skip to Main Content

A NOTE ON THE REDUCTION IN ECONOMIC BENEFITS DUE TO INDUCED TRAFFIC, FOLLOWING ROAD INFRASTRUCTURE INVESTMENT IN

A NOTE ON THE REDUCTION IN ECONOMIC BENEFITS DUE TO INDUCED
TRAFFIC, FOLLOWING ROAD INFRASTRUCTURE INVESTMENT IN

CONGESTED CONDITIONS

L Kane B Eng (Hons) Cardiff M Sc Eng Cape Town

ABSTRACT

This note sets out the rationale for a more rigorous consideration of induced traffic in
transport modelling and economic assessment in congested conditions. Induced traffic is the
term given to the release of suppressed demand for travel because of reductions in user costs.
The need for this adjustment arises from significant economic evaluation errors which can
occur when the impacts of induced traffic on the user costs of existing traffic are not
considered. The note takes widely cited work from the UK, and applies the lessons to South
African practice.

It explains, from first principles, the interplay between supply and demand curves, the notion
of induced traffic and how South African practice has traditionally dealt with this. It goes on
to detail UK research which demonstrates that the South African approach could result in an
overestimate of benefits in the order of 20-30%. A brief discussion of the future importance of
this matter is given.
Keywords: Induced traffic, economic evaluation, congestion, elasticity,

Kane, L (2006)
1

A NOTE ON THE REDUCTION IN ECONOMIC BENEFITS DUE TO INDUCED
TRAFFIC FOLLOWING ROAD INFRASTRUCTURE INVESTMENT IN

CONGESTED CONDITIONS

Introduction

This note sets out the rationale for a more rigorous consideration of induced traffic in
transport modelling and economic assessment in congested conditions. Induced traffic is the
term given to the release of suppressed demand for travel due to reduction of user costs. The
need for this adjustment arises from significant economic evaluation errors which can occur
when the impacts of induced traffic on the user costs of existing traffic are not considered.
The note takes widely cited work from the UK, and applies the lessons to South Africa
practice.

If these issues are not considered, then significant overestimation of benefits from road
investment can occur.

Background

A typical simple approach to economic assessment is described in Figure 1, where user costs
per trip are shown in relation to volume of trips from point a to b.

Kane, L (2006)
2

The user costs have decreased from C0, the do-nothing scenario, to C1 in the do-something
scenario. The demand for the facility is assumed in this case to remain constant (as is the case
with a simple transport model, with no land-use transport interaction, and only growth
exogenous to the scheme taken into account). This assumption that demand for travel is
unaffected by changing user costs is the so-called fixed-matrix approach, and is standard
practice in models used in South Africa today. The so-called demand ‘curve’ which represents
the fixed matrix is in this case a straight vertical line. In this fixed matrix case, the estimate of
benefits is simply the shaded area in the graph, extended over the evaluation period and
discounted to some chosen year.

As we know, life is not as simple as the fixed matrix approach presupposes, and demand for
transport is elastic; that is, as the costs decrease, the demand increases, and so it is more
realistic to draw the demand curve as shown in Figure 2.

Kane, L (2006)
3

In this case, calculating benefits due to the scheme is simply a case of calculating the areas
shaded. The rectangle indicates the benefits which accrue to existing traffic. The triangle
indicates new traffic ‘induced’ due to elastic demand, and also accruing benefits thanks to the
scheme. (This is informally known as the ‘rule-of-half’.)

All of this is widely understood and documented in South African practice (see, for example,
Pienaar et al. 2002).

Kane, L (2006)
4
Introducing the speed-flow relationship

Every traffic engineer is taught the speed-flow relationship, which defines speed behaviour of
traffic under varying flow conditions. Figure 3 illustrates it for two conditions, a do-nothing
scenario and a do-something scenario. For the do-nothing scenario the maximum speed
possible is S0, whereas the do-something scheme has a higher maximum speed, S1. Notice
that in this simple case, speed is constant and free-flowing up until the capacity of the road is
reached, beyond which the speed deteriorates rapidly. (This is a reasonable simplification, for
the purposes of economic evaluation, of the complex dynamics of traffic behaviour.)

This speed-flow relationship can be redrawn as a cost-volume relationship, as shown in
Figure 4, with the Figure 2 elastic demand ‘curve’ overlaid.

Kane, L (2006)
5
The reduction of benefits due to induced traffic

In most economic evaluations the assumption is made that the roads evaluated are operating
below capacity, that is, in the free flow and constant cost section of the curves described
above. Closer reflection, however, reveals this assumption to be wrong for large parts of the
urban road network. Indeed, in many cases road investment merely moves a network from
being severely congested to slightly less congested. This scenario can be explored as in
figures 5 to 7.

Figure 5 shows the fixed-matrix assumption, as line FD, superimposed on the congested part
of the cost curve. In this case the benefits are calculated using the shape bound by ABDEF.

Kane, L (2006)
6

FED gives the estimate of benefits due to induced traffic, using the rule of half described
above.

Calculation of the actual benefits requires a move away from the fixed-matrix assumption
towards a more rigorous approach to the modelling of transport. The calculation of costs
which arise when elastic demand is integrated into the modelling process gives the value of G,
the point at which the demand curve intersects with the congested cost curve in Figure 6. In
reality, then, the economic benefits are bounded by the shape AFGH.

Kane, L (2006)
7

Kane, L (2006)
8

Figure 7 combines Figures 5 and 6, and illustrates the overestimate of benefits which is taking
place whenever the fixed-matrix assumption is applied to a congested urban area, and then the
results from this assumption are taken through to the economic evaluation stage.

Empirical observations of this phenomenon

For many years the standard forecasting approach used for modelling in the UK made use of a
fixed-matrix approach. Furthermore, since elastic demand was not acknowledged as being
significant, the benefits accruing to induced traffic (the so-called rule-of-half outlined above)
were also not included. It was only with the publication of the SACTRA (Standing Advisory
Committee on Trunk Road Assessment) report (1994) that these errors were identified as
significant. The next section gives a brief review of the empirical evidence of these economic
evaluation errors.

In the work of Williams and Moore (1990), a network was tested under various assumptions
of congestion and demand elasticity. A ‘Delta’ value was calculated:

Area HIDB – area FGI
Area AFDB

The tests found that in most cases actual benefits were within 10% of the fixed-matrix
approach estimates. However, in situations of medium to high congestion (that is, volume to
capacity ratios in the base case of 0.75 and above), or in situations where elasticity was
greater than -1.5, then the fixed-matrix modelling estimate led to a significant overestimation
of benefits. Equivalent South African tests would reveal even wider discrepancies between a
fixed-matrix estimate and the calculation of actual benefits. This is because in South Africa

Kane, L (2006)
9

economic evaluation typically includes the benefits to induced traffic, even under fixed-matrix
assumptions, whereas in the UK the approach was more conservative and this was not the
case.

In later work from SACTRA, in a full-scale model of the city of Cardiff, Williams found that
with medium elasticity and congestion assumptions the reduction in the fixed-matrix benefit
assumptions was 20-30%, although this could be even greater under heavy congestion and/or
high elasticity

Elsewhere, Litman (2005: 18) has examined the implications of including the traffic induced
by user cost reductions into the modelling and subsequent economic analysis of schemes, and
found that in all cases the impact on the B/C ratio was significant, and in one case the
examined scheme changed from having a positive effect to having a negative one.

Conclusions

The outcome of the SACTRA work was a significant revision to recommendations for
modelling practice in the UK, which now states that the “principal requirements for [road] modelling” are:

 A road traffic assignment model to provide estimates of traffic flow, speed and cost
changes;
 A means of estimating the volume of traffic induced by the expansion of road capacity
or by the road capacity released by travellers transferring to public transport” (Transport
Analysis Guidelines 2004)

Kane, L (2006)
10

The South African advice has been less categorical, and although it has been stated that:
“because of the ‘dynamic’ nature of the increase in user cost with increasing traffic volumes,
it is essential to take special care when determining the traffic volume for the calculation of
annual user cost if each year’s user cost is to be reflected accurately” (Hromic 1995: 2.22),
there are no guidelines on what this should mean for modelling practice.

More recent guidance (Pienaar & Botes 2002: 3-13) describes induced traffic as one of three
types of future traffic growth (alongside normal and diverted), but only mentions the benefits
which would accrue to induced traffic and not the very significant costs which induced traffic
growth could impose on other traffic, as described above. Indeed these authors state
elsewhere (2002: 3-5) that “it is well known that road user benefits increase exponentially
with an increase in traffic volume”, a statement which is at odds with the research
documented in SACTRA for congested conditions.

As South Africa continues to grow and the urban areas continue to experience growing
congestion, traffic induced by reductions in user costs will become more significant. The
benefits from investment in road infrastructure in urban areas will become quickly eroded due
to induced traffic (Behrens & Kane 2004). Without modelling which adequately takes
induced traffic into account, the cost-benefit analysis results will appear far more positive
than is realistic and could lead to an unfortunate misallocation of funds away from other
priorities.

Kane, L (2006)
11

References

Behrens R & Kane L 2004. Road capacity change and its impact on traffic in congested
networks: evidence and implications. Development Southern Africa, 21(4), October: 587-602.
Hromic, M. (Ed.) 1995. Guidelines for conducting the economic evaluation of urban
transport projects. 2nd ed. Cape Town: Core City of the Cape Town Metropolitan Transport
Area (CTMTA).
Litman, T 2005. Generated traffic and induced travel: Implications for transport planning.
Accessed online at http://www.vtpi.org/gentraf.pdf on 19 September 2006.
Pienaar, W J & Botes, F J 2002. Guidelines for conducting the economic evaluation of urban
transport projects. 3rd ed. Cape Town.
Standing Advisory Committee on Trunk Road Assessment (SACTRA) 1994. Trunk roads and
the generation of traffic. London: HMSO.
Transport Analysis Guidelines 2004. Major scheme appraisal: cost benefit analysis. Unit
3.9.2. Accessed online at
http://www.webtag.org.uk/webdocuments/3_Expert/9_Major_Scheme_Appraisal_in_LTPs/3.
9.2.htm on 19 September 2006.
Williams, H C W L & Moore, L A R 1990. The appraisal of highway investments under fixed
and variable demand. Journal of Transport Economics and Policy, 24: 61-81.

Acknowledgements

Thanks to Professor Anthony Leiman of the University of Cape Town for comments on an
earlier draft.

Back To Top