Are driverless cars a game changer?

A common observation by many historians I’ve read goes like this: “they failed to understand just how important such-and-such was going to be in the future”. In many cases, “such-and-such” is a decisive technology that went unrecognized until it ended up completely changing the game.

Well, I think one technology that’s being grossly under-estimated today is Driverless Cars (DCs). If they could deliver fully on their promise, they’d have an enormous impact and bring a triple bottom line improvement to our cities – more efficient, more equitable and better for the environment.

There’s plenty of commentary around on driverless cars and I wrote at length on their potential as recently as May 31, in Are driverless cars coming?  In that piece, I discussed the current state of the technology and some of the formidable technical, social and legal obstacles to a driverless car fleet.

However as we know from the history of electricity, public sanitation, the car, the computer, inoculation, the pill and many other innovations, it’s very hard to deny an irresistible idea. Given enough time – say the 30 year horizon typical of current planning strategies – it’s possible the sheer weight of benefits DCs promise our cities will provide the motive force to overcome these obstacles.

The key potential benefits are:

  • Expansion in the effective capacity of the road system – at least double and perhaps eight times as much, with consequent savings in infrastructure provision
  • Time savings from faster journeys – technology can manage vehicle interactions and speeds more efficiently than human drivers (although there might be a trade-off with capacity here)
  • Almost complete elimination of serious injuries and fatalities associated with accidents
  • More productive use of in-vehicle journey time compared to conventional cars
  • Greater mobility for those who cannot drive e.g. the unlicensed, disabled, drunk

These are potentially enormous private and social benefits. In addition, the warrant for owning a private vehicle would be greatly reduced in a world of DCs. If a total or substantial shift to DC-sharing were achieved, the size of the urban car fleet would be reduced by an order of magnitude. There would be many benefits:

  • Lower environmental impact because many fewer vehicles would need to be manufactured
  • Less public and private space devoted to parking – this could greatly enhance the quality of public spaces and even residential streetscapes
  • Better matching of vehicle type to need, resulting in lower resource and environmental costs e.g. many DCs could be single seaters
  • Lower cost of travel due to eliminating need for vehicle ownership and removing the “status” component
  • Reduced noise, pollution, emissions and energy consumption by virtue of having a more efficient “standard” set of vehicles
  • The opportunity to rationalise the way travel is paid for by introducing a new pricing ‘paradigm’ – all standing and variable costs, including externalities, could be incorporated in a distance-related tariff (this isn’t intrinsic to DCs, but the changeover to a new paradigm provides the opportunity)

There are other potential strategic benefits too. Driverless cars could greatly reduce (though not eliminate) the need for public transport. This would offer a number of potential advantages:

  • Faster, safer and more private travel for those who currently use public transport – many travellers would enjoy very significant time savings
  • A higher proportion of the total cost of providing transport in the city could be borne directly by DC users rather than, as at present, by taxpayers

DCs aren’t just a replacement for the car, they’re a potential game-changer for the entire urban transport task. Read the rest of this entry »

Is The Age providing fair comment on transport issues?

Prospective corridors for road tunnel (Eddington Report)

I take an agnostic view of freeway proposals – I don’t assume apriori that they’re all bad or all good. I prefer to look at the evidence first before deciding if a proposal has merit or is a poor idea. But it seems there are some who will overlook evidence to the contrary if it undermines their ideological view.

Like Kenneth Davidson in his column in The Age on Monday, Why the east-west road tunnel is a stinker, I have some misgivings about the tunnel proposed to connect Melbourne’s Eastern Freeway with the Western Ring Road. The Victorian Government has just submitted a proposal to Infrastructure Australia, seeking funding to develop the tunnel idea further.

My key concern is the anaemic benefit-cost ratio. But Mr Davidson, who’s a senior columnist at The Age, goes further. He reckons the proposed tunnel recommended in 2008 in the Eddington Report is a “stinker” and a “confidence trick”.

That’s because an earlier study undertaken for the Bracks government in 2004, the Northern Central City Corridor strategy (NCCC), found most of the traffic coming off the Eastern Freeway heads into central Melbourne. It found only 15% is bound for the northern or western suburbs. “In other words”, Mr Davidson says:

the public justification for the east-west link – that it would take traffic away from the central business district – was a confidence trick……The first question (Eddington) should have asked was where did the 2004 study go wrong.

I don’t think anyone disputes the NCCC study was negative about the case for the tunnel. Nor is Mr Davidson the first to raise this objection. The “gotcha” Mr Davidson seizes on with such alacrity is that Sir Rod Eddington apparently ignored the NCCC study’s key finding.

But it seems it’s Mr Davidson who’s doing the ignoring. The Eddington Report actually does consider the NCCC study. Moreover it deals with it in a way that is prominent and impossible to miss by anyone with their eyes open (read it – Chapter 5, page 129).

The Report argues it’s a myth that nearly all Eastern Freeway traffic is destined for the inner city. It says the NCCC produced diagrams that present “a distorted view of traffic distribution (and further NCCC modelling for a future link would have identified and addressed this issue)”.

In a section titled, ‘Myth 2: nearly all the Eastern Freeway traffic is destined for the inner city’, It argues the NCCC study didn’t look beyond the capacity of existing roads or the ultimate destination of traffic once it left the NCCC study area.

First, given the roads in question, the traffic distribution (identified in the NCCC study) is not surprising: at the end of the freeway, there are ten freeway standard traffic lanes (five each way). By the time traffic reaches Macarthur Avenue in Royal Park, the corresponding ‘connection’ is a two-lane road (one lane each way). The traffic distribution is as much a function of the roads available, which progressively reduce in capacity towards the west, as it is a reflection of the demand for a particular direction of travel.

Secondly, when the (Eddington) Study Team analysed how traffic from the Eastern Freeway is distributed (with the analysis closely matching the NCCC distribution), it revealed that around 40 per cent of the daily traffic from the freeway travels beyond the central city area – to the south and the west. That is the case with the current network: in the future, EastLink will add a new dimension.

The Eddington Report also argues (page 137) the NCCC study focussed on Eastern Freeway traffic and didn’t fully consider traffic using adjacent streets instead. Moreover, it didn’t recommend against the tunnel because insufficient vehicles would use it, but rather because the high cost of construction yielded an inadequate benefit-cost ratio. Read the rest of this entry »

Are public transport trips longer (than car trips)?

Non-work trip distances (km) by car and public transport (%) - data from VISTA

I’ve frequently mentioned that trips by public transport in Melbourne are longer on average than those by car. The exhibit above illustrates that difference using data from the Department of Transport’s VISTA data base.

It shows the length of non-work trips in Melbourne by private vehicles versus public transport. For all practical purposes, it’s cars on the one hand (in red), versus trains, trams and buses combined on the other (in blue).

A ‘trip’ refers to travel between main activities. Where multiple modes are used on one trip a single ‘main mode’ is defined. For example, if a person drives from home to the station, catches a train to the city, and walks to their destination, their mode for the trip is the train.

It’s important to get these figures in perspective from the outset. There are almost fourteen times as many non-work trips made by car as are made by public transport. This reflects the dominance of private transport in Melbourne.

It’s evident from the exhibit non-workcar trips are substantially shorter – just compare the two fitted curves. In fact 56% are less than 5 km, whereas the corresponding figure for public transport is 27%. At the other end of the scale, 5% of non-work car trips are longer than 30 km, compared to 9% of public transport trips.

Work trips (i.e. commutes) tend to be longer than non-work trips in the case of both private and public transport, but again public transport trips are longer on average. For example, 40% of commutes by public transport exceed 20 km. The corresponding proportion for car commutes is much lower, at 27%.

I’ve shown non-work trips in the exhibit because they account for around four fifths of all trips (defined as per above). They are not as dominant in the case of public transport, but still comprise two thirds of all trips by trains, trams and buses.

Trains are the key reason public transport trips are so much longer – people travel considerably further on metropolitan trains on average than they do on buses and trams. I don’t have data to hand that separates out trains (I hope to get my hands on it shortly), but the difference between cars and trains would be much more dramatic.

It’s not that the train out-competes the car for long trips. In fact many more people still prefer to use the car for long trips than take any form of public transport. For example, while 7% of non-work car trips are over 25 km compared to 14% of public transport trips (they’re virtually all train), the car trips number 534,000 per day. The comparable number of trips made by public transport passengers is just 80,000.

Rather, people who choose to travel by train are more inclined to use it for longer distances, on average, than car travellers. I’d say that’s largely because Melbourne’s train system is designed to transport people across long distances – primarily from the suburbs to the city centre.

Cars are flexible and in non-congested conditions can get travellers to a range of dispersed destinations, both near and far. Melbourne’s train system on the other hand is relatively inflexible by comparison. It essentially leads to one destination. Read the rest of this entry »

Are these really the most (and least) liveable suburbs in Melbourne?

Liveability index for 314 Melbourne suburbs (from The Age)

I never read those glossy magazine inserts in The Age (who does?) but on Friday I made an exception for The Melbourne Magazine because it promised to tell me “the most liveable suburb in the world’s most liveable city”.

The Age’s Our liveable city project ranks the “liveability” of 314 suburbs from top to bottom and claims to reveal “Melbourne’s best suburbs, the ones improving the most, and where you should buy your next home”.

Liveability is defined initially as “the general quality of a place that makes it pleasant or agreeable for people to reside in”. Fortunately, someone saw that was a tautology and wouldn’t be of any practical help. So liveability is subsequently defined by 13 measurable criteria.

These cover topography, traffic, crime, cultural facilities and parks, as well as proximity to a range of amenities – the CBD, beaches, public transport, schools, restaurants and shops. Scores out of five on each criterion are added to give an overall summary score for each suburb. Each suburb is subsequently ranked from 1 to 314 on an Index of Liveability.

Unsurprisingly with this sort of exercise, there was a lot of criticism from readers, with many pointing to apparent anomalies in the rankings. One said, “once I saw Footscray was rated higher than Middle Park I stopped reading”.

Disagreement is inevitable. People are different and so it’s hard to get consensus on just what does, or doesn’t, make a place liveable. That shouldn’t be surprising – an elderly couple, for example, is likely to have a very different definition of liveability to that of a young single. Throw in further differences, say in education, income or ethnicity, and it gets much more complex.

There is a much more straightforward and reliable way of establishing the relative liveability of suburbs. That simply involves measuring what people are prepared to pay to live in them i.e. property prices. It doesn’t require complex measures and weights (not that The Age bothered with the latter). In fact it sidesteps entirely the hardest and most intractable question of all – defining apriori just what liveability is.

Moreover, it provides a clear ranking and allows us to measure the size of the difference between suburbs. And it’s based on the actual decisions of hundreds of thousands of householders. Knowing that the average property value in South Yarra is three and a half times higher than in Hallam is a much more useful and valuable piece of information about the relative merits of the two suburbs than knowing one ranks 350 places ahead of the other on the Index of Liveability.

There’re distortions in the market so property values aren’t a perfect representation of the relative liveability of Melbourne’s suburbs. I would argue however that this approach involves infinitely fewer compromises than the methodology used by The Age. Of course it wouldn’t make a very interesting story when there’s the option available to the newspaper of bringing some “science” to the issue.

That’s not to say exercises like the Index of Liveability don’t have value. They can be useful to establish just why residents think one suburb is more liveable than another. Knowing why properties average $1,300,000 in South Yarra and $369,000 in Hallam would be very important information for policy-makers.

Interpreted this way, I think The Age’s attempt is actually better than many of the commenters are prepared to concede (even though I’m annoyed that very little information about the methodology is disclosed). You can argue the toss at the margin – for example, I suspect proximity to schools only matters in the case of certain institutions – but by and large the criteria are a reasonable compromise. Read the rest of this entry »

What does ‘random’ look like?

Which plot is random?

The exhibit shows two seemingly similar patterns – but one of them is random and one isn’t. Can you tell which is which? More in a moment.

I’ve taken these plots from Steven Pinker’s new book, The better angels of our nature: why violence has declined. It’s in a Chapter titled The statistics of deadly quarrels where he discusses the statistical patterning of wars and takes a small detour into “a paradox of utility”, specifically our tendency to see randomness as regularity with little clustering.

This cognitive illusion has relevance to all disciplines but is of particular interest to anyone interested in spatial issues, as a couple of these examples show.

Professor Pinker, who’s a psychologist at Harvard, cites the example of the London blitz, when Londoners noticed a few sections of the city were hit by German V-2 rockets many times, while other parts were not hit at all:

They were convinced that the rockets were targeting particular kinds of neighborhoods. But when statisticians divided a map of London into small squares and counted the bomb strikes, they found that the strikes followed the distribution of a Poisson process—the bombs, in other words, were falling at random. The episode is depicted in Thomas Pynchon’s 1973 novel Gravity’s Rainbow, in which statistician Roger Mexico has correctly predicted the distribution of bomb strikes, though not their exact locations. Mexico has to deny that he is a psychic and fend off desperate demands for advice on where to hide

Another example is The Gambler’s Fallacy – the belief that after a long run of (say) heads, the next toss will be tails:

Tversky and Kahneman showed that people think that genuine sequences of coin flips (like TTHHTHTTTT) are fixed, because they have more long runs of heads or of tails than their intuitions allow, and they think that sequences that were jiggered to avoid long runs (like HTHTTHTHHT) are fair

The exhibit above shows a simulated plot of the stars on the left. On the right it shows the pattern made by glow worms on the ceiling of the famous Waitomo caves, New Zealand. The stars show constellation-like forms but the virtual planetarium produced by the glow worms is relatively uniform.

That’s because glow worms are gluttonous and inclined to eat anything that comes within snatching distance, so they keep their distance from each other and end up relatively evenly spaced i.e. non-randomly. Says Pinker:

The one on the left, with the clumps, strands, voids, and filaments (and perhaps, depending on your obsessions, animals, nudes, or Virgin Marys) is the array that was plotted at random, like stars. The one on the right, which seems to be haphazard, is the array whose positions were nudged apart, like glowworms

Thus random events will occur in clusters, because “it would take a non-random process to space them out. The human mind has great difficulty appreciating this law of probability”. Read the rest of this entry »

Could major housing developments be outside activity centres?

We need to start thinking about new ways of increasing housing supply in the established suburbs. As I’ve noted a number of times now, activity centres aren’t delivering much and infill housing, though it’s putting in a sterling effort, is probably at full stretch. These strategies are still important, but additional sources of supply are needed.

Much of the current thinking confines major developments to a narrow range of locations, primarily activity centres and major transport corridors where key infrastructure, particularly rail lines, already exists. Many have argued these locations have the potential to accommodate enough apartments to house all of Melbourne’s projected growth for decades to come. I don’t doubt they could in theory, but in reality they’re not doing enough.

I think there are lessons we can learn from the study of infill housing I discussed last time. One is the decisive importance of land – developers working at all scales need to be able to easily acquire or assemble sites that are of an appropriate size and aren’t encumbered by high-value improvements.

Developers don’t want to be rejected or delayed by unhappy neighbours or sent packing by councils that impose restrictive limits and conditions on development. And they need to offer housing types that are attractive to the market in that locality – not everyone who’d be inclined to live in the middle ring suburbs would want to live in an apartment.

I think there’s another option worth investigating that addresses many of these concerns. Although it’s just an idea at the moment, it involves a reverse strategy – encouraging residential and commercial development in areas that usually have poor infrastructure, but large sites and compliant neighbours. Rather than require infrastructure to be in place first, it involves retrofitting infrastructure like public transport in order to create new living areas.

It relies on the existence of sites in the suburbs which are large, in single or limited ownership, have few neighbours, and are either undeveloped or have relatively low-value improvements such as warehouses. Some sites that meet these criteria are attached to large public sector organisations like tertiary institutions and utilities, but are surplus to requirements.

As an example, I’m familiar with a tertiary institution in an Australian capital city that’s located within 20 km of the CBD and has more than 100 ha of surplus developable land (that’s five times the size of Melbourne’s E-Gate). It is in a single ownership and already has the extensive infrastructure and services required to serve a student population and workforce of many thousands.

It’s in an attractive environment and could potentially be redeveloped as a major regional activity centre at relatively high densities. It isn’t near a train line, but is well-connected by buses to the rail network with the potential to add more to suit the needs of a very large permanent resident population. I’m aware professionally of some opportunities in Melbourne that are consistent with this example.

However most of the prospective sites are likely to be used at present for storage, distribution and manufacturing activities. They’re not brownfields sites though, because they’re not disused. The idea is that rezoning would provide owners with an incentive to redevelop their properties for intensive residential and commercial uses.

The extended area running through Clayton/Monash/Glen Waverley is a possible candidate in Melbourne. It has a number of large industrial sites that could potentially be redeveloped given an appropriate inducement. This region is particularly important because it already has the largest concentration of jobs in suburban Melbourne (albeit at relatively low density compared to the CBD) and is a prime candidate for development as a Central Activities Area.

While it isn’t well-served by rail, it could potentially be retrofitted as part of the proposed Rowville rail project. More plausibly, it could be serviced at a lower up-front cost by an expanded bus rapid transit system similar to that now connecting Monash University with Glen Huntley Huntingdale station.

The overall idea of retrofitting relies on the sheer size and intensity of projects to provide the incentive for redevelopment and to justify investment in infrastructure. The scale of projects would also be a key way of differentiating developments from neighbouring uses, some of which may retain their non-residential character, at least in the short term. Read the rest of this entry »

Can Melbourne depend on infill housing?

Geography of infill developments in City of Monash, 2000-2006 (source: Phan et al, 2008 - graphic via Grattan Institute)

We know that the inability to increase significantly the supply of dwellings within established suburbs is a key failing of strategic planning in Melbourne. Simply put, there’s not enough housing to make established suburbs affordable for all the people who would like to live in a relatively accessible location.

We also know that activity centres aren’t pulling their weight in the task of increasing supply (see here and here) and that the burden of supply is instead falling on small-scale infill development, much of it dual occupancy projects. So it’s worth looking further at the nature of infill housing.

A study by Monash University’s Thu Phan, Jim Peterson and Shobhit Chandra , Urban infill: the extent and implications in the City of Monash, examined new developments in the municipality over the period 2000-06. They defined infill primarily as projects where two or more new dwellings were constructed on sites formerly occupied by detached houses. A total of 1,483 projects were identified, ranging in size from two dwellings to 178.

The study revealed a number of interesting aspects about this middle suburban municipality.

First, it found new dwelling supply is dominated by small projects. One project built more than 178 dwellings and three built between 40-77 dwellings, however 98% of projects involve just 2-7 seven dwellings (and we can be pretty confident they’re heavily weighted toward the smaller end).

Second, projects are dispersed, not concentrated. As shown in the exhibit, proximity to major trip generators is uncorrelated with location of projects. Just 5% are within 400 metres of a Principal, Major or Specialised activity centre, and only 10% are within 400 metres of a rail station. Moreover, the authors found projects located within 400 metres of an activity centre are smaller on average than those in more distant locations.

Third, developers tend to be opportunistic rather than strategic – they wait for properties to be offered for sale and assess each one on its potential for redevelopment. Thus the geography of infill development is shaped largely by what comes on the market rather than by any sort of deterministic planning policy.

Fourth, the size of lots and the age of the existing house is a more important influence on the location of infill development than proximity to an activity centre or rail station. The average infill site is relatively large (700 to 900 sq m) and the majority of existing dwellings are relatively old i.e. built between 1945 and 1965. Lot sizes close to rail stations are smaller – and hence less amenable to redevelopment – than those further away, probably reflecting the different periods of development.

Thus not only are activity centres failing to expand housing supply in accordance with the precepts of Melbourne 2030, but the great bulk of new housing being built in Monash isn’t located close to activity centres but rather is dispersed (relatively uniformly too judging by the exhibit i.e. non-randomly).The dispersed pattern will worry some, but I don’t see it as a big issue. Read the rest of this entry »