New urban agenda

Election time. How do the two main parties (National, Labour) intend to address urban issues?

National wants a new urban planning act to replace the RMA. This act will see faster development and more opportunities for housing and development in a way that improves the quality of the local environment, and makes them great places to live and work, apparently. Meanwhile, National wants more roads and more use of private-public partnerships in the cosy form of urban development authorities funded through special purpose vehicles (aka known as user pays).

What about Labour? It says that the Auckland urban growth boundary should be removed and density controls abolished. New developments, both in Auckland and the rest of New Zealand, will be funded through innovative infrastructure bonds. Labour also wants to slow down foreign investment is housing and reduce immigration. A capital gains tax is a possibility. Light rail  and public transport seems to get more support than roads, with the use of petrol tax to fund this. Labour wants to build more state homes.

Both sets of policies suffer from their own contradictions: National’s basic desire to ‘free things up’ sounds positive, but you get the feeling that this will get warped into freeing things up on the edge of the city. But the desire for a user pays approach to infrastructure and the inability to build roads fast enough will mean that land and housing on the edge of the city will remain expensive. Labour’s push towards public transport, removing density barriers and building more state houses perhaps better favours intensification, but they want to ban foreign buyers. So where is the money going to come from? Is Labour serious about quality urban design, better open space and improved amenities in existing urban areas? Do they know how much money will be needed to sort out the infrastructure? 

Does either party address the issue of finance and the tighter lending rules being put in place by the main banks? Add in the difficulties of over-providing for growth (NPS-UDC), maintaining and enhancing freswater (NPS-FWM) and avoiding significant natural hazard risks (Sec 6), then its not an easy mix.

Put simply, to help ‘solve’ our housing crisis we need to generate a glut of housing. A glut of apartments may be an easier task than a glut of stand alone houses. Apartments have to be built in ‘lumps’, once committed you have to build the apartment, you cant stop halfway. For stand alone houses, it is easier to slow production in response to changes in demand. Only if migration suddenly drops might an oversupply of stand alone houses eventuate.

We managed to generate a bit of a glut of apartments back in the mid 2000s as many small, cheap apartments got built in the Auckland CBD. They were built for investors and financed by the (dodgy) non bank finance sector (that subsequently collapsed). The apartments were also built because of the shift in office space demands down to the waterfront, leaving many sites on the southern and western fringe of the CBD with no obvious use. There is a theory that the failure of the finance sector that funded these apartments (often leaky buildings) has seen investors turn off new builds and instead focus on existing houses, hence the steady rise in the median house price since the mid 2000s.

The Australians have recently generated a glut of apartments in their central cities. The factors that have propelled this glut seem much the same as our mid 2000s boom . Factors like low interest rates, lack of alternative investments and growing demand for inner city living has seen a big inflow of money into apartments, so much so there are concerns about a big oversupply. Planning has helped lubricate the wheels. In their case foreign finance has been critical. As one report (1) puts it:

the big kicker in recent years has been the rise of the overseas investment.  This has powered the Australian apartment market on both the consumer and the developer sides.  The entry into the Australian housing market by the overseas development sector seeking more stable conditions to undertake their business, supported by a relaxed foreign investment policies and our increasingly permissive planning framework, is a new addition to our housing market.

Much of the apartment development is occurring in the former industrial areas that surround their central cities, perhaps equivalent to Newtown and Kingsland in Auckland, where the decline in manufacturing has seen large areas of land become free. Rezonings and regeneration plans have helped to give the market some certainty, but the main force appears to be foreign investors looking for a return.

We need investment in housing, so how do we shift investor demand from existing stand alone houses to new apartments? Rather than try to force out investors with restrictions on foreign capital and minimum deposits, do we instead need to harness the capricious investors for society's ends? 

How do we support finance from the non bank sector? Has the mid 2000s collapse of the finance sector that funded the CBD apartment mini boom burned too many people for them to return to the new apartment market? We need foreign investment. Do we need to deliberately tip the playing field by limiting some types of foreign ownership (existing houses) but enable other forms of investment (new apartments in designated areas?). Do we need a coherent plan to free up more land in former industrial areas for redevelopment? One action could be more new industrial areas on the edge of the city to help promote re-locations. Increased height in the industrial areas that ring the CBD is another move (subject to good design and incorporation of public benefits like some affordable housing). Do we need some aggregation of fragmented land holdings? Do we need improvements to the public realm in the areas of change - better open spaces, improved road environments, better facilities for walking and cycling? Improved urban design must be a bottom line this time around.

Are either party thinking of these types of actions? Are either really thinking of the next urban age?

Note 1

http://blogs.unsw.edu.au/cityfutures/blog/2017/03/im-forever-blowing-bubbles-has-the-sydney-apartment-market-finally-burst/

Urban redevelopment: trigger points

Urban Auckland needs to intensify. No doubt about that. The question is where and how.

Is there some simple relationship that can help identify where redevelopment is likely to occur, for example where a step up from townhouses to apartments is likely to be feasible?  Good urban planning says redevelopment and intensification should happen around town centres and transport hubs. In these places there is access to public transport and a range of amenities. Often town centres have some ‘ragged’ edges that can be redeveloped without intruding too strongly into neighbouring suburbs. But are town centres the natural place to intensify?

The Urban Development Capacity NPS is likely to kick off a whole lot of detailed feasibility testing of development. In Auckland, this testing has already got down to site-by-site GIS analysis.  I keep on thinking there must be a simpler way to help determine appropriate zoning envelopes to provide for urban intensification than lots of detailed modelling.

The actual relationships that determine where intensification may feasibly occur are fairly straightforward. To produce a house, cost equals land price plus construction costs. Construction costs vary between a stand alone house and an apartment, but do not vary across space. What does vary across space is land values. As land values rise, then the land area per dwelling needs to reduce to keep houses relatively affordable - a house on a 800 sqm section on the edge of town should have a similar value to a houses on a 400m sqm section closer to town. At some point land gets so expensive that it makes sense to ‘go up’ - to replace land with capital.  On a per square metre rate, construction costs are much higher for an apartment, so land values have to be quite high to makes it cheaper to build the apartment than a stand alone house.

For example, if we take a 1,000 sqm section and assume a land value of $500 per sqm, then the section has a value of $500,000. On that plot we might be able to build 4 dwellings or 12 apartment units. The dwellings have a land area of 250 sqm or $125,000 per lot. The apartments have a (nominal) share of the land area of 83 sqm, or $41,500 each. But the construction costs of the apartment will be higher - there is more floorspace and build costs on a per sqaure metre rate may be up to twice that of the stand alone house. Those higher construction costs are not off-set by the lower land costs per apartment unit, as the following table demonstrates.

Table 1

Component	Stand alone house	Apartment	Notes
Land Area (sqm)	1,000	1,000
Land value ($) per sqm	$   500.00	$  500.00
Land value	$   500,000	$   500,000
Building coverage (%)	0.35	0.35
Storeys	2	5	4 storeys plus basement
Floor area (sqm)	700	1750
Costs to build ($ per sqm)	$ 2,200	$ 4,500	Assumes build costs, fees and charges and finance costs
Total cost of build	$  1,540,000	$ 7,875,000
Total cost (land plus build costs)	$  2,040,000	$  8,375,000	No allowance for developers profit
Average floor area per unit(sqm)	200.	150.	Gross area - includes garages, common areas in apartment
Number of units	3.5	11.7
Average cost of units	$  582,857	$ 717,857

In this case, you can’t build an apartment that is cheaper than a house.I have based the build costs on MBIE's development feasibility calculator.

If we adjust the land value so that it is much greater, say $2,000 per sqm, then the apartment becomes cheaper than the stand alone dwelling:  

Table 2

Element	Stand alone house	Apartment
Land Area (sqm)	1,000	1,000
Land value ($ per sqm)	2,000	2,000
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
Total cost of build	$  1,540,000	$   7,875,000
Total cost (land plus building)	$  3,540,000	$   9,875,000
Average floor area (sqm)	200	150
Units	3.5	11.7
Cost	$  1,011,429	$   846,429

While the apartment is not cheap, it is cheaper than the house.

If we hold build costs steady, at what land value does the apartment get cheaper than the house? Based on the above numbers, the following graph shows the crossover point is at about $1,100 per square metres. However we are comparing two different products -  a stand alone house with a floor area of 200 sqm with its own garden and an apartment which may just have a balcony and be half the size in terms of actual living area. There is a theory that to make up for not having a garden and a smaller living area, people want to see a discount in price between the apartment and the dwelling. If the price is very similar, then why buy the apartment? Some may buy the apartment because it is easier to maintain, but if given a choice, then most people go for the space. Generally if the apartment is 10 or 15% cheaper than the house, then that may sound like a reasonable deal. If the discount is about 15%, then the land price needs to be more like $1,900 per sqm.

Figure 1

Two key assumptions are the floor area of the house and the apartment, along with their respective build costs. These assumptions can be varied, and so the crossover point would also vary. Exactly what the numbers should be is likely to be a matter of some debate, I could imagine.  

If the crossover point is close to $2,000 per square metre, then the area of the city that has these land values is actually quite small. It is really only the inner Isthmus area and the eastern seaboard. Two problems are immediately apparent: most of the inner Isthmus is covered by heritage suburbs, while most people will say don't build tall buildings along the coastline.

Could you make the point at which the apartment is more attractive relative to the house at a lower land value if you made the apartment building bigger and shrunk the size of the units? Well yes you can, but then you may need to accept a bigger discount between the house and the apartment. If we assume 45% site coverage for the apartment, a gross floor area of 120 sqm per unit, then at a land value of $1,000 per sqm, an apartment would be about 20% cheaper than the stand alone house.

Table 3

	Stand alone house	Apartment
Land Area (sqm)	1,000	1,000
Land value ($ per sqm)	$1,000	$1,000
Total land value	$  1,000,000	$1,000,000
Building coverage	35%	45%
Storeys	2	5
Floor area	700	2,250
Costs to build per sqm	$ 2,200	$  4,500
Cost of build	$   1,540,000	$ 10,125,000
Total cost	$   2,540,000	$ 11,125,000
Average floor area (sqm)	200	120
Number of units	4	19
Average cost of units	$ 725,714	$   593,333

In this case you end up with quite a big building and more, smaller, units. Not a great mix. The units probably get harder to fit in and are more likely to be single aspect, reducing liveability.

And at some point you run into problems with car parking. Even with the first scenario of the 12 unit apartment, if there were 1 car park per unit, then 12 cars would occupy about 420 sqm of space. This is more than the 350 sqm of space within the 35% footprint of the basement level, so a couple of car parks need to be at ground level, or not provided for.

If you increase the number of units, then parking becomes a nightmare to fit in. Under the ‘larger  box/smaller unit’ scenario, 19 car parks would occupy about 665 sqm of space. So at least half of the car parks may have to be on the ground, or in a much bigger basement. Either way, there is much less land for garden and landscaping. Car parking does not need to be provided, but most people want a car park (or want to use the space to stash the bike, kayak, golf clubs and camping gear).

The other problem is that it is hard to find large enough sites for apartment buildings. You also need to pay for the existing house on the land. But with high land values and an older housing stock, the value of improvements may be quite low. The $2,000 per sqm section may end up being a total of $2.2m - $200,000 for the house and $2m for the land.

The more likely option is that as land values rise the size of the section will shrink - go from a town house to a terrace type house. The following table estimates a terrace type product - 5 units . I have put up build costs a little.

Table 4

	Terrace house	Apartment
Land Area (sqm)	1,000	1,000
Land value per per sqm ($ per sqm)	1,500	1,500
Land value	$ 1,500,000	$   1,500,000
Building coverage	35%	35%
Storeys	2	5
Floor area (sqm)	700	1750
Costs to build per sqm	$    2,400	$    4,500
Cost of build	$  1,680,000	$   7,875,000
Total cost	$  3,180,000	$   9,375,000
Average floor area (sqm)	150	130
Units	5	13
Cost	$ 681,429	$   696,429

The end result of the above calculations is the classic density / land value gradient: urban density edges up as land values rise. As some point, density takes a sharp upward tack, once it becomes cheaper to build an apartment, compared to a town house.

Figure 2

In simple terms, you could say that there are three bands of density, set by land values and build costs. These are set out in Table 5. Should this be the rule of thumb to 'drive' zoning?

Table 5

Land Value per m2 (area median)	Development Type
Up to $1500 per m2	Two storey stand alone houses, town houses and infill type units
Between $1500 and $2000 per m2	Two or Three storey terrace or row houses on sites less than 200m2 per unit
Over $2,000 per m2	Four storey plus apartments