While the risks in road construction have fallen steeply since the first BOT roads of the National Highways Authority of India (NHAI), they remain much higher than they need be. The risks of land acquisition are particularly large in populous areas and those with poor law and order. The process of handing over sections for construction only after acquisition of land by the state lowers the risk, but the obverse of such risk-avoidance rather than risk-reduction is that few of the roads planned in these “problem” areas would be built on time unless the approach to land acquisition changes radically.

In other areas, where land acquisition is not the problem and the private sector is waiting, the risks can substantially come down if the value of roads can be enhanced. These would also allow markets to be brou-ght in to realise the economies of specialisation without upsetting the incentive structures that a BOT/annuity-financed road entails.

Values are currently held down by several factors, such as poor design, that is seriously limiting. Either too many exits and entries or too few of them destroy value. Given the variety of vehicle types, the variance of traffic, more than the volumes, creates congestion. Therefore, service roads and lane separation that do not affect high speed on the main highway, rather than multi-laning, would enhance service levels on roads considerably.

Similarly, adequate numbers of over- and under-passes need to be anticipated upfront. Today, many completed sections of the Golden Quadrilateral are being dug up for construction of under-passes and may again be dug up for construction of service roads!

Tolling sections do recognise the needs of the consumer but are driven on administrative basis. Similarly, there is much value reduction by interspersing good roads with bad roads and bottlenecks. Thus, the value of a sequence of GGGBBB is many times that of GBGBGB; where G is a good section of 10 km and B a poor section. The latter being the case today, there is far too much value reduction. Poor traffic discipline, especially by truck and local ferries, takes away the value by introducing high risks. More than 80% of trucks do not have tail lights and no traffic inspector thinks this is wrong. Can’t toll points debar such trucks?

Surprises further add life-threatening risks to road users. Night driving is all but impossible for cars. Unbuilt service points, service points on the wrong side, dubious quality of petrol further take away value. So, to the user, the tolls seem like a major cost.

Most important, the near- complete absence of coordination between NHAI and city governments results in multi-lane highways ending in the periphery of cities, leaving the user wasting all the time gained on the highway. BOT, while having the right structure, forces traffic risks upon the operator. Similarly, the annuity model means the government bears the traffic risks. Neither option is the best.

Untolled sections result in little recourse for the government to impose service levels and quality of road surface upon the constructor. One important option is to use shadow tolling, not just based on the number of lanes but on actual measurement of traffic by independent agencies. Thus, even if a road is not tolled, it may be worthwhile for the government to pay the annuity amou-nts depending upon actual traffic carried.

Similarly, why force traffic risks upon the operator? After all, roads, like other networks (gas pipelines or power transmission), show significant network flow externalities—like the traffic on a road section could decline because some other section has improved or demand has changed—even increased. So section- wise traffic risks can be lowered through pooling. This means that markets, where such risks can be traded on the lines suggested by Tilotia and Pawar in a path-breaking paper (in The India Infrastructure Report 2004: Ensuring Value for Money, OUP) can become important.

If the government can measure the traffic on most national highway sections, tolled or otherwise, and these can be linked to revenue streams, then speculators with better information and analysis on traffic flows (and there are large-scale econ-omies here) can emerge allowing road operators to hedge their traffic risks in an exchange market. This means traffic risk can be unbundled from O&M and construction risks without destroying the incentive compatibility of models like BOT. Thus, a BOT operator would offload his traffic risk on the market, while someone specialising in understanding traffic could take on these risks.

Given that India—except for interior Chhattisgarh, the Himalayas and the deserts— has very high population density, the network aspect of roads is bound to be large making it all the more imperative that either the government or development finance institutions take the lead in promoting markets to hedge traffic risk.