[sustran] A problem with infrastructure studies

Kerry Wood kerry.wood at paradise.net.nz
Wed Jul 12 13:31:33 JST 2000


In this neck of the woods a problem has recently come up in a transport
corridor study, which might be of general interest. It involved double
counting of roading benefits. 

The regional authority did a corridor study, looking at both road
and rail: full marks so far. This was for a radial route, extending the
motorway section of a highway from the suburbs of a small city, out to a
mixture of suburban and peri-urban sprawl along a narrow (sometimes very
narrow) corridor between hills and the sea.

The existing road and rail links follow the coast but the proposed
highway runs inland for some 20 km, bypassing some settlements.

The study had some rather shaky-looking features:

-	15% of the roading costs (on the existing highway) were included in the
'Do minimum' case, and so excluded from the cost side of the
cost:benefit equation.

_	The new highway was included in all the roading cases modelled, except
'Do minimum'.

-	The new highway will remove only about 20% of traffic from the old
highway, and this will be replaced by growth fairly quickly.

-	Looking at the capital costs and the congestion savings to road
users, suggests that the rail options are very good. The smaller one
gives two thirds of the road congestion savings of the highway for one
twelfth of the cost, and the more radical rail scheme gives twice the
road congestion benefits for half the cost of road building. An
intermediate scheme might be better than either of the modelled proposals.

The unfavourable results for the rail-only options are supposedly
because they will need subsidies. However, the present value of rail
costs is not so very different from the capital costs, suggesting that
the (unstated)
subsidies are reasonable. 

However, there does seem to be one major discrepancy: the tolls and
petrol tax used to partially fund the new highway are treated as income.
As a colleague points out, there has been a serious confusion between
two questions: 

-	What is the return on investment? (looking at the costs and benefits
to the region).

-	How will the investment funds be raised? (looking at the problem of
funding an infrastructure project). 

The regional council should have started by asking the first question:
What is the return on investment? 

They would then have followed general practice in assessing roading
benefits, plus an adaptation of that practice to cover rail passengers.
Tolls, petrol tax and perhaps rail fares should have been ignored. In
national (or regional) terms they are neither costs not benefits: they
are user
contributions towards costs, and are an internal transfer. They will be
less than the calculated benefits if the project is worthwhile, but they
are part of those benefits: adding them to in to the benefits is double counting.

If the answer to the first question was that there is an acceptable
return on investment, the regional council could then have legitimately asked
the second question: How will the investment funds be raised? 

The objective would then be to fund the construction and operation/subsidy
of road and rail projects. Users could legitimately be charged a
proportion of the benefits that they would receive from the project, so
at this stage the tolls, petrol tax and fares could be considered as
income to the project,
although they would still not be income to the nation or region.

What the regional council did in the event seems to have been to ask
both questions at the same time, so they ended up including tolls,
petrol tax and fares as a benefit to the country/region, in addition to
the benefits calculated using standard methods.

The treatments above are fully appropriate for tolls, but a petrol tax
or rail fares may need slightly different handling:

-	A special-purpose petrol tax is proposed, payable by all who buy
petrol in the region, whether or not they use the new highway.
Purchasers on the other side of the region could reasonably argue that
they gain very
little benefit and should not pay, but the same is true for central area
road users (like me), who seldom use the radial route. This means that
some road users will be subsidising others. This problem could
theoretically be avoided by using tolls for all road user contributions,
but there is a practical problem. If an economic toll were charged, much
of the traffic would continue to use the old road, unless it were tolled
too, but there is no appropriate tolling point: it is used by local
traffic throughout its length. 

-	Rail fares may need to be treated in the same way as tolls, and
deducted from the rail benefits. However, this is less clear than for
tolls because there is no well-established framework for modelling rail
passenger use, and it is not clear from the council's report just what has
been done. 

My actions on finding this problem show just how insidious it can be. I
asked for peer review, and one reviewer pointed out that I had made the
same mistake for rail fares!

Another difficulty with the regional council's approach is that if using
petrol tax to fund a new highway is legitimate, then it is at least
equally legitimate to use a petrol tax to fund rail improvements which
achieve the same objective more cheaply. 


-- 
Kerry Wood
1 McFarlane St, Wellington 6001, NZ
Phone +64 4 971 5549  Mobile  +64 21 115 9346




More information about the Sustran-discuss mailing list