Where's the evidence that time savings from new roads deliver tangible economic benefits?

Professor Alan Wenban-Smith (p18, 'Viewpoint', Local Transport Today, 19 March 2010)

In October last year we reproduced, (with permission), a letter which appeared in 'Local Transport Today' (LTT) by Professor Alan Wenban-Smith in which he cited the Mersey Gateway as a prime example of how the present transport appraisal process fails to function as it is intended to (see archive). On the same theme, which he expands upon, he again refers to it in an article which appeared last month (LTT 541) which we reproduce below.

Professor Wenban-Smith, an independent consultant who specialises in integrated spatial and transport planning, was an expert witness for the main opposition case to the Mersey Gateway which was mounted for last year's public inquiry by the North West Transport Roundtable and Friends of the Earth.

The Eddington Report in 2006 recommended a new focus on the economic purpose of transport investment, in a move away from the traditional focus of reducing congestion. Nevertheless, in current appraisal practice user time savings continue to dominate the estimation of economic benefit, and to inflate the Cost-Benefit Ratio that is the key determinant of investment priority.

The 1999 SACTRA (Standing Advisory Committee on Trunk Road Appraisal) report Transport and the Economy makes an important observation on this issue: "... in general, the value of direct transport benefits must decline if indirect economic benefits are to grow". Economic theory predicts that direct transport benefits at the outset, even if they are eroded by changing travel and locational choices, should (in a perfect market) end up as economic benefits of equal value. However, the economic impact does not lie in the time savings themselves, but in the chains of causation that they set going. SACTRA listed these as: (a) rationalisation of production / distribution/land-use; (b) improved labour market catchments/costs; (c) lower cost of production; (d) stimulating inward investment; (e) unlocking sites; and (f) multiplier effects on all of these.

Provided they have been correctly valued, the eventual economic benefits will be equal in value to the original time savings, however long and tortuous the processes in between. Recent work on imperfect market effects (such as agglomeration) suggest that these are more likely to add to the benefit than reduce it.

In conventional appraisal practice, endorsed again by recent WebTAG revisions, time savings are derived from comparison of model results in 'do nothing' and 'do something' scenarios. Results are dominated by work time, which is valued at average employer costs (worth roughly four times as much as leisure time). Thus valued, time savings account for the overwhelming bulk of the quantified and monetised impacts of transport, and their contribution to the critical Cost-Benefit Ratio (CBR) means they dominate transport investment priorities.

So what's not to like? The reason for my concern is that in any particular case it is very difficult to see how an economic benefit of the size indicated by time savings will in fact arise. This difficulty is most clearly exposed where a Wider Economic Impact report has been undertaken, to identify benefits to Regeneration Areas (RAs). Such benefits can arise from relocation of activity to RAs from elsewhere, so they will generally be greater than the net benefit to the UK economy (and for this reason are kept separate from the main appraisal).

For schemes where regeneration is the major justification, it is generally found that the RA benefits are much smaller than the projected value of user time savings. At the Mersey Gateway bridge inquiry last July the net present value of time savings to business users was #222m, but the gross economic benefit to the RA was only 1,233 jobs (a net present value of about #15m at NAO rates). Similarly, in figures given at the Bexhill-Hastings link road inquiry in November, business user benefits were valued at #107m while gross RA job gains would be worth about #22m (and only about 20% of these were likely to be net gains).

These very large discrepancies demand an explanation. In both cases the main transport effects modelled were felt within the RA, yet diligent search could not show how economic benefits anywhere near as large as the initial time savings would appear.

We are driven to the conclusion that time savings are being grossly over-valued. One possible cause is not far to seek: conventional practice is to value even the shortest time savings at the same rate as longer ones. This requires a logical absurdity: gains below the threshold of perception should nevertheless lead to changes in economic behaviour of the kinds listed by SACTRA. Nevertheless this position continues to attract official support on the grounds that even small savings might sometimes be crucial, and that excluding them led to practical difficulties. Interestingly, an investigation done by the Institute for Transport Studies at Leeds in 2003 reported that no behavioural effect could be observed for non-work time savings less than five minutes. A high proportion of time savings found are of this order - over 70% in the Bexhill-Hastings case. The behavioural effect of work time savings was not examined, and depends on economic theory rather than observation.

Whatever the conclusion on this point, the fundamental objection remains: if the very large user time savings benefits typically attributed to road schemes are exchanged over time for economic benefits, it should be possible to identify where these will show up. This matters: the effect of user time savings has been to skew UK transport investment towards roads for almost 50 years, with devastating effects on our urban economies compared with continental counterparts.

So, what's the alternative? My view is that the pursuit of greater precision in appraisal through refinement of models and appraisal guidance is a chimera. After several decades of continuous refinement, diminishing returns have set in. The problem does not lie with appraisal as such, but with the over-dependence on modelling and appraisal as a substitute for a more balanced approach to strategic planning. The 2004 MVA Consultancy report to DfT, The integration of Regional Transport Strategies with spatial planning policies (co-authored by Denvil Coombe and myself) discusses this issue in more depth. With an appropriate strategic planning process to plug into, appraisal could do its proper job: choosing between alternative ways of achieving specific, limited aims. The way that the appraisal tail wags the planning dog makes transport an odd one out in terms of its planning practices, even within the UK.