The city as a labour market: 3

I ended my previous blog on the city as a labour market with the call for planning to improve affordability and mobility. Sounds reasonable. “Mobility” is the ability to reach any area of a metropolitan area in as short a travel time as possible, and “affordability” is the ability of households and firms to locate in whichever area they deem will maximize their welfare” (See note1).  But whose mobility and accessibility should be improved? And can you do both at the same time - improve mobility and affordability?

Improved mobility will help improve people’s welfare. The more jobs that are accessible, the better the job prospects and in theory at least, better productivity and regional GDP.  Improved affordability means that more people can move into or stay in a (growing) big city and participate in the economic benefits. If house prices are too high, then growth may stall and people may move out, so maintaining affordability is important.  

This is the basic idea of the city as labour market as advanced by Alain Bertaud. In his view, productivity (and hence GDP) is not just determined by city size, but also by its layout and transport system.

Bigger cities are supposed to be more productive than smaller cities. A city that doubles in size should see per capita GDP increase by about 115%, all else being equal.  The evidence from Auckland is somewhat mixed. It is a bit more productive that other cities in NZ, but not by much, on a per capita basis.

Table 1: Regional per capita GDP 2016

Region	Regional GDP per capita (2016)	Population estimate (2016)
Auckland	$58,717	1,593,100
Wellington	$54,565	501,100
Canterbury	$55,727	594,100

(Source: Stats NZ)

If we compare Auckland with the other two metro regions in NZ (Wellington and Christchurch), then Auckland's per capita GDP is about 6 to 7% higher, yet Auckland’s population is more than double the other two regions.  Could the answer to this ‘disjoint’ lie in the arrangement of jobs and houses in Auckland?

In this blog I want to look more closely at the mobility side of the mobility/affordability equation.

Alain Bertuad contends that workers’ mobility – their ability to reach a large number of potential jobs in as short a travel time as possible, is a key factor in increasing the productivity of large cities and the welfare of their workers. Large agglomerations of workers do not insure high productivity in the absence of worker mobility. The time spent commuting should, therefore, be a key indicator in assessing the way large cities are managed.

To help explain the need to improve accessibility, Alain Bertaud sets out a simple city, as per the diagram below. This shows a cross section through a city that extends between points ‘a’ to “e’. Employment is concentrated at points “b”, “c” and “d” with housing in between. The city takes 2 hours to traverse from edge to edge.  

Figure 1: The city as a labour market

If each job location (b, c and d) had 33% of the cities employment, and 1 hour was the maximum commute, then people living between b and d could access 100% of the jobs, but people living between a and b or d and e could only access 66%.

So to improve accessibility, either more people need to live in between ‘b’ an ‘d’, or accessibility from ‘a’ to ‘b’ and ‘d’ to ‘e’ to the rest of the city needs to be improved.

Of course, the edges of a city are never going to have the accessibility of the centre. There is a travel time penalty for living out on the edge. This penalty is reflected in the land prices. Land prices are higher in the middle / centre due to the better accessibility and therefore higher demand. Land prices and hence house prices are lower on the edge due to higher transport costs.  Improve accessibility to the edges and affordability probably goes down. Affordability is then improved by moving further out again to areas of lower accessibility, and so it goes on. The alternative is that relative housing affordability is maintained in the face of improved accessibility and higher land prices by less land and floor area area per person (aka intensification).  

In Auckland, we get constant calls to improve affordability by adding more land onto the edge of the city, but without improved accessibility, is that just a short term fix, at best? This is setting aside the more basic problem that Auckland's built up area is probably at the edge of its commuter shed, as it is.

The 33% of employment in the 3 areas of employment in the simple model above is unrealistic. The centre is always likely to have a bigger share.  If points “b” and “d” only had 25% of employment each, with 50% at point “c”, then people living between “a” and “c” actually have 75% of employment within an hour. Interestingly, trying to get a more even split of employment between the three nodes actually reduces accessibility. Concentration tends to increase accessibility.

Looking at Auckland, the pattern of employment in Auckland is skewed towards the centre and south. Below is a the number of jobs per Local Board area, with the Boards arranged roughly north to south. I have dropped off the ‘rural’ Boards of Rodney, Great Barrier, Waiheke and Franklin.  

Figure 2: Number of jobs by Local Board area, 2016

North-West

Central

South-East

There are two main points of concentration - the central isthmus and the ‘mid-south’.

In percentage terms, about 25% of jobs are in the central Isthmus area covered by the Waitemata Board.

Figure 3: Percentage of regional jobs, Auckland Local Board, 2016

The Local Board areas from the centre out to the north and west accommodate about 50% of regional jobs, while the Local Board areas from the centre to the south and east cover about 75% of regional jobs. So is the south and east the best place to expand housing capacity?

In terms of growth and change, the northern and western parts of the city have added some jobs. If we look at the change in employment between 2001 and 2016 by Local Board we get the following picture.

Figure 4: Change in number of jobs, 2001 to 2016, by Local Board

A bit of a bump in the Albany / North Harbour area, but not much when compared to population growth. However putting more people down south may not be the best place either.  Geography does come into play here. Auckland is elongated north to south by its Isthmus form. The south extends further to the south than the north extends to the north.

Previously I have referred to the Marchetti constant - the one hour commute per day (ie 30 mins each way) as being a very common metric as to what most people find tolerable. In a bigger city, one hour per day may be a bit of a luxury, but it is nevertheless a valid rule of thumb. How far can you get in 30mins? Depends upon the transport system. I have used a 15km radius as a rough and ready estimate of a 30 minute commute (ie assume an average speed of 30km per hour).

As a check, go to mapnificent (see note 2), plug in 30 mins and you get the following map for Auckland central. Mapnificent shows you the area you can reach with public transport from any point in a given time (although Im not too sure of the details as to the travel times involved, but it looks about right).

This is about a 10km radius. A 30 min PT shed is probably a bit of a smaller area than a 30min car trip. The same map for South Auckland is as follows:

For access to employment, I have taken an estimated mid point on the North Shore and an estimated  mid point in the south and drawn two 15km radius circles. See Figure 5 (note, my GIS skills are not that hot, so for some reason the circles have come out more like a lozenge than a circle, but you get the idea. The lozenge shape actually better fits Auckland’s geography than a circle).

Figure 5: 15 km radius estimates - north and south.

The North Shore 15 km radius reaches most of the Isthmus. The southern based 15km radius just touches the southern part of the Isthmus.

Applying this approach to northern, central, western and southern sectors, then we get the following figures as to the number of jobs within a 15km radius.  

Figure 6: Number of jobs within 15km of central point, by urban sector, 2001 and 2016

Figure 7: Percentage of regional jobs within 15 km radius

The south-south east has the lowest number of jobs close by. Just over 40% of regional jobs are within a 15km radius compared to 50% for the north and west. Being in the central area is by far the best place to be. No surprises it is also the most expensive place to live.

However jobs in the south are growing, so that is positive.  Warehousing and distribution are the two industrial sectors that Auckland has a bit of an advantage in, compared to the rest of the country. The map below shows the location of warehousing and transport jobs, as of 2016.  

Figure 8: Number of jobs in transport and wholesale, 2016

There is a concentration in the central south. When we look at the home location (as of 2013) for those working in the warehouse and transport sectors (ANZSIC F and I), we get the following pattern. This map shows the percentage of all workers resident in an Area Unit who are employed in the F and I sectors. Across the region as a whole, about 11.5% of workers are employed in transport and warehousing.  

Figure 9: Percentage of workers in wholesale and transport sectors, 2013, by home location

While there is quite a strong correlation between workplaces and home location in the south for those in the F and I sectors, there are also plenty of workers resident in the north and west who presumably need to travel to the southern Isthmus.

On the other side of the coin, access to the higher paying jobs in the business services sector located in the centre, from those living in the south  is not as good as for those living in the centre, west and north.   

Auckland is kind of grappling with these issues. The Auckland Transport Alignment project (see note 3) identified that two key growth distribution trends are:

• Population growth is spread throughout the Auckland urban area and extends into major future urban growth areas to the north, northwest and south of the existing city. Nearly a third of population growth is projected to occur in areas beyond 20 km of the city centre.

• Employment growth is highly concentrated in a few locations, particularly the city centre, the Airport and other major metropolitan centres. Over a third of employment growth is projected to occur within 5km of the city centre. The growth in service sector jobs, which often prefer to locate in major centres to benefit from agglomeration, is a key force behind the projected concentration of employment growth.

Accessibility from the west and south to employment is likely to decline in the future, as congestion worsens.  

These accessibility projections highlight a significant unevenness to future employment accessibility and a growing polarisation of access to employment in the future. By 2046 more than a million people will be living in the western and southern parts of Auckland, nearly half the region’s population. However, these areas see relatively little improvement in their access to employment over time, particularly by private vehicle. The wider implications of these areas being at least partly excluded from the benefits of Auckland’s expanding employment base over the next 30 years are potentially significant, particularly given they include parts of Auckland with higher levels of deprivation, as well as a number of key future urban growth areas.

I'm not too sure about the extent of concentration of jobs in the central area that the Alignment project has assumed. If population growth is to occur in the southern or western edges (as per the call by Minister Twyford to remove the RUB), then fast, frequent transport links into the central Isthmus will be critical to regional productivity and associated housing, social and economic outcomes. And by fast, I mean more than a shoulder bus lane. Even the train needs speeding up - it takes 50 minutes to go from Papakura to Britomart. Furthermore, that investment in transport links needs to go in now, ahead of development if it has any hope of shaping land use and transport patterns. No sign of that, either.

But with improved accessibility to and from the south and west, will land prices rise and hence affordability drop, unless density of housing increases?

1. http://marroninstitute.nyu.edu/uploads/content/Cities_as_Labor_Markets.pdf
2. https://www.mapnificent.net
3. Auckland Transport Alignment Project - Foundation Report.

Update on trigger points

In my post of the 20 August 2017, I looked at urban development trigger points based on land values. This was to inform zoning based on land value. I came up with the following typologies, based on some rough estimates of what type of unit is cheaper to build. The basic idea is that as land values rise, then land area per dwelling unit needs to drops to maintain affordability. Floor area may also get a bit smaller. At some point, higher land values mean that the higher construction costs of apartments sees a dwelling unit that is cheaper to produce than a stand alone or terrace house, even when those types of houses have a much cheaper per square metre build cost. My rough guess was as follows:

Land value up to $1,500 per square metre 

Two storey stand alone houses, town houses and infill type units

Land value between $1,500 and $2,000 per square metre:

Two or three storey terrace or row houses

Land value over $2,000 per square metre:

Four storey plus apartments.

I provided some basic calculations as to costs of producing a stand alone house versus an apartment to get to these numbers, along with assumptions about floor area. For example, the above bands are based on fairly generous apartments and terraces, not shoe boxes.  

In those calculations I did not include developer’s profit. Quite why, I’m not too sure. This is after I had questioned the absence of developer's profit in other calculations of the cost of land use regulations (see 6 August 2017).

So I have had a recalculation.  

Developer’s profit is often based on build costs. My understanding is that 20% of costs is a pretty standard rule of thumb as to what banks will expect as a developers' margin. This is the figure used by the Auckland IHP in its feasibility testing. The recent Auckland Mayoral Taskforce on Housing suggested that developer's margin was in the order of 18 to 22% of total costs (land and building). Whether the developer ever sees this profit is a different issue - this is just a feasibility test.

Including developer’s profit doesn’t really change the basic approach that I set out above.  As the profit is a ratio of build costs, the basic relationships between land values and build costs do not change. Below is my recalculation of the $2,000 per square metre land value scenario, with developer’s profit included.  

The calculation results in an apartment of about $850,000. Interestingly, Colliers recently said that the average selling price for an apartment in places like Albany and Stonefields was $824,000 and up to $ 1.32 million in the CBD.

In my calculations, apartments are about 15% cheaper than stand alone houses, and therefore likely to be a marketable product, when land values are $2000 per square metre. Note, I have also fiddled with the average apartment size (gross floor area) to make it 130 square metres rather than 150 square metres. The 130 square metres is  my guess as to an actual living area of about 90 square metres, with the rest made up by a car park and common access areas (like lobbies, corridors and stair wells).  

	Stand alone house	Apartment
Land area (sqm)	1,000	1,000
Land value $ per sqm	$         2,000	$           2,000
Total land value	$  2,000,000	$   2,000,000
Building coverage	35%	35%
Storeys	2	5
Floor area (sqm)	700	1750
Costs to build per sqm	$         2,200	$           4,500
Cost of build	$  1,540,000	$   7,875,000
Developer’s profit	$     308,000	$   1,575,000
Total cost	$  3,848,000	$ 11,450,000
Average floor area (sqm) per unit	200	130
Units	4	13
Cost	$  1,099,429	$      850,571

To put the above land values in context, here is a map from Council’s GIS system of land values per square metre for the inner Western Bays, based on 2014 valuations. The red areas are the areas with land values over $2,000 per square metre. These land values will be updated soon with a revaluation occurring this year. Growth in prices between 2014 and 2017 will be reflected in these revaluations. The red areas are likely to have got a bit bigger.  

Much of the red is in the Ponsonby / Freemans Bay / Herne Bay area. The coastal edge is mostly red. The darker orange areas like Pt Chevalier are between $1,500 and $2,000 per square metres. It is a pretty quick transition in land values out from the CBD.

If the above map is compared with the one below, which shows recent building consents for apartments and terraces (via the RCG Development Tracker website), then there is a pretty clear co-relation with the red areas.

Great North Road has seen some apartment development. Land values appear to be on the cusp of the $2,000 benchmark, and so tends to confirm the analysis. Great North Road also has the benefit of the former industrial sites. It is interesting how the apartments have developed on the western side of the road, but not the eastern side.

The analysis also suggests that most of Ponsonby and Herne Bay would probably transition over into apartments, if the zoning allowed.

The important point is that apartments are not really that cheap (they are relatively more affordable compared to a stand alone house). They are also really a high land value, inner city or coastal thing. Now of course the costs of the apartment could come down by a cheaper build and smaller floor area per unit, but that tends to lower quality, and narrower the number of buyers.

What is more likely to dominate the orange areas will be townhouses and terrace units - 2 or 3 storey units that share common walls.  More on this soon.

New urban agenda 2

“This government will take steps to improve our resource management system, with better spatial planning and better enforcement. An urban development agency will be introduced, and more emphasis placed on public transport and light rail.

This government will remove the Auckland urban growth boundary and free up density controls. New developments, both in Auckland and the rest of New Zealand, will be able to be funded through innovative new financing methods like infrastructure bonds. This government will also give Auckland Council the ability to implement a regional fuel tax”.

A couple of blogs back I made some off the cuff remarks about the National and Labour parties election statements on urban areas. I said Labour had the more consistent sets of policies. I may have been a bit hasty, given the above? Quite a lot in those few lines above. They are from the new government's speech from the throne setting out their legislative ambitions.  

Better spatial planning sounds good. But to what end? Will efficient and equitable urban areas be the goal (or just speedier development?) Will climate change adaptation get a serious look in? The question is how you link the spatial planning into the RMA system. There have been mutterings of an urban planning act, but no mention of that.

Urban development agencies also sound good, and with a left leaning government, hopefully they will have more of a public good function to them than the national government’s proposed use of such agencies to attract private investment and speed things up. However such agencies need funding (big amounts of funding) if they are to amalgamate land and promote urban redevelopment.

More emphasis on public transport and light rail. Tick. We need more mixed use zoning along collector and arterial roads that allow for small workplaces as well as apartments above.

Regional fuel tax. I think we are quickly getting to the point where more ‘flat rate’ type taxes to fund transport are not really much use, apart from raising a bit more revenue. But with the growth of electric vehicles and ever more fuel efficient petrol and diesel engined cars, then even as a source of revenue you begin to wonder about petrol taxes. Time for charging for use of the transport network that is route and time specific.

Infrastructure bonds. The idea of government using its size and power to borrow money and back long term bonds to fund  infrastructure upgrades and extensions is probably sensible. The question is who pays the interest and capital - the taxpayer, the rate payer or the user of the infrastructure to be provided? It should really be the user these days. But there are tricky issues to sort out in brownfields areas as to catch up, renewals and growth related costs of infrastructure. Making the user pays in greenfields areas will not make housing any cheaper in these areas, but it is necessary that we limit any subsidies to create an even playing field between expansion and redevelopment.

Free up density controls. This has kind of happened already in Auckland. However there are some issues to resolve, not least of which are urban design issues with the re-emergence of the modern day ‘blocks of flats’ that sit side onto the street,  occupying most of a site with the units looking out over the neighbours back yard. Somehow we need to keep hold of back yards and concentrate more of the bulk of development to the front (road side) of sites.

I think there is also a case to re look at the zoning of the ‘middle ring’ of suburbs - those about 5 to 15km out from the CBD. There is an argument that there should be more three storey development in these areas, with 4 stories along collectors. Busy arterial road corridors (more than 20,000 vehicles per day)  are probably better suited to business uses these days. This all needs to be supported by more investment into open spaces and better street environments, stormwater management and the like. In fact these areas requires a massive investment of funds if compact Auckland is to work.  Finding a funding source is key.    

The odd one out is the removal of the Auckland urban growth boundary (which I guess means the newly minted Rural Urban Boundary or RUB). The RUB is not a fixed things these days - it can get shifted. By what is to be gained from its removal?

New towns? While ideas of new towns sound attractive, they always struggle. Good transport links are needed back into the main metro area as the towns are never self sufficient in terms of jobs, shops or services like education and health.  The transport links would need to be fast if the commute time penalty was not to be too high which suggests new dedicated routes (read big dollars), while any rapid transit link can’t serve all regional employment destinations.  Equally important is how the new town is funded. Ideally, the government buys the land involved at above rural land values but below future urban values; rezones the land, puts in the infrastructure and then sells the land onto developers at the urban value, with the difference used to fund the infrastructure. But is the government really going to go out and force a bunch of rural landowners to sell up at what to them will be a value they think it less than what they deserve? And does this mean the sections will be that much cheaper than might otherwise be the case? On the positive side, at least some of the uplift involved is used for a public purpose. But you don’t need to remove the whole RUB to advance this type of approach.

The other reason for removal of the RUB is to increase competition between landowners over who gets to subdivide. If all land within about 20km of the edge of Auckland is up for grabs, then there may be less of a premium to land that could be subdivided. While attractive, the idea misses the point that there is a form of spatial monopoly at play. Auckland tends to grow along three main corridors (north,north-west and south) so land close to these corridors is always going to have a head start over other land. The more likely outcome is public investment being scattered across the countryside, busy trying to serve a number of new housing ‘hot spots’ crying out for attention because of their poor roads and limited infrastructure. This also gets back to the issue of the infrastructure bonds - who pays?

The RUB is a tool to achieve a number of outcomes. It is possible that this type of omnibus method gets replaced by a bunch of more targeted (possibly more effective) methods. If you look at what the RUB is trying to do, then it is not too hard to think up other (and perhaps more targeted means) of managing the different issues.

Issue	RUB	Alternative
Poorly priced transport systems. Cities can expand more than they should if transport network use is under priced	Slows outwards expansion of urban areas,	Road pricing / congestion taxes
Environmental protection - significant landscapes and environments that contribute to sense of place	Seeks to avoids urban growth in very sensitive / valued areas	‘No go’ areas identified and protected by specific statute (a bit like Waitakere Ranges)
Infrastructure funding and efficient roll out networks	Releases land for development once commitment to funding is in place	Spatial planning to set out programme of network infrastructure extensions, with partial user pays in the areas to be served. Full user pays for out of sequence network extensions
Maintenance of rural production, rural landscapes	Provides certainty for rural productive activities	Open space / landscape covenants or similar

There is also an important timing issue. If the RUB is to be ‘removed’, then it should only be removed after all of the other actions are in place, such as:

1. Better spatial planning that defines ‘no go areas’
2. Secure funding and implementation of public transport / light rail projects
3. Sorting out how to pay for infrastructure
4. Road tolls  / congestion charges (a petrol tax is a poor second here)
5. Urban redevelopment authorities in place and underway
6. Some sort of rural landscape protection scheme - not just outstanding landscapes.

Hopefully there is some recognition of these critical timing issues.