the regional controls. Similar human capital models are estimated in Hyslop and Yahanpath (2005), Dixon and Mare (2006), and Maani and Maloney (2004) using the Income Survey. Unsurprisingly, our results for the demographic and employment variables are consistent with those reported in these papers. Our full results are available by request.
(24) In most representative surveys, random samples of households are drawn from geographical clusters called PSUs. Ideally, standard errors should account for this type of clustering. All regression models in this paper are estimated using the Stata statistical package (version 9). The regress command with the robust cluster (PSU) and [pw=svyweight] options is used to generate weighted regression results with Huber/White standard errors.
(25) See Koenker and Hallock (2001) for more information about quantile regression.
(26) The reported standard errors are adjusted to account for the clustering of survey households at the primary sampling unit (PSU) level. One hundred replications are used for bootstrapping standard errors. These models are estimated using the qreg command in Stata with the bootstrap option for calculating standard errors.
(27) The exceptions to this approach were the GDP associated with the ownership of owner occupied dwellings, which was allocated to each region according to its share of the national population, and the GDP not assigned to an industry which was split across regions using each regions share of the national labour force. See NZIER (2004) page 10.
(28) In fact, we find that our choice of different geographical areas compared to the regional areas used in the NZIER and National Bank estimates appears to make little difference. When we estimated our measures for the regional council areas, we find similar results with Auckland ranked 1st or 2nd among the 12 regions in terms of income and wage levels and wage growth but 8th or 9th in terms of income growth.
(29) There are two caveats here. First, in competitive markets wages reflect the marginal product of labour and not total value added per hour worked which is the usual measure of labour productivity. Secondly, if markets are non-competitive, wages may not reflect marginal products.
(30) The two most recent reports are available at http://www.nationalbank.co.nz/economics/regional/default.htm.
(31) Examples of official SNZ series available on a regional basis are retail sales, employment and population. An example of a non-SNZ series available on a regional basis is the quarterly survey of business opinion undertaken by the National Bank. Other examples of regional data that the National Bank uses are new car registrations, job ads, dwelling approvals, commercial building consents and accommodation guest nights.
(32) The National Bank (2004) has nevertheless published a graphical time series of real activity per capita for Auckland, Waikato, Taranaki and Southland from 1974:1 to 2004:3 based on its index of regional economic activity. This shows the three 'dairy' regions outgrowing Auckland since around 1996.
Geoff Lewis (+) and Steven Stillman ([club]) *
(+) New Zealand Treasury, email@example.com;
([club]) Corresponding author; Motu Economic and Public Policy Research, Level 1, 93 Cuba Street, PO Box 24390 Wellington New Zealand; firstname.lastname@example.org
* We thank Ron Crawford, Richard Fabling, Dean Hyslop, Arthur Grimes, John McDermott, Dave Mare, Megan Claridge, Warwick Terry, an anonymous referee and seminar participants at the New Zealand Treasury for helpful comments. The views, opinions, findings, and conclusions or recommendations expressed in this paper are strictly those of the authors and do not necessarily reflect the views of the New Zealand Treasury, Motu Economic and Public Policy Research, or Statistics New Zealand. Access to the data used in this study was provided by Statistics New Zealand under conditions designed to give effect to the security and confidentiality provisions of the Statistics Act 1975.
Table 1--Mean Economic Performance Across Regional Areas Real Annual Real Hourly Real Hourly Labour Wage- Wage- Income W/S Workers All Workers v. Auckland No Controls 1997-2004 Wellington 3,336 ** 0.84 0.61 (1096) (0.43) (0.51) Christchurch -3,954 ** -2.53 ** -2.74 ** (833) (0.38) (0.46) Other Urban -5,638 ** -3.22 ** -3.41 ** (663) (0.30) (0.34) Rural -5,920 ** -3.74 ** -2.95 ** (687) (0.30) (0.35) R-Squared 0.02 0.03 0.01 v. Auckland Controlling For Demographic Covariates 1997-2004 Wellington 108 -0.32 -0.46 (627) (0.28) (0.32) Christchurch -6,429 ** -3.11 ** -3.23 ** (458) (0.24) (0.31) Other Urban -6,517 ** -3.31 ** -3.39 (412) (0.20) (0.23) Rural -6,810 ** -3.59 ** -2.79 ** (442) (0.20) (0.27) R-Squared 0.22 0.20 0.10 v. Auckland Controlling For All Covariates 1997-2004 Wellington -510 -0.42 -0.38 (426) (0.23) (0.28) Christchurch -5,065 ** -2.58 ** -2.50 ** (327) (0.21) (0.27) Other Urban -4,710 ** -2.60 ** -2.60 ** (284) (0.16) (0.19) Rural -0,718 ** -2.58 ** -2.39 ** (311) (0.16) (0.22) R-Squared 0.46 0.26 0.14 Observations 123,731 74,395 84,875 Weekly Hrs Receiving Employed Worked if Benefits Employed v. Auckland No Controls 1997-2004 Wellington 0.048 ** -0.22 -0.016 (0.009) (0.24) (0.011) Christchurch 0.029 ** -1.47 ** 0.012 (0.008) (0.28) (0.010) Other Urban 0.004 -1.56 ** 0.050 ** (0.007) (0.19) (0.008) Rural 0.020 ** -0.42 0.024 ** (0.007) (0.23) (0.008) R-Squared 0.00 0.00 0.00 v. Auckland Controlling For Demographic Covariates 1997-2004 Wellington 0.014 * -0.34 0.003 (0.006) (0.22) (0.006) Christchurch -0.007 -1.51 ** 0.035 ** (0.005) (0.27) (0.006) Other Urban -0.016 ** -1.37 0.051 ** (0.004) (0.19) (0.005) Rural -0.002 -0.43 0.024 ** (0.005) (0.23) (0.005) R-Squared 0.11 0.12 0.09 v. Auckland Controlling For All Covariates 1997-2004 Wellington NA -0.06 0.006 (0.22) (0.005) Christchurch NA -1.05 0.029 (0.27) (0.005) Other Urban NA -0.89 ** 0.042 ** -0.20 (0.004) Rural NA -0.92 ** 0.027 ** -0.24 (0.004) R-Squared 0.17 0.28 Observations 123,942 93,590 123,942 Real Annual Real Hourly Real Hourly Labour Wage- Wage- Income W/S Workers All Workers v. Auckland No Controls Growth Between 1997 and 2004 Wellington -4,303 -2.40 -1.91 (3956) (1.61) (1.90) Christchurch 1,772 -1.46 -0.77 (2865) (1.49) (1.64) Other Urban 658 -1.54 -1.49 (2513) (1.24) (1.42) Rural 785 -2.13 -0.39 (2420) (1.42) R-Squared 0.02 0.03 0.01 v. Auckland Controlling For Demographic Covariates Growth Between 1997 and 2004 Wellington -3,254 -1.68 -1.30 (2083) (0.90) (1.21) Christchurch 1,320 -1.56 -0.99 (1640) (0.97) (1.15) Other Urban 275 -1.43 -1.42 (1434) (0.81) (1.01) Rural 1,124 -1.63 0.02 (1480) (0.85) (1.01) R-Squared 0.22 0.20 0.10 v. Auckland Controlling For All Covariates Growth Between 1997 and 2004 Wellington -2,504 -1.36 -1.17 (1394) (0.73) (1.03) Christchurch -615 -1.67 -1.17 (1325) (0.77) (0.99) Other Urban -739 -1.46 * -1.36 (1015) (0.66) (0.85) Rural 470 -1.48 * 0.28 (1165) (0.72) (0.85) R-Squared 0.46 0.27 0.14 Observations 123,731 74,395 84,875 Weekly Hrs Receiving Employed Worked if Benefits Employed v. Auckland No Controls Growth Between 1997 and 2004 Wellington -0.023 -0.46 0.030 (0.037) (1.01) (0.043) Christchurch 0.049 1.10 -0.009 (0.028) (0.90) (0.040) Other Urban 0.044 0.38 0.011 (0.027) (0.78) (0.033) Rural 0.027 -0.30 0.025 (0.027) (0.89) (0.032) R-Squared 0.00 0.00 0.01 v. Auckland Controlling For Demographic Covariates Growth Between 1997 and 2004 Wellington -0.016 -0.14 0.010 (0.022) (0.93) (0.023) Christchurch 0.049 * 1.22 -0.013 (0.020) (0.82) (0.025) Other Urban 0.040 * 0.76 0.012 (0.017) (0.78) (0.019) Rural 0.027 0.24 0.017 (0.018) (0.90) (0.018) R-Squared 0.11 0.12 0.09 v. Auckland Controlling For All Covariates Growth Between 1997 and 2004 Wellington NA -0.44 0.005 (0.88) (0.018) Christchurch NA 0.89 0.008 (0.78) (0.021) Other Urban NA 0.67 0.029 (0.75) (0.016) Rural NA 0.45 0.028 (0.91) (0.016) R-Squared 0.17 0.28 Observations 123,942 93,590 123,942 Notes: All amounts are in June 2004 Dollars. Standard errors which account for clustering at the PSU level are in parentheses. All estimates are weighted using the sampling weights provided by Statistics NZ. All regressions include controls for the year of the survey. Estimates that control for demographic covariates also include the following variables: each individual's gender. age, ethnicity, immigration status, and qualifications. Estimates that control for all covariates include the following additional variables: each individual's employment status. occupation. and industry. Coefficients with * are significant at the 5% level and those with ** at the 1% levels. Table 2--Real Hourly Wage Distribution for Wage/Salary Workers Mean 10th Pctile 25th Pctile v. Auckland No Controls 1997-2004 Wellington 0.84 0.45 ** 0.65* (0.43) (0.13) (0.26) Christchurch -2.53 ** -0.34 ** -0.69 ** (0.38) (0.11) (0.23) Other Urban -3.22 ** -0.57 ** -1.08 ** (0.30) (0.08) (0.18) Rural -3.74 ** -0.78 ** -1.38 ** (0.30) (0.10) (0.18) Pseudo R-SQ 0.03 0.01 0.01 v. Auckland Controlling For All Covariates 1997-2004 Wellington -0.42 -0.02 -0.10 (0.23) (0.08) (0.07) Christchurch -2.58 ** -0.74 ** -0.98 ** (0.21) (0.12) (0.09) Other Urban -2.60 ** -0.83 ** -1.10 ** (0.16) (0.09) (0.07) Rural -2.58 ** -0.97 ** -1.19 ** (0.16) (0.07) (0.07) Pseudo R-SQ 0.26 0.11 0.16 Observations 74,395 74,395 74,395 Median 75th Pctile 90th Pctile v. Auckland No Controls 1997-2004 Wellington 1.03 ** 1.86 ** 2.47 ** (0.32) (0.51) (0.93) Christchurch -1.00 ** -1.70 ** -3.39 ** (0.27) (0.39) (0.79) Other Urban -1.53 ** -2.45 ** -4.33 ** (0.15) (0.28) (0.62) Rural -2.23 ** -3.08 ** -5.06 ** (0.18) (0.29) (0.61) Pseudo R-SQ 0.01 0.02 0.02 v. Auckland Controlling For All Covariates 1997-2004 Wellington -0.31 ** -0.26 -0.45 (0.11) (0.24) (0.33) Christchurch -1.59 ** -2.05 ** -2.86 ** (0.08) (0.16) (0.33) Other Urban -1.64 ** -2.14 ** -3.04 ** (0.07) (0.12) (0.23) Rural 1.71 ** 2.15 ** 2.67 ** (0.08) (0.13) (0.23) Pseudo R-SQ 0.21 0.23 0.24 Observations 74,395 74,395 74,395 Mean 10th Pctile 25th Pctile v. Auckland No Controls Growth Between 1997/98 and 2003/04 Wellington -2.40 -0.19 -0.84 (1.61) (0.40) (0.56) Christchurch -1.46 -0.37 -0.46 (1.49) (0.21) (0.54) Other Urban -1.54 -0.12 -0.14 (1.24) (0.20) (0.37) Rural -2.13 -0.16 -0.32 (1.29) (0.30) (0.33) Pseudo R-SQ 0.03 0.01 0.01 v. Auckland Controlling For All Covariates Growth Between 1997 and 2004 Wellington -1.36 -0.28 -0.62 ** (0.73) (0.33) (0.21) Christchurch -1.67 * -0.25 -0.41 * (0.77) (0.26) (0.19) Other Urban -1.46 * -0.08 -0.23 (0.66) (0.15) (0.16) Rural -1.48 * -0.14 -0.27 ** (0.72) (0.19) (0.10) Pseudo R-SQ 0.27 0.11 0.17 Observations 74,395 74,395 74.395 Median 75th Pctile 90th Pctile v. Auckland No Controls Growth Between 1997/98 and 2003/04 Wellington -1.05 -1.85 -3.65 (0.70) (1.27) (2.21) Christchurch 0.05 -0.93 -2.97 (0.77) (1.78) (1.82) Other Urban -0.05 -1.45 -4.42 ** (0.64) (0.95) (1.48) Rural -0.66 -1.73 * -5.30 ** (0.51) (0.88) (1.41) Pseudo R-SQ 0.01 0.02 0.00 v. Auckland Controlling For All Covariates Growth Between 1997 and 2004 Wellington -0.87 ** -0.72 -0.70 (0.31) (0.73) (0.79) Christchurch -0.76 ** -0.59 -1.30 (0.23) (0.43) (0.66) Other Urban -0.53 * -0.69 * -1.55 * (0.22) (0.30) (0.63) Rural -0.41 -0.38 -1.32 (0.25) (0.45) (0.68) Pseudo R-SQ 0.21 0.23 0.24 Observations 74,395 74,395 74,395 Notes: All amounts are in June 2004 Dollars. Bootstrapped standard errors which account for clustering at the PSU level are in parentheses. All regressions include controls for the year of the survey. Estimates that control for all covariates also include the following variables: each individual's gender, age, ethnicity, immigration status, qualifications, employment status, occupation, and industry. Coefficients with * are significant at the 5% level and those with ** at the 1% levels. Table 3--Real Hourly Wage Distribution for All Workers Mean 10th Pctile 25th Pctile v. Auckland No Controls 1997-2004 Wellington 0.61 0.31 * 0.55 * (0.51) (0.13) (0.22) Christchurch -2.74 ** -0.49 ** -0.87 ** (0.46) (0.08) (0.19) Other Urban -3.41 ** -0.69 ** -1.14 ** (0.34) (0.09) (0.16) Rural -2.95 ** -1.60 ** -1.83 ** (0.35) (0.06) (0.15) Pseudo R-SQ 0.01 0.01 0.01 v. Auckland Controlling For All Covariates 1997-2004 Wellington -0.38 0.01 -0.24 * (0.28) (0.06) (0.11) Christchurch -2.50 ** -0.93 ** -1.16 ** (0.27) (0.15) (0.10) Other Urban -2.60 ** -0.94 ** -1.25 ** (0.19) (0.11) (0.08) Rural -2.39 ** -1.12 ** -1.39 ** (0.22) (0.07) (0.12) Pseudo R-SQ 0.14 0.08 0.10 Observations 84,875 84,875 84,875 Median 75th Pctile 90th Pctile v. Auckland No Controls 1997-2004 Wellington 0.88 ** 1.42* 2.36 (0.32) (0.59) (1.23) Christchurch -1.23 ** -1.95 ** -3.66 ** (0.33) (0.43) (1.04) Other Urban -1.72 ** -2.90 ** -4.82 ** (0.17) (0.36) (0.77) Rural -2.32 ** -3.01 ** -3.14 ** (0.19) (0.39) (0.76) Pseudo R-SQ 0.01 0.01 0.01 v. Auckland Controlling For All Covariates 1997-2004 Wellington -0.34 * -0.46 * -0.70 (0.17) (0.20) (0.40) Christchurch -1.73 ** -2.17 ** -3.29 ** (0.10) (0.27) (0.45) Other Urban -1.80 ** -2.26 ** -3.34 ** (0.05) (0.17) (0.06) Rural -1.85 ** -2.25 ** -2.90 ** (0.09) (0.26) (0.25) Pseudo R-SQ 0.13 0.14 0.17 Observations 84,875 84,875 84,875 Mean 10th Pctile 25th Pctile v. Auckland No Controls Growth Between 1997/98 and 2003/04 Wellington -1.91 0.05 -0.63 (1.90) (0.30) (0.49) Christchurch -0.77 0.41 0.47 (1.64) (0.49) (0.42) Other Urban -1.49 0.26 0.18 (1.42) (0.31) (0.46) Rural -0.39 0.71 * 0.34 (1.42) (0.31) (0.44) Pseudo R-SQ 0.01 0.01 0.01 v. Auckland Controlling For All Covariates Growth Between 1997 and 2004 Wellington -1.17 -0.34 -0.59 * (1.03) (0.24) (0.30) Christchurch -1.17 0.20 -0.23 (0.99) (0.31) (0.24) Other Urban -1.36 0.14 -0.13 (0.85) (0.23) (0.15) Rural 0.28 0.25 -0.05 0.85 0.29 (0.17) Pseudo R-SQ 0.14 0.08 0.10 Observations 84,875 84,875 84,875 Median 75th Pctile 90th Pctile v. Auckland No Controls Growth Between 1997/98 and 2003/04 Wellington -1.04 -2.08 -1.30 (0.79) (1.76) (3.22) Christchurch 0.35 -0.50 -0.62 (0.82) (1.27) (2.43) Other Urban 0.08 -1.04 -2.54 (0.58) (1.01) (1.75) Rural -0.22 -0.92 -0.83 (0.58) (1.20) (2.07) Pseudo R-SQ 0.01 0.01 0.01 v. Auckland Controlling For All Covariates Growth Between 1997 and 2004 Wellington -0.74 ** -0.40 0.27 (0.27) (0.86) (0.76) Christchurch -0.28 0.31 0.24 (0.24) (0.56) (1.49) Other Urban -0.18 -0.36 -0.76 (0.34) (0.26) (0.69) Rural -0.11 0.08 0.1 (0.21) (0.44) (0.93) Pseudo R-SQ 0.13 0.14 0.17 Observations 84,875 84,875 84,875 Notes: All amounts are in June 2004 Dollars. Bootstrapped standard errors which account for clustering at the PSU level are in parentheses. All regressions include controls for the year of the survey. Estimates that control for all covariates also include the following variables: each individual's gender, age, ethnicity, immigration status, qualifications, employment status, occupation, and industry. Coefficients with * are significant at the 5% level and those with ** at the 1% levels. Table 4--Mean Economic Performance Across Auckland Zones Real Annual Real Hourly Real Hourly Labour Wage- Wage- Income W/S Wrkers All Workers v. Northern No Controls 1997-2004 Western -5,200 ** -3.63 ** -3.65 ** (880) (0.45) (0.61) Central -848 -0.46 -0.21 (994) (0.46) (0.60) Southern -5,806 ** -3.01 ** -3.52 ** (967) (0.46) (0.60) R-Squared 0.01 0.02 0.01 v. Northern Controlling For All Covariates 1997-2004 Western -864 -1.24 ** -1.29 ** (567) (0.31) (0.49) Central 454 -0.26 -0.17 (637) (0.36) (0.53) Southern 114 -0.18 -0.53 (532) (0.34) (0.49) R-Squared 0.47 0.26 0.16 Observations 27,211 16,106 18,431 No Controls Growth Between 1997/98 and 2003/04 Controlling For All Covariates Growth Between 1997 and 2004 Weekly Hrs Employed Worked if Receiving Employed Benefits v. Northern No Controls 1997-2004 Western -0.025 * 0.64 0.043 ** (0.011) (0.56) (0.010) Central -0.034 ** 1.89 ** 0.047 ** (0.010) (0.55) (0.010) Southern -0.052 ** 1.13* 0.079 ** (0.012) (0.48) (0.013) R-Squared 0.00 0.00 0.01 v. Northern Controlling For All Covariates 1997-2004 Western 0.014 0.80 0.007 (0.011) (0.57) (0.009) Central -0.013 1.49 ** 0.023** (0.008) (0.54) (0.007) Southern -0.003 0.88 0.018* (0.009) (0.50) (0.007) R-Squared 0.13 0.15 0.28 Observations 27,289 19,952 27,289 Real Annual Real Hourly Real Hourly Labour Wage- Wage- Income W/S Wrkers All Workers v. Northern No Controls Growth Between 1997/98 and 2003/04 Western 2,803 0.23 -0.71 (3398) (1.94) (2.33) Central 3,368 0.54 0.37 (3882) (1.82) (2.05) Southern 2,714 0.04 -0.13 (3500) (1.68) (1.94) R-Squared 0.01 0.02 0.02 v. Northern Controlling For All Covariates Growth Between 1997 and 2004 Western 1,765 0.05 -1.44 (2121) (1.26) (1.82) Central 1,654 -0.14 -0.08 (2327) (1.25) (1.54) Southern 710 -0.12 -0.29 (2082) (0.92) (1.31) R-Squared 0.47 0.26 0.16 Observations 27,211 16,106 18,431 Weekly Hrs Employed Worked if Receiving Employed Benefits v. Northern No Controls Growth Between 1997/98 and 2003/04 Western 0.023 4.69 * -0.018 (0.037) (2.09) (0.036) Central 0.042 2.89 -0.024 (0.036) (1.81) (0.036) Southern 0.055 1.81 -0.089 (0.046) (1.86) (0.047) R-Squared 0.00 0.01 0.010 v. Northern Controlling For All Covariates Growth Between 1997 and 2004 Western 0.018 4.38 * -0.014 (0.035) (2.04) (0.027) Central 0.017 2.68 -0.006 (0.030) (1.70) (0.025) Southern 0.029 2.09 -0.062 * (0.037) (1.78) (0.024) R-Squared 0.13 0.15 0.280 Observations 27,289 19,952 27,289 Notes: All amounts are in June 2004 Dollars. Standard errors which account for clustering at the PSU level are in parentheses. All estimates are weighted using the sampling weights provided by Statistics NZ. All regressions include controls for the year of the survey. Estimates that control for all covariates also include the following variables: each individual's gender- age, ethnicity, immigration status, qualifications, employment status, occupation, and industry. Controls for employment status, occupation, and industry are excluded from regressions examining whether individuals are employed. Coefficients with * are significant at the 5% level and those with ** at the 1% levels. Table 5--NZIER Estimates of Regional Economic Performance RGDP Growth: RGDP Per Labour Produc- Annual Capita Growth: tivity Growth: Average Annual Average Annual Average Region 2000-2004 1998-2003 2000-2004 Northland 1.2 2.0 0.0 Auckland (rank) 3.1 (6=) 1.1 (11th) 1.0 (4=) Waikato 2.5 1.8 0.5 Bay of Plenty 2.1 2.3 1.0 Gisborne- 3.0 3.5 1.2 Hawke's Bay Taranaki 3.1 4.0 0.5 Manawatu- 5.1 4.6 0.6 Wanganui Wellington 3.9 1.6 1.5 Upper South 2.2 0.7 0.6 Island Canterbury 4.8 3.7 0.8 Otago 3.8 1.8 1.0 Southland 3.8 1.9 1.7 New Zealand 3.5 2.3 0.9 Unemployment: Nominal GDP Average Rate in Per Capita: Year Year to June 2004 to March 2003 Region 6.0 $26,600 Northland 3.9 (4th) $30,750 (9th) Auckland (rank) 3.6 $31,100 Waikato 5.6 $28,500 Bay of Plenty 5.1 $26,600 Gisborne- Hawke's Bay 4.5 $33,100 Taranaki 4.8 $29,900 Manawatu- Wanganui 4.9 $36,700 Wellington 3.3 $35,800 Upper South Island Canterbury 4.1 $35,650 Otago 5.1 $31,300 Southland 3.1 $33,700 New Zealand 4.3 $32,100
|Printer friendly Cite/link Email Feedback|
|Author:||Lewis, Geoff; Stillman, Steven|
|Publication:||New Zealand Economic Papers|
|Date:||Jun 1, 2007|
|Previous Article:||Is there an exchange rate channel in the forward-looking Phillips Curve? A theoretical and empirical investigation.|
|Next Article:||National R&D and patenting: is New Zealand an outlier?|