Wednesday, March 28, 2012

Cities, Cars, People: is changing car use a function of new urbanism?


Thinking dense
One cornerstone for urban designers and planners seeking to transform the polycentric or suburban city of the 20th Century into something resembling the high density city of the 19th was a cross-city comparison by Newman and Kenworthy and successors. [1]  They argued that this proved automobile dependence is a function of city density.  It followed that regulating for greater residential densities and increasing the capacity of public transport systems to avoid the congestion that would follow if people continued to drive themselves would improve the sustainability of cities.

Of course, any comparison with the overcrowded and unhealthy cities of an earlier century is unfair: today’s density is achieved with higher standards of private and public space, and much enhanced transit and sanitation.  And many, probably the majority, of 21st century citizens in high income nations can escape the confines of the urban environment on occasional sojourns to country or coast (or beyond), unlike their 19th Century or developing world counterparts.  They can even find repose in the midst of 24/7 city hubbub in their own in-house media centres.

But can we really build urban policy on the Newman and Kenworthy analysis?  Especially given evidence that car use is declining anyway?

Questionable correlation
There are still questions over the original analysis and it successors.  Cross-cultural effects, physical geography, differences in economic structure, incomes, wealth, and growth all intervene in the relationship between city density and car dependence.  And cause and effect are hard to pin down. 

Perhaps more critical: the leap from observing relationships across cities at a point in time to regulating travel behaviour, housing ,and consumption choices into the future assumes that individual behaviour is a microcosm of collective behaviour. This fallacy of inference has long been recognised by the biological and sociological sciences.  And the likelihood of getting policy wrong by making such an assumption is far greater when dealing with populations of people, with their diverse circumstances, beliefs, values, and means, compared with, say, populations of penguins. 

Is it this blind spot that has made it so much more difficult to get people out of their cars or their low density houses than anticipated by urban reformists?

The city as a time warp
One problem is that analyses of city density and car dependence are usually static.  Plotting urban form and transport consumption at a particular point in time – the mid/late 20th century in the Newman and Kenworthy case – embodies particular patterns of technology, wealth, and, behaviour.  Consequently, their urban prescription is based implicitly on the 9 to 5 work day; single city centres that focus urban employment, exchange, and consumption; and the nuclear family with its distinctive housing and service demands; all urban artifacts that have been breaking down since the 1960s.

But the times they are a-changing
In a 2011 paper the authors acknowledge that things are changing as international evidence shows rates of car use beginning to decline in parts of the world.  A partial view of what they are changing from, though, sustains a deterministic explanation of the why and what they are changing to:

“technological limits set by the inability of cars to continue causing urban sprawl within travel time budgets; the rapid growth in transit and re-urbanization which combine to cause exponential declines in car use; the reduction of car use by older people in cities and among younger people due to the emerging culture of urbanism and the growth in the price of fuel which underlies all the above factors”.[2]

The view remains time-bound; even the reference to exponential decline is a simplistic inference of the relationship between public transport and car use taken from a cross section of cities in 1995. 
Individual agency barely gets a mention.  Any description of an “emerging culture of urbanism” needs to be embedded in the reality of evolving patterns of wealth, income, and consumption and even in simple demographics to determine just how real and significant it is.

Growing old and driving more
What are the grounds for the claim that older people are reducing their car use, for example? I took a quick look at the evidence for New Zealand.  It is certainly not the case here.  The rate of growth in driving has been higher among older age groups than among younger – with decline most evident among the under 45s.  
Is it so different in the other ageing societies from which Newman and Kenworthy draw their examples?

Figure 1: Changes in Annual Driving Distance by Age, New Zealand 1990-2008


Fewer kilometres doesn’t mean less dependence
What does go a long way to explaining declining car travel in the aggregate is the fact that older people don’t drive as much younger people, and populations in western cities are simply getting older.  It’s simple maths – as the population ages car usage will go down, despite a greater propensity to drive among older cohorts. Again, look at the evidence from New Zealand:

Figure 2: Automobile Dependence by Age Group, New Zealand 2004-2008



Car usage appears to decline after age 44, rapidly after retirement age, 65. 

Why does car use fall with age?
There are a number of reasons why this may be so.  From 45 years on households have fewer transport-dependent children.  Mature families may have more localised social networks.  A greater share of recreation may be neighbourhood based.

On retirement work trips disappear and incomes, discretionary dollars and consumption fall.  The capacity for more shared travel and trip planning increases as households age.  Diminished car use doesn't necessarily mean that  households are less automobile dependent.  They just doesn’t generate as much travel demand.

These explanations don’t depend on particular urban designs.  Yet Newman and Kenworthy claim that diminished driving happens because “older people move back into cities from the suburbs”.  This is not consistent with the common observation of people’s preference to age in place.[3]  (For the New Zealand evidence, see my posting Ageing in the City).

Moving into the centre - a one-way street?
And their notion “the children growing up in the suburbs would begin flocking back into the cities rather than continuing the life of car dependence” rather simplifies a historically specific event: the transition of sons and daughters of the baby boomers from young adulthood, advanced education, and job seeking to the career and housing paths associated with their movement into more stable relationships.  As they age, it is highly likely that suburban preferences re-emerge, sustained by the capacity to purchase and operate a private vehicle.

Generation X boosted inner city dwelling over the past two decades, and Generation Y will do so, to a lesser extent, for another decade.  The 15 to 24 year age group also coincides with the age of greatest automobile independence (illustrated for New Zealand in Figure 3).  But don’t expect this historically-specific phenomenon to sustain some sort of indefinite culture of city consolidation, and I wouldn’t bet the fiscal bank on expensive transit systems designed around the assumption that it will. 

These are passing generations: their successors will be that much smaller and facing a somewhat different world.[4]

Figure 3: Use of non-Automotive Modes by Age Group, New Zealand 2004-2008


Who are we planning for?
Of course, there are plenty of exceptions to prove the rule: but that is the point.  Diverse communities have diverse expectations and behaviours.  And they are continuously changing, in composition, in form, and in behaviour. 

The failure of modernity lay in its assumption of conformity and convergence, compounded by the conceit that we could regulate for it.  And planning for what is little more than a statistical construct – the auto-independent city – risks blinding us to the richness and opportunity of alternatives, of lifestyle, of environmental stewardship, of urban design, and of mobility.

If we start with the behaviour of individuals and households our designs for sustainable cities may be less deterministic and our planning less didactic, better informed, lighter in touch, and a lot more effective in meeting the long-term needs of evolving urban communities.


[1]           Newman, P and Kenworthy J (1989) Cities and Auto Dependency: A Sourcebook. Gower, Aldershot
                Newman, P and Kenworthy, J Sustainability and Cities: Overcoming Automobile Dependence, Island Press, Washington, D.C.
[2]        Newman P and Kenworthy J (2011) “‘Peak Car Use’: Understanding the Demise of Automobile Dependence”, World Policy Transport and Practice, 17, 2, 31-42
[3]        Pynoos R, Caraviello R, and Cicero C (2009) “Lifelong Housing: The Anchor in Aging-Friendly Communities”, Journal of the America Society on Aging, 33, 2, 26-32
[4]         For New Zealand, check the numbers

Wednesday, March 21, 2012

Can we do Bigger Better? Firm Size and the Quest for Productivity

Big enough to prosper
Look at this from the Economist:

Britons and Americans are used to lionisations of the small businessman. This praise is often misplaced; it is not so much small firms that drive growth and job creation so much as small and young firms on their way to becoming much larger. Where small firms are most common, as around Europe's southern periphery, their prevalence is sign of uncompetitive markets and low productivity.

The Economist suggests that the key to productivity is letting companies grow, because bigger companies deploy capital and labour more effectively.  We might add, from our remote New Zealand perspective, that size also gives firms the ability to work offshore markets, expanding beyond a constrained  domestic market.  

In this way, size begets size.  And big firms provide the seed bed for innovation (as I argued in Growing a productive urban economy) creating a virtuous cycle - at least up to a point. (For very large organisations there often comes a time when management diseconomies reverse efficiency gains).

What Role SMEs?
In New Zealand we extol the small and medium enterprise sector (SME) when maybe we should be doing what we can to get SMEs to bust out and grow.  And it’s fashionable to argue that if more were to locate within Auckland– our primary city - businesses could reap external economies of scale (so-called agglomeration economies), accessing more labour, more skills, and more services.   

But it’s more fundamental than that.  Rather than promoting lots of small business units sharing services and competing in a crowded labour market we really need to do what we can to ensure that more are actually growing.

The Size of New Zealand Business Units
The rest of this post looks at the size distribution of New Zealand business units using Statistics New Zealand February 2011 data. It then looks for evidence that location in Auckland might favour manufacturing by favouring bigger firms. 

The statistics tell us that 65% of New Zealand’ Geographic Business Units didn’t employ anyone.  (A GBU is defined by Statistics New Zealand as “a separate operating unit engaged in New Zealand in, or predominately one, kind of economic activity from a single physical location or base.) Let’s set them aside.

So what about the other 35%?  Well, the bad news is that New Zealand’s business units are underwhelmingly small.  63% of them provide 5 or fewer jobs, 78% fewer than ten! 

Looking at the other end of the spectrum, of 175,150 GBUs only 2,425 (1.4%) employed more than 100 people, and another 3,540 employed between 50 and 100 (2.0%).  Between them, though, these bigger units provided 44% of the country’s jobs.

The Analysis
I looked at the size distribution of business units for two “regions” – Auckland (New Zealand’s only city of over 1 million people) and the rest.  Auckland accounts for 29% of the country’s business units and 33% of its employment. 

I plotted the share of GBUs and employees across six size categories for Auckland and the rest of New Zealand.  The share of  business units in a given size category is plotted in the first two bars and the share of employment in the second two for each of the six size categories.  And in each case the share of the country’s units or employment outside Auckland (Rest of New Zealand) is plotted first ( in blue), to the left of the share inside Auckland (in red). 



So what does the data tell; us?  Business units in Auckland tend to be slightly larger. Even though over 75% still employ fewer than ten people (compared with 78% elsewhere) there is a marginal bias towards larger organisations, with 35% of Auckland jobs in units employing over 100 people compared with 31% for the rest of New Zealand.

This is consistent with the propensity of businesses in a dominant city to be larger, given a labour resource that can support greater expansion and a larger local market.

Digging Deeper
This exercise has been repeated for manufacturing, and within manufacturing for the dominant food processing sector (67,850 employees nationally) and for the next three sectors, equipment and machinery manufacturing (26,060 employees), metal products (21,650), and wood products (15,970). The charts appear at the end of this post.

Manufacturing
In manufacturing the picture changes.  The rest of New Zealand category has a greater proportion of both small and large enterprises than Auckland.  Consequently, only 32% of Auckland’s manufacturing employees are in firms with over 100 employees, compared with 42% elsewhere in the country.

Apparently the benefits of urban scale do not translate into larger manufacturing enterprises, at least at this level of generalisation.  This  raises doubts over the productivity of firms in the city,  and is consistent with my earlier evidence that concentration of a sector in  Auckland does not necessarily favour its growth .

Food Manufacturing
This anomaly is explained in part by the nature of the food products sector, which accounts for 36% of all manufacturing employment in New Zealand.  It is dominated by large dairy and meat processing works in rural areas, small towns, or provincial cities.  The nature of these long-standing and globally competitive primary processing activities mean that efficiencies accrue outside major urban centres in factories that tend to be labour, land, and capital intensive. 

Wood Products
Wood product manufacturing is much the same.  The 21% of New Zealand’s units located within Auckland tend towards medium size (10-50 employees).  Elsewhere in the country 27% of units of employ over 100 people.

So scale, expertise, and a commitment to exporting in primary processing mean that productivity in manufacturing may be healthiest outside  Auckland.  It is also reflected in a heavy commitment to exporting from  secondary centres.  Despite concerns that New Zealand food and wood processing do not produce a lot by way of highly transformed, high value products, sustaining and promoting their scale outside the main urban areas has been critical to their continued competitiveness.

Certainly Auckland (and Christchurch) plays an important role in these  industries, first through housing  higher order or specialised services they might draw on, including logistics, finance, legal services, and research; and, second, as a  location for smaller spin-off firms that undertake more specialised manufacturing and marketing using primary products as their raw materials.

The urban manufacturers
What about sectors that entail a greater degree of product transformation? Metal product and machinery manufacturing fit this description and are, by contrast with primary processing, urban activities.  In New Zealand, though, they are modest both in scale and performance. With a small number of notable exceptions there are few units capable of competing internationally.

Metal product manufacturing tends to take place mainly in medium-sized units, from 10 to 100 employees, particularly in Auckland, which dominates the sector.

The manufacture machinery and equipment has a higher proportion of very small units, and more employees in the small number of large units.  Again, Auckland accounts for a disproportionate share of national business units and employment.

The size distribution of units in these sectors, then, may be tied to the  scale of the  markets they are located  in – making them relatively poor prospects for productivity-driven international sales.  Their capacity to move beyond small scale is most likely limited.

So what does it mean?
If scale is a condition of productivity (and exporting), New Zealand industry faces a major disadvantage.  That’s hardly news.  So far, it has not achieved a lot by way of global performance outside primary processing sectors where scale and accumulated expertise build on a natural production advantage.  That’s not new, either. 

Making the point that in New Zealand size, innovation, and productivity happen outside Auckland is an important reminder, though, that much as we might want to pursue planning and policy fads based around urban agglomeration and density, that’s unlikely to offset our intrinsic disadvantages of small scale and remoteness. 

So what can we do?
According to the World Bank rankings New Zealand is already third in the world for ease of doing business.  Perhaps the answer is to ensure that it is also easy to grow a business here. 

Among other things, a policy fixation which promotes places as the fonts of productivity and innovation – and Auckland as the solution to boosting New Zealand’s economic standing -- may have to change.  The reality is that productivity is associated with organisations –  large and growing organisations, and their capacity to engage in networks ("production and distribution chains") of growing businesses.

The challenge for urban policy makers is to ensure that local conditions, the rules, regulations, and charges  affecting firms, and  the quality and availability of land, labour and capital  do not impede local investment and growth.  How this might be done should  be central to the Auckland Spatial Plan.  I see little evidence that it is.