6. Quantifying Agglomeration Impacts in Auckland
The London research outlined in the previous section has shown that there are linkages between accessibility and density, and density and average earnings. In this section, we replicate this approach in the Auckland context to see whether similar relationships exist.
6.1 General Approach and Data Sources
This work is primarily based on Census data and information from the Auckland Regional Transport (ART) and Auckland Public Transport (APT) models for 2001. While information on employment density and average earnings by workplace area is available for almost the entire area within the region, the coverage of the transport models is more limited and is largely confined to the metropolitan areas. Attention has been focused on these areas. As in the London work, earnings have been taken as a proxy for productivity, with higher earnings reflecting higher value added/productivity.58
6.2 Employment Density and Productivity
6.2.1 The Auckland Region
The relationship between employment density and average earnings for the four cities within the Auckland region (Auckland, Manukau, North Shore and Waitakere) is set out in Figure 6.1. Given the rather dispersed nature of the employment density data, we use a log transformation to assist in presentational terms. Within the constraints of this work, we have concentrated on the overall picture and we have not at this stage attempted to break down the relationships to consider the particular components which contribute to different productivity levels in different areas.
Figure 6.1: Auckland Cities 2001: Employment Density and Earnings
→ Full size version of Figure 6.1 [71 kB JPG]
The pattern that emerges for the cities in the Auckland region is that at low densities there is little apparent relationship between employment density and productivity. However above a minimum density of 1 job per hectare (Log density = 0) a positive relationship appears.
The CBD zones typically have high densities and high average earnings. However, in Newmarket (circled in the graph) the high proportion of employment in the retail sector, which typically has lower wage rates, means that the average for the area is relatively low despite a high density of employment.
6.2.2 Auckland City
The area with higher employment densities on the right hand side of Figure 6.1 is dominated by Auckland City, and the position here in terms of employment densities and average wages is illustrated graphically in Figures 6.2 and 6.3. This analysis excludes the islands in the Hauraki Gulf.
Figure 6.2: Auckland City 2001: Density of Employment
→ Full size version of Figure 6.2 [182 kB JPG]
Employment density is highest in and around the Auckland CBD, although there is a pocket of relatively high-density employment in the Ellerslie North area and to a lesser extent in the industrial belt between Mount Wellington and Onehunga.
Figure 6.3: Auckland City 2001: Average Earnings
→ Full size version of Figure 6.3 [227 kB JPG]
To a large extent, the pattern of average earnings reflects the pattern of employment densities. Average earnings are again high in the CBD and surrounds, and are also relatively high in the area between Mount Wellington and Onehunga. There are also a few zones which have high earnings without particularly high employment densities, such as Kohimarama East, Epsom South and One Tree Hill Central, but in general the numbers in these zones are small.
The overall linkage between employment density and average wages in Auckland City is set out in Figure 6.4.
Figure 6.4: Auckland City 2001: Employment Density and Average Earnings
→ Full size version of Figure 6.4 [53 kB JPG]
As Figure 6.4 demonstrates, within Auckland City the linkage between density of employment and average earnings is strong. The outlier with an employment density of about 1.25 and average earnings of just over $20,000 is Sandringham North, which is dominated by the retail centre at St Lukes with the low wage rates typical of the retail sector.
For the position as a whole, an increase in employment density from 10 per ha to 100 per ha (from 1 to 2 on a log scale) average earnings increase by about 30 per cent, implying an elasticity of about 0.03, broadly in line with the experience from overseas.
6.2.3 Other Cities in the Auckland Region
A similar analysis has also been undertaken for the other cities in the Auckland region and the relationships defined are set out in Figures 6.5 to 6.7. It should be noted that the defined relationships concentrate on the urban areas and exclude areas where the density of employment is less than 1 per hectare, typically in the more rural parts of these cities. There are no zones within Auckland City where the density is below 1 job per ha.
Figure 6.5: Manukau City 2001: Employment Density and Average Earnings
→ Full size version of Figure 6.5 [55 kB JPG]
Figure 6.6: North Shore City 2001: Employment Density and Average Earnings
→ Full size version of Figure 6.6 [48 kB JPG]
Figure 6.7: Waitakere City 2001: Employment Density and Average Earnings
→ Full size version of Figure 6.7 [44 kB JPG]
In the other cities the strength of the relationships and their statistical robustness is less than for Auckland City, possibly reflecting the nature of their growth often as overspill areas for various types of industry or dormitory suburbs of Auckland City and the more limited time for these to reach a position of equilibrium. However, in all cases there is a positive relationship between employment density and average wages. This suggests that agglomeration impacts are widely experienced across the region, although the magnitude of these varies by area.
6.2.4 The Position in Wellington
A similar analysis has been undertaken for Wellington City and the results are set out in Figure 6.8.
Figure 6.8: Wellington City 2001: Employment Density and Average Earnings
→ Full size version of Figure 6.8 [37 kB JPG]
Again, there is a fairly strong relationship between employment density and average wage rates by area, both in terms of the gradient of the fitted line and the statistical robustness of this. The line for Wellington City is higher than that for Auckland City, but its gradient, the response of productivity to changes in density is less marked, possibly reflecting the different mixes of activities in the two cities.
6.3 Accessibility and Employment Density
6.3.1 Introduction
Having considered the relationships between employment density and average wages or productivity within the Auckland region, the next step is to consider the effects of changes in transport accessibility, since this provides a method by which the agglomeration process can be influenced. The work described above indicated that in particular in Auckland, constrained transport accessibility particularly reflecting increasing congestion on the links into the CBD appeared to be limiting growth in employment. Measures that would allow improvements in accessibility may therefore contribute to increased levels of economic activity.
Transport accessibility measures the extent to which particular areas have access to opportunities at a greater or lesser distance, normally defined within a particular catchment area. It therefore has two components:
- The way in which the catchment area is defined.
- The opportunities within the defined catchment areas.
6.3.2 Measures of Accessibility
There are a number of alternative definitions of accessibility. For the purpose of this analysis we have followed the approach developed in London which defines accessibility on the basis of the numbers of residents and workers that can access a particular location by car and by public transport. The analysis for Auckland is based on the accessibility of the census area units (CAUs) within the Auckland cities and he definitions have been based on a 30-minute journey by private car and on a 75-generalised cost minute journey by public transport. This measure for public transport recognises the different components of the complete journey (walking and waiting as well as the time spent travelling on the public transport vehicle) and the perceived weights attached to each. The information used for this is derived from output from the ART and APT models operated by Auckland Regional Council. The values selected to define the catchment areas are based on the average journey times by the two modes. The accessibility index calculated for each CAU measures the proportion of the regional totals for population and employment which are within the defined catchment areas of that zone.
6.3.3 Auckland Region
The position for the Auckland region as a whole with aggregation of the data points by levels of accessibility is set out in Figure 6.9.
Figure 6.9: Auckland Cities 2001: Accessibility and Employment Density
→ Full size version of Figure 6.9 [37 kB JPG]
This shows that at low levels of accessibility, changes in accessibility appear to have little effect on employment density but at higher levels, the effect becomes stronger. This mirrors the position identified for London as set out above in Appendix B, Figure B.2. The curve, with relatively high levels of employment density around the accessibility index of 0.25, reflects the position in the major centre of Takapuna, which although it has relatively low accessibility, has a relatively high density of employment. This suggests that in particular cases, factors such as urban amenity, in the case of Takapuna a location close to the sea, may also play a part in determining employment locations, with workers being prepared to trade-off an improved environment around their workplace against more difficult accessibility. While transport accessibility appears to be an important factor influencing employment density, in specific instances other factors may enter into the equation.
The position for Auckland City has been considered in more detail. Accessibility levels in 2001 are set out in Figure 6.10.
Figure 6.10: Auckland City 2001: Accessibility
→ Full size version of Figure 6.10 [220 kB JPG]
The patterns of accessibility reflect the presence of the motorway network and also the concentration of public transport routes in the central area. As a result, the figure demonstrates high levels of accessibility in the central area and also along the Southern Motorway corridor to the south. It is noteworthy that Ellerslie South is an area of high accessibility, which appears to be reflected in the relatively high employment density and high average wages set out in Figures 6.2 and 6.3.
For the City as a whole the simplified linkage between accessibility and employment density is set out in Figure 6.11.
Figure 6.11: Auckland City 2001: Accessibility and Employment Density, Simplified Relationships
→ Full size version of Figure 6.11 [36 kB JPG]
Again, this displays a similar pattern to the position described for the region as a whole with employment density only increasing slowly at low levels of accessibility but increasing more sharply as accessibility levels increase.
The more detailed position is set out in Figure 6.12.
Figure 6.12: Auckland City 2001: Accessibility and Employment Density, Detailed Relationships
→ Full size version of Figure 6.12 [56 kB JPG]
Within Auckland City, the relationship between accessibility and employment density appears to be strong and is reasonably robust statistically, although the straight line fitted to the data may underestimate the response at high levels of accessibility. Further investigation of this issue would probably be beneficial.
Back to Top