Jaimini Bhagwati: Why bother about IPO pricing?

We need to try out IPO auctions, on an experimental basis, as a way of improving price discovery.

Initial public offerings (IPOs) in India and most other countries are usually underpriced. A stock issue is deemed to be underpriced if the closing price on the first day of listing is higher than the IPO price. A review of the literature on the subject confirms that even though the bookbuilding methodology is an improvement over fixed price IPOs, issues continue to be significantly underpriced. For example, on average, underpricing in the US from 1980 to 2001 was 18.8%. In the dotcom boom years of 1999 and 2000, underpricing was much higher at about 72% and 56%, respectively.

In the context of the recent IPO "scam" in India, various solutions have been proposed to detect multiple applications made in the name of the same person to corner large allocations from shares reserved for individual retail investors. The suggested solutions include improvements in regulatory policing. However, we know from past experience that systematic mispricing engenders fraud, regardless of regulatory efficiency. It is likely, therefore, that scamsters would devise ever more sophisticated schemes to get around regulatory controls if an IPO is perceived to be significantly underpriced. It will be argued in this article that along with tighter regulation we need to reduce underpricing in Indian IPOs by allowing auctions, in addition to bookbuilding and fixed price issues.

Bookbuilding was introduced in India in 1999 and it has been steadily gaining greater acceptance vis-à-vis the traditional fixed price IPOs. Last year, well over 50% of Indian IPOs used the bookbuilding methodology while the remaining were fixed-price issues. In the bookbuilding process, underwriting banks, in consultation with institutional investors and the issuer, estimate a price band for the stock to be put on sale. Bookbuilding IPOs used to allow some allocation discretion to the lead managers but this has been reduced since September 2005. If there is oversubscription, allocations are now made to institutional and retail investors as per quotas on a pro-rata basis. Allocations in a fixed price IPO are made on a purely pro-rata basis.

In the US, almost all large IPOs use the bookbuilding process. If there is oversubscription, allocation is done by the underwriters at their discretion. In the 1990s, IPOs were mostly auctioned in Japan and France. More recently, both bookbuilding and auctions are being used in Japan and France.

In comparison with the underpricing in bookbuilding issues, it is likely that auction IPO prices would be closer to first day listing prices. For example, in the Open IPO auction system, pioneered by William Hambrecht in the US, investors specify the price and number of shares they are willing to purchase. The issuer allocates shares in a descending order of prices till the amount of shares to be issued is exhausted. The price quoted on the last bid becomes the price at which demand is deemed to meet supply and all bidders pay this price for their allotments. If the issue is oversubscribed, allocation is done on a pro-rata basis providing for equitable treatment across investor categories.

An example of a major auction IPO is the 2004 US$1.9 billion Google issue. It was a limited success in terms of efficiency in price discovery since the listing price was about 18% higher than the IPO price. Some have argued that for the IPO of a path-breaking IT company, the 18% price differential was relatively modest. Several forms of auctions have been proposed for IPOs, e.g. the Ausubel variants. However, due to space constraints, the different forms of auctions and probable effectiveness in terms of price discovery and fairness of allocation will not be discussed in this article.

India is not alone in instances of wilful misallocation of IPO shares. In the US, during the dotcom boom years of 1999-2000 several leading investment banks allocated IPO issues on a favoured basis to institutional and high net worth investors. This was done for return favours such as pre-agreed further purchases after the stock was listed and/or repeat business. The US capital market regulator, Securities and Exchange Commission (SEC), fined JP Morgan and others for violating Rule 101 of the SEC's Regulation M. It is evident, therefore, that it is impossible to even prevent international investment banks, with much to lose in terms of reputational risk, from engaging in wrongdoing if there is systematic underpricing.

Arguments against IPO auctions come most stridently from investment banks. Their argument is that since it is impossible to price issues accurately, underwriters end up taking unquantifiable risk. Additionally, to ensure sufficient investor interest it is appropriate to have a conservative bias towards underpricing. Another argument used is that if the IPO issuer is not worried about leaving money on the table, why should regulators worry since retail investors benefit from underpricing. We need to worry since systematic mispricing usually spawns fraudulent behaviour.

Auctions have been used for price discovery in various other sectors of the economy in which it is difficult to estimate value. For instance, auctions are used to allocate telecom spectrum allocation and compensation for reductions in carbon dioxide emissions. Internet auction platforms such as Ebay are widely used for sale of items at the retail level. It is likely that IPO auctions will reduce transaction costs for issuers.

It is clear that given the incentives of higher fees, and a greater probability of repeat business with institutional investors as compared to issuers, underwriters are likely to resist initiatives to experiment with IPO auctions. It would, therefore, be worthwhile to make a beginning with Indian issuers who are willing to try the auction route with smaller IPOs. It should be logical for less well-known issuers to use the auction route for their IPOs to get the best possible value for their stock. Further, if the listing price is close to the auction IPO price, reserving shares for retail investors would not be required.

To summarise, bookbuilding results in underpricing IPOs. Such a pricing bias leads to a scramble for allotment. Under these circumstances, no matter how vigilant or efficient the regulating agencies, it is difficult if not impossible to stamp out wrongdoing. This is particularly true in India as we do not have national identity numbers, and our justice system finds it difficult to secure convictions. Consequently, we need to try out IPO auctions, on an experimental basis, as a way of improving price discovery.

j.bhagwati@gmail.com