Initial Public Offerings
Incentives to Underprice

When a company is launched on the stock market, it may be expected that the original owners would want to receive as high a price as possible for their initial public offering (IPO). In fact, according to Julian Franks, there are strong incentives for `insiders', the company's directors, to underprice the issue as a means of retaining control. Franks, who is Professor of Finance and Director of the Institute of Finance and Accounting at London Business School and a Research Fellow in the CEPR Financial Economics programme, was speaking at a London lunchtime meeting on 5 December. His talk was based on his CEPR Discussion Paper No. 1211, `Underpricing, Ownership and Control in Initial Public Offerings of Equity Securities in the UK', written with Michael Brennan. The meeting was sponsored by Merrill Lynch under CEPR's Corporate Membership programme.
An IPO is said to be underpriced if the price at which the security trades in initial dealings exceeds the offer price at which it was sold to investors. IPO underpricing has been confirmed by researchers in many different countries, and there are many theories to explain why the owners of a company may want to sell shares to outsiders for less than the apparent maximum price achievable. Franks focused on how underpricing can be used by insiders to retain control, and how the separation of ownership and control evolves as a result of an IPO. He argued that pre-IPO shareholders, who derive private benefits of control, will have incentives to underprice so as to ensure oversubscription and rationing in the share allocation process. Rationing allows discrimination between applicants for shares, and limits the block size of new shareholdings. This, in turn, reduces the possibility of management being subject either to close scrutiny by a larger shareholder or to a hostile takeover.
This `control' theory suggests several propositions. First, underpricing and the consequent oversubscription is used to discriminate in the rationing process against large applicants and in favour of the smaller investor. Diffuse shareholdings intensify the free-rider problems that impede a hostile change in control. Second, using the rationing process to create a more diffuse post-IPO shareholding makes it more costly to assemble large blocks of shares. Large shareholders could assemble large blocks in the secondary market, but such purchases may not be profitable if the change in ownership and control is expected and the price rises to anticipate the gains. So if rationing is used to create diffuse shareholdings, it should be expected that smaller blocks emerge subsequent to IPOs with greater underpricing. Third, if directors obtain private benefits from control and also set the issue price, it should be expected that the lower the fraction of underpricing costs borne by directors, the greater the underpricing.
Testing these hypotheses is assisted by the formality of the UK new issue process in which potential purchasers must generally submit quantity demands at a fixed price specified by the issuer. In the event that the issue is oversubscribed, the available shares, and therefore the gains from underpricing, must be allocated by a formal rationing scheme. Franks used data from a sample of 69 UK IPOs over the period 1986–9; 64 of these were fixed price offerings. He found that holdings by directors were reduced by a third – from 42% of the number of shares outstanding prior to the IPO to 29% seven years later. In contrast, holdings by other (private) shareholders were virtually eliminated over the same period, declining from almost 42% of the pre- issue number of shares to less than 3%. Holdings by other (institutional) investors fell almost as dramatically. This suggests that non-directors see the IPO as a vehicle for disposing of their shares, and although their ownership at the post-IPO stage remains substantial, it is only temporary.
Private companies and large public corporations represent opposite extremes of the relationship between ownership and control. The IPO is a key step in the evolution of a management-owned firm into the public corporation and of the separation of ownership and control. Franks's results indicated that in less than seven years, almost two-thirds of the offering company's shares had been sold to outside shareholders, substantially advancing the separation process. There was substantial rationing in many of the new issues in his sample. The rationing frequently discriminated against the large investor and in favour of the small investor, and this discrimination was positively related to the degree of underpricing.
Franks next turned to the results of the control hypothesis. The first and second propositions suggest that the greater the underpricing, the easier it is for firms to introduce rationing in the share allocation process, and therefore to render it more difficult to assemble large blocks after the IPO. The control hypothesis predicts that the larger the underpricing, the more diffuse the shareholding structure and the smaller the size of post-IPO blocks. When the dependent variable in Franks' analysis was the largest outside holding or the total of large outside holdings, there was strong evidence that underpricing tends to prevent the formation of large blocks of shares in the hands of outside shareholders.
The costs of underpricing are borne by the pre-IPO shareholders, but may be borne differentially by different investor groups. The distinction is important because directors may derive greater private benefits of control, and may have more influence in setting the issue price than other shareholders. If the shares sold in the IPO all come from non-directors, directors may be more inclined to underprice. Hence, the third proposition suggests that the lower the fraction of underpricing costs borne by directors, the greater is the underpricing. An important assumption is that directors and not other insiders set the issue price and therefore determine the extent of underpricing. In Franks's analysis, underpricing was used to calculate the costs borne by directors, so he used the rate of oversubscription as the dependent variable since the higher the level of underpricing, the greater the level of oversubscription. The results showed that the fraction of costs borne by directors was not a significant variable in explaining underpricing, although the coefficient had the right sign.