Traffic congestion will continue until we're game to tax it

The governments of NSW and Victoria lost zero time in rejecting the Grattan Institute’s proposal that all state governments introduce “congestion charging” in their capital cities. But don’t imagine this unpopular idea will go away. It will keep coming back until we buy it.

Australians and their political leaders have a record of trembling on the brink for decades before belatedly accepting the inevitability of upgrades to the tax system. Take value-added tax.

A full quarter century passed between the first official report recommending a VAT – which the Whitlam government rejected at the same time it made the report public - and its introduction by John Howard in 2000, rebadged as the more euphemistic goods and services tax.

Economists had lots of fancy economic-efficiency arguments for changing to a broad-based, single-rate tax on consumer spending but, in the end, it was quite pragmatic, revenue-protecting arguments that won the day.

The High Court had ruled various state government indirect taxes to be unconstitutional, and the growth in collections from the federal government’s ramshackle wholesale sales tax was falling further and further behind the growth of the economy, as more of every consumer dollar was being spent on (untaxed) services rather than goods.

An eventual move to charging motorists directly for using the roads could also be prompted by the declining effectiveness of the present tax system. As ever more of our car fleet moves from petrol-powered to electricity-powered, receipts from the nation’s main tax on motoring – the federal excise on petroleum – will wither away.

But that’s not an argument used in the Grattan Institute’s report – written by Marion Terrill – advocating a move to congestion charging. Indeed, Terrill makes it clear she’s not talking about a general road user charge – that is, charging that covers the cost of building new roads and maybe also the costs of wear and tear to roads, accidents and so forth. (Even though such a general charge for road use may well be what we end up with.)

No, Terrill is only on about charges designed to reduce excessive congestion.

So why might we get charges directed solely at reducing congestion? Because all of us hate it so much and because, even if it doesn’t increase in coming years as cities get ever bigger, you can be sure we’ll all believe it’s got a lot worse.

And, finally, because congestion charging is the most certain – and the cheapest – way to actually reduce congestion, not just promise to.

State politicians have gone for decades claiming to be reducing congestion by spending billions on new freeways (and a lot fewer billions on expanding public transport), but it hasn’t happened.

Why not? Partly because our cities keep getting bigger, but mainly because, in Terrill’s words, “most city-dwellers find car travel more appealing and convenient than other means of travel”.

Initially, a new freeway is much faster than the roads it replaces, but that just attracts more people who’d prefer to travel by car. They keep flooding in until the congestion increases the delay to the point where it’s about as bad as it was before.

By contrast, we know that congestion charging really works. You’d still have to build more freeways and railways as the city grew, but many fewer.

Terrill argues that congestion charges could be introduced in three stages. First is “cordon charging” where drivers pay to cross a boundary into a designated zone, such as a CBD. Next “corridor charging,” where drivers pay to drive along an urban freeway or arterial road. Then network-wide, distance-based charging, where drivers pay to drive within a designated network or area, on a per-kilometre basis.

She says there are three reasons why now’s the time to get started. First, many people say that congestion charging couldn’t be introduced without a big improvement in public transport. Well, that’s just what we’re getting.

In recent elections, the winning party promised spending on public transport of $72 billion in Victoria, $42 billion in NSW and $13 billion federally.

Specifically, Melbourne is getting the Suburban Rail Loop and the Airport Rail Link. Sydney’s getting Metro West, Metro City and rail to Western Sydney Airport.

Second, the technology for congestion charging is getting cheaper and better all the time. Third, there’s now enough global experience - not just Singapore, London and Stockholm, but also Malta, Gothenburg and Milan, with Jakarta and New York on the way - to show that congestion charging works and that, despite initial opposition, is soon accepted as a big improvement.

Terrill says that, in Sydney’s morning peak, for example, up to 21 per cent of trips are for “socialising, recreation or shopping”. A congestion charge wouldn’t raise much revenue. It wouldn’t have to be high to deter enough people to reduce road use in key parts of the city during peak hours. And remember, such charges are designed to be avoided.

It’s true that motorists with lots of money could easily afford to pay the charge, whereas people on modest incomes couldn’t. But the claim that a charge would be unfair is exaggerated.

If the charge was imposed on cars entering the CBD, only 3 per cent of Melbourne households would pay it on a typical day. And not many of those would be poor. The median income of full-time workers driving to work in the CBD is $1980 a week in Melbourne (and $2450 a week in Sydney). Sound poor to you?

But if you’re still not convinced by Terrill’s arguments, here’s a more radical proposal. The economists’ Coase Theorem implies it shouldn’t matter whether you impose the charge on workers required to start work in the CBD during peak hour or on their employer doing the requiring.

After all, workers have little or no ability to change the time they must start or leave work, but their bosses do.

Read more >>

Why more expressways don't fix traffic jams

When Marion Terrill, of the Grattan Institute, set out to find out how much commuting times had worsened in Sydney and Melbourne, she discovered something you’ll find very hard to believe. But it would come as no surprise to transport economists around the world.

Everyone is sure traffic congestion has got much worse in recent years. This is only to be expected since Sydney’s population grew at the annual rate of 1.9 per cent, and Melbourne’s rate grew even faster, at 2.3 per cent, between the censuses of 2011 and 2016.

Both cities have grown much faster than the Australian population overall. People are crowding into our big cities, much to the disapproval of many people already living there.

Why are they piling into already-crowded cities? For reasons economic geographers call “economies of agglomeration”. One way for countries to get richer is for their businesses to pursue economies of scale; another way is for businesses and their workers to pursue the gains from agglomeration – a fancy word for piling things together.

There are three kinds of agglomeration economies. They come from matching (in a big city, people are more likely to find a job, while businesses are more likely to find the particular workers they need; there can be greater specialisation), sharing (less idle capacity in, say, car parks, or waiting around for customers), and learning (more workers for you to see and imitate; knowledge and know-how shared face-to-face).

Sharing, matching and learning can occur in two ways. When a lot of firms in the same industry gather in the same city, or just because a lot of people and firms are located together, making the city large enough to justify, for instance, heart and lung transplant centres.

Of course, along with the great benefits of crowding together go the costs of crowding together - such as feeling terribly crowded.

There are more people per square kilometre living in the centres of our big cities than there were five years ago. Sydney’s population density has increased by 23 per cent – and Melbourne’s by a mere 46 per cent.

And surely more crowding means more traffic congestion. But this is where Terrill and the co-author of her report, Hugh Batrouney, found their first strange fact. Between the last three censuses, from 2006 to 2016, there’s been virtually no change in the distance between where people live and where they work, measured as the crow flies.

Next surprise came from the HILDA survey – household income and labour dynamics in Australia – which, among other things, asks people how long they spend commuting.

In the four surveys between 2004 and 2016, for both Sydney and Melbourne there was no change in the fact that a quarter of workers had one-way commutes lasting no longer than 15 minutes. One half of workers had commutes no longer than 30 minutes.

When you take it up to the experience of three-quarters of workers, there was some increase over the years in Sydney, but only a small increase in Melbourne. Other figures, from Transport for Victoria, tell a similar story.

So, we all think the increasing traffic volume is leading to greater delay and, hence, longer commute times, but the best available actual measures of commute times say they’re little changed.

Find that hard to believe? Well, as I say, few transport economists would. Why not? Because it fits well with what they call “Marchetti’s constant”. Marchetti was an Italian physicist credited with discovering the empirical truth that the average time spent by a person on commuting is about an hour a day – 30 minutes each way.

The amazing truth of this “constant” has been shown by many studies of many cities around the world.

And it fits with another empirical regularity known as the “Lewis-Mogridge position”, formulated by those gents in 1990: “traffic expands to meet the available road space”.

The government notices that traffic is particularly congested on a certain road, so it builds a big new expressway. When it opens, the time taken to get from A to B falls dramatically. But when people realise this, more of them stop travelling to work by public transport and start going by car.

So many people do this that the speed gain disappears within months, even weeks. The time taken to get from A to B goes back to about what it was before the expressway was built.


The only change is that a higher proportion of workers are able to go by car. The traffic jam is often just shifted to another place on the road network.

Getting back to road congestion in Sydney and Melbourne, how can the gap between what we think has happened and what actually happened be explained?

One possible part of the explanation is that although the traffic really is heavier, making trips less pleasant, this doesn’t prolong the time of the trip as much as we think it has.

But the main explanation – both in Oz and in other countries – is that commuters adapt to the greater congestion.

They take evasive action by moving to a job that’s closer to home, or moving to a home that’s closer to the job. Or they stop going by car and start using public transport.

One thing that really has changed with our bigger cities is more crowded trains and buses.

It’s as though each of us has our own internal, unconscious regulator that draws the line at 30 minutes and, when that limit is exceeded, prompts us to take steps to get travel times back down to where they should be.

Terrill and Batrouney are clear on this: in neither city was enough new infrastructure built between 2011 and 2016 to explain why the huge population growth didn’t lengthen commute times.

The government didn’t fix it, you and I did. Which says we ought to be wary of thinking the obvious – and only - solution to greater crowding is greater spending on transport infrastructure.

Read more >>