PDF: 1832 KB - Bureau of Infrastructure, Transport and Regional ...


PDF: 1832 KB - Bureau of Infrastructure, Transport and Regional ...

Appendix B

Measuring congestion


Congestion costs are measured in many different ways. The measurement is

important because it strongly influences both the estimate of congestion costs and

the ranking of policy options to deal with congestion (as discussed from page 3).

Equally important is the uncertainty ranges for each measure. While uncertainty is

not systematically reported, a recent analysis of the congestion costs in Australia

illustrates its significance:

… the costs imposed on Australian society by urban traffic congestion are likely to fall

in the range of $5 to $15 billion for current levels (BTRE 2007, p. 18).

There is no one definition of congestion costs and no agreed measure of congestion:

each approach has its strengths and weaknesses. The relevance of any measure is

determined by the end use. This appendix explores the different approaches to

measuring congestion costs with particular emphasis on the implications for policy



Methodology and assumptions

Aggregate versus network modelling

Congestion cost estimates are usually developed using complex models incorporating

a variety of simplifying assumptions and approximations. The models themselves

may be ‘top-down’ (aggregate) or ‘bottom-up’ (using detailed network models).

The distinction between the two approaches reflects, in part, the trade-off between

accuracy and practicality.

Network models, including microsimulation models, provide the greatest level

of accuracy as they aim to replicate the traffic flows on all of a city’s major roads.

However, they are very ‘data hungry’ and computationally demanding. Given the cost

involved, they invariably raise the benchmark question of whether the costs can be

justified by the output.

Aggregate models are less data and computationally demanding, but accuracy is

sacrificed to the point that the usefulness of the results is in doubt:

‘… in general, an aggregated analysis that does not consider congestion separately

on each road link is unlikely to provide a credible estimate of the costs of congestion’

(BTCE 1996b, p. 478).


More magazines by this user
Similar magazines