After spectrum, why not auction roads?

We need to radically overhaul the existing, abused system of selecting concessionaries to build roads and the best way forward is to adopt bidding for present value of revenue.

India, advised prudently by Dot Econ, one of the world’s foremost auction theorists, conducted one of the most successful global auctions of spectrum which got the treasury $15 billion recently.

One of the questions that begs an answer is that if we can auction one kind of a limited national resource aka spectrum, why not another, namely roads?

If yes, how should such an auction be designed and what objectives should it seek to achieve?

India needs world-class highways and needs them quickly.

China has built 55,000 miles of highways in roughly half the time, about 11 years, that the US took to build its own similar network, whereas India can barely boast of about 12,000 km of world-class highways even 10 years after the start of the National Highways Development Programme (NHDP).

Private participation in highways is essential for three primary reasons.

First, to bring in capital, which the government is quite significantly short of.

Second, to bring in execution excellence by completing roads on time and within budget, which the government does not do very well. Third, to bring world-class quality in highway design and construction.
    
The puzzle is how to align the incentives of the developer to achieve these three objectives.

Developers in practice usually get into the bid at any cost, thereafter goldplate the costs and overinvoice to pull out about 15-20% upfront to recover their equity, then buy land around the proposed exits to be able to get windfall gains and finally renegotiate to get even further substantial gains, all at the expense of the taxpayer.

They also sometimes indulge in quality shading so as to be able to cut costs and increase the project IRR (internal rate of return).

In India, there is a permanent expectation on part of the concessionaire and the concensionee to renegotiate to give and take huge bribes at every step and thus the bids happen at unrealistic, artificially low tolls.
    
Thus, the taxpayer gets expensive, poor or uneven quality roads, and with fixedterm concessions where either the toll or the concession period is fixed, the gains from privatisation are easily squandered by opportunism and renegotiation.

Most private infrastructure concession contracts are renegotiated. Thus, the auction design has to be such that it prevents opportunism and renegotiation.
    
The often-adapted solution is one what is called a Demsetz auction where firms compete for the field rather than competing in the field through a process that mimics competition.

In other words, a government which wants to purchase a service or an asset such as a good quality road, they will bid and award it to the lowest bidder, which is the classic Demsetz auction or competing ‘for’ the field.

The other option is to award it to two agents and then allow them to compete ‘in’ the field’.

However, what this article proposes is a third method, which is in some ways a variant of this classical method.
    
The government is said to be contemplating a system based on a new index like the IRR and doing away with the annuity model altogether.

While the latter would be welcome, for the fiscal burden it imposes on the government is also a ‘cost’ to be added to that of the highways itself, the former is subject to debate.

It should frankly not be the government’s concern what IRR a developer gets. The concept of a cut-off IRR after which a project is considered unviable doesn’t take care of the key issues that hold up highways from coming up mentioned above.

Instead of making developers compete with one another in bringing down the IRR that gets actually selected, they should be made to bid on the basis of a present value of revenue (PVR) method based on the present value each bidder expects to earn from the project and this would minimise, if not eliminate, opportunistic renegotiation of the contract.

How it happens is as follows.
    
The authority would simply set a maximum toll and specify quality standards and monitoring specifications for toll equipment, not done at all today.

For instance, there are electronic logs available that show road availability and actual number of vehicles plying at any point in time in almost every new toll road in the world.

Thus, if a road is down for maintenance or if there’s accident on a road, the cameras time stamp it and stream it to logs which capture the downtime based on which penalty is imposed on the operator.
    
In response to the pre-announced toll schedule information instead of bidding on a toll as is usually the case, bidders will announce the present value of tolls for each package.

The concession would be won by the firm bidding for the least present value of toll revenue (and not the least IRR).

What sets apart the PVR method is that the concession will come to an end when the present value of toll revenue is reached and more importantly, each firm’s bid reveals the revenue required to earn what can be called a ‘normal profit’ and thus reduces post-contract negotiations.

Under the PVR method, while the concessionaire will still bearing the risk, he will not have the incentive to renegotiate windfall gains at the expense of the taxpayer nor would he make losses, considering the long-run demand.
    
We would typically expect the least PVR chosen to be below the reserve IRR (can be 18% and 21% as per Chaturvedi Committee's report), just as in the 3G auctions, the price for bandwidth was sold was substantially about the reserve price fixed by the government.

What is also important is the discount rate used to calculate PVR, which should ideally be as close to the weighted average cost of capital and should be specified in the bid document itself.

Lastly, the concession period can be readjusted based on measured actual demand, allowing flexibility to the government to adapt to poor or good traffic.
    
The global experience in highway concessions reveals clearly that a fixed contract based on either a fixed toll or duration has only led to renegotiation where often the best-connected firm, and not the best firm, wins and opportunism is rampant at a huge cost to the taxpayer.

Evidence from the UK and China reveals that when developers compete for the field instead of competing in the field, then it mimics competition in a natural monopoly and aligns developers interest with that of the citizens and more importantly, reduces the onus on the government to be vigilant which it often is unable to perform diligently, either by accident or by design.