The recent 2G spectrum scam has raised an important question: How can the allocation of scarce public resources to private enterprises be institutionalized? In this opinion piece, three professors — Ravi Bapna from University of Minnesota’s Carlson School of Management, who is also head of the Srini Raju Centre for IT and Networked Economy at the Indian School of Business in Hyderabad; Alok Gupta, who is also a professor at the Carlson School of Management; and Arun Sundararajan from New York University’s Stern School of Business — maintain that the overarching objective of any institutionalization effort should be to maximize social welfare and that a nuanced analysis of the business environment along with a thoughtful, scientific methodology are critical for the success of such resource allocation.

For a nation that has no tradition of auctioning scarce public resources to private-sector organizations, the recent outcry over the 2G spectrum scandal is both intriguing and encouraging. It’s a sign of hope when the citizens of India’s vibrant democracy are shown a better way of governance, they accept nothing less, and will exercise their collective public voice if they come across egregious infractions. When duly amplified by a boisterous media, this voice seems to have the power to topple a raja or rani [king or queen] or two. (Although we may have to wait until the next general election to ascertain how persistent the power of the people is.)

If the slivers of airwaves providing 3G mobile services — to a young and mobile population in an economy that’s growing at close to double-digit rates — can ring up more than US$22 billion for the country’s exchequer, one wonders about the fair economic value for all other air, water and land resources that belonged to you-the-nation, but was allocated to you-name-it company for a you-name-it purpose using — to put it kindly — less than transparent and economically efficient mechanisms. FM radio spectrum, deep-sea drilling, mineral rights, timber, highway and byway construction, and gambling licenses are just a few of a number of examples that come to mind.

The issue of corruption aside, it is fascinating, nonetheless, to focus on what transparent mechanisms could uncover about the true underlying economic value of India’s untapped resources and to think about developing a normative view of how we-the-nation might institutionalize the allocation of scarce public resources to private-sector enterprises. We hope to initiate this dialog by presenting a set of objectives, choices and tradeoffs.

Maximizing Social Welfare

The overarching objective for decision-makers in this context should be to maximize social welfare. However, when one closes the economics textbooks and maximizes social welfare in practice, it’s clear that the process can be quite nuanced. In the case of mobile spectrum, the resource allocation mechanism needs to be efficient in two ways. First, it should be able to identify the operators that are the most likely to roll out mobile services successfully to as a much of the population as possible in the shortest possible time, with great enough profitability to ensure the service’s sustainability — after all, there’s little value in bankrupting operators while maximizing spectrum prices. The mechanism will also ideally need to facilitate a “price discovery” process of supply and demand, which provides the exchequer with revenues commensurate with the ex-ante unknown value of the national resource, in order to give the population (especially the portion not using mobile technology) fair returns on the resource it collectively owns. Note how a mechanism that maximizes revenues could achieve neither of these goals.

On both fronts, the controversial 2G allocation process fell miserably short. If mobile spectrum licenses on a first-come, first-served basis are allocated to, say, construction companies that have no prior expertise in rolling out such services and they then hoard spectrum, the nation fails to roll out its critical information infrastructure to large sections of the population.

This is a huge social loss, which is reflected in the cumulative effect of the small, but numerous morsels of friction that persist despite a fast-growing, technologically connected economy. For example, a farmer might not be able to get a critical weather forecast before fertilizing a field, which might subsequently be washed away by unseasonal rain, or taxi drivers might not be able to pick up a few extra fares if their dispatchers are operating inefficiently. Add these up over the hundreds of millions of digitally disenfranchised Indians, and the welfare gains that might never come to fruition are massive.

Looking Forward

Can one way of looking at the damage caused by the 2G scandal be through the lens of where the opportunities now lie? We need to be forward-looking while keeping the welfare objective in mind, rather than focusing on the failure through the rear-view mirror. Any attempt to retract licenses will be challenged legally and lead to a protracted, inefficient stalemate, which will slow down the rollout of critical mobile infrastructure even more. Perhaps one step forward to enforce the welfare-enhancing rollout could be by mandating a use-it-or-lose-it rule — that is, insist that all allocated spectrum is used and monitor usage based on a simple, observable measure. Meanwhile, a secondary market could be facilitated for reallocation, giving operators with capital access to spectrum, and perhaps recovering some of the money the government left on the table.

Contrast the 2G “beauty contest” and the first-come, first-served approach (which portends to be a US$40 billion political scandal), with the more recent and transparent 3G auction. While the mobile operators did fuss about prices being too high during the 3G auction, the overall approach was positively received because of its transparency. No backroom dealing was suggested. The potential for efficient mechanisms in India seems immense, even if they seem complicated from afar.

How does this inform us more broadly about allocating public resources to private-sector firms? Are auctions of the type conducted for the 3G spectrum the way forward? Our answer is yes, albeit a qualified yes. Preserve the same broad objectives: Efficiently allocate licenses to firms that generally have the best business plans to use the resource; protect consumers from a monopoly by ensuring sufficient competition; and help the exchequer recover something resembling a fair value for the resource.

Auctions, while clearly not perfect, seem to have properties that are the most favorably aligned with these objectives. The rub is that the government has no simple way to ascertain ex-ante which firm has a superior business plan and is always going to be out-maneuvered by private firms in beauty contests, especially for resources used in new markets. But backroom wheeling and dealing — hard to avoid in private allocations — can be minimized by transparent price discovery based on clearly stated auction rules.

Need for Transparency

But the design of an auction is complex. Selling interconnected and complex billion-dollar public resources in auctions is a far cry from the traditional neighborhood or eBay auctions. For example, none of the winners of the 3G auction in May received a pan-Indian license comprising all 23 circles (or regions), and the average number won was only 10. The so-called simultaneous ascending auction (SAA) that was used evidently seems to have resulted in operators’ inability to have bundles of regions allocated to them. Vodafone won Chennai and the rest of Tamil Nadu but failed to win Karnataka and Andhra Pradesh.

The mechanism did not allow operators to bid for these three circles as a package. While the SAA is attractive in some ways, especially in simplifying early-stage price discovery, it could have been more efficient to close the auction with a final round, in which package bids were allowed and Vickrey-style (sealed-bid) discounted prices computed (where the seller pays the second-best price) — the latter mitigates the winner’s curse by ensuring that bidders pay only as much for a bundle as the competition forces them to pay for that bundle.

Making a definitive connection between the allocation inefficiency and the auction process needs a full analysis of the bid history (and we urge the Department of Telecommunication to release this). But it serves to remind us that while desirable, such auctions are also by no means a panacea for achieving efficiency.

Transparency and partnering with universities can help. The spectrum auctions in the U.S. have shown that predatory (retaliatory or blocking) bidding practices by large players, such as AT&T, suppressed bids from smaller bidders and competitors in earlier auctions. The discovery of such behavior and its consequences were possible only because the data was made public and the Federal Communications Commission worked closely with academic researchers to design features that mitigated these behaviors in the subsequent auctions for Advanced Wireless Services (AWS) in 2006 and the 700 MHz band in 2008.

Generally, every natural resource that is considered in the future for private allocation will have its own set of constraints and context. Designing these into a real-time auction mechanism, while keeping the broader goal of both exchequer revenue and overall welfare in mind, requires a nuanced analysis of a business environment as well as a thoughtful, scientific methodology with extensive laboratory testing and research, one that should precede how the final allocation mechanism is designed and run.

While taking us in the right direction, auctions need to be constructed with care. It might be useful here to take a leaf out of Ludwig Mies van der Rohe’s book. Although he was the architect of museums rather than resource allocation mechanisms, he did have a point when he cautioned us to pay careful attention to the fine points, especially when creating something we hope will be admired for its clarity and simplicity. Or to paraphrase his oft-paraphrased quote, the devil — when tackling inefficiency — really is in the details.