1.2 Empirical Studies on IPO Puzzles
1.2.2 IPO share allocation and oversubscription
As the offering price is the same to all investors, and the demand generally exceeds the supply in the primary market, it is entirely the underwriters' call as to how to allocate IPO shares to investors. Koh and Walter (1989) perform the rst careful examination on the allocation issues. They test the allocation of IPO shares related to the underpricing level during the period of the 1970s and 1980s. Their research is conducted with a sample of data from the Singapore IPO market, a special market in the sense that all investors apply for the same number of shares and all have an equal chance. In addition, the allocation policies are publicly available. Nevertheless, in other markets, such as that of the U.S., the allocation policy of underwriters is strictly restricted as regards the public.
From their investigation, the authors nd that the uninformed investors' returns are not signi cantly different from the risk-free rate of return in the market. Their observation is consistent with the prediction that underpricing as a source of excess return eventually produces an equilibrium offer price, which can attract uninformed
investors to the issue. Moreover, their analysis suggests that the offer price in primary markets is set above the uninformed investors' estimated price.
Saunders and Lim (1990) con rm the allocation problem with updated data from the Singapore market. Lee, Taylor and Walter (1996) suggest that according to the requirement of the Singapore Stock Exchange, smaller investors are systemat- ically favourably allocated for IPO shares. In addition, compared with the demand of small investors, the demand of the large investor is signi cantly more sensitive to the realised level of underpricing. Lee et al. also document that the oversubscription level is not a function of the underpricing level. In the sample used of 128 IPOs in Singapore, the median level of over subscription is 14.1 times.
Hanley and Wilhelm (1995) present direct evidence on IPO shares allocation, which supports the argument that institutional investors dominate the short-run pro t in offerings. Using con dential IPO share allocation data from 38 IPOs underwrit- ten by a single underwriter during the period from 1983 to 1988, the authors nd that approximately 70% of the shares are allocated to institutional investors. Across IPOs with different levels of realised underpricing, i.e. "hot" (high underpricing level) or "cold" (low underpricing level or no signi cant mispricing), the institutional in- vestors' share of the allocation is always at the same level.
On the contrary, the Wall Street syndicate manager, quoted byBusiness Week6,
estimates that the institutional investors take 80% of shares in highly underpriced
IPOs and 60% of shares in normally underpriced IPOs. In the same article, John Markese (president of American Association of Individual Investors) says, "After the institutional graze the market, there isn't a whole lot left. We've basically suggested that if you can get [some shares in an offering], you probably don't want it."
Lee et al. (1999) con rm that institutional investors pro t a lot from their favourable allocations of IPO shares. This research is based on the sample of 132 IPOs underwritten between July 1973 and December 1992. The result shows that the demand by large investors is positively associated with the realised underpric- ing. They repeat the test applied by Hanley and Wilhelm (1995) and nd that the proportion of the allocation to institutional investors and the underpricing level are positively related, though not signi cantly so.
Ljungqvist and Wilhelm (2002) report evidence found on allocation of IPO shares from the revision activities from offering prices. With the revision informa- tion of the IPO offering price, they nd that institutional investors are rewarded with favourable allocations of the underpriced shares for providing their private informa- tion on the IPO shares to underwriters. Moreover, they nd that the proportions of favourable allocations have a nonlinear relationship with both the revisions and the level of underpricing. Given that all other variables are the same, there is weak evi- dence that the underpricing diminishes when institutional investors receive more than 10% of the average allocation.
It is understandable that research into the allocation of IPO shares is performed in a limited number of markets, because the allocation information of IPO shares is con dential in most global nancial markets. After all, empirical studies mainly focus on two points. First, is there any signi cant relationship between the degree of underpricing and the distribution of shares among different groups of investors, for example institutional investors and individual investors? If such a relationship exists, to what extent, do the underwriters favourably allocate the shares to certain investors and why is this the case? Second, in different IPOs, such as highly underpriced ones, can the favourably-allocated investors gain information about the underpricing level in advance, and adjust their demand accordingly for the shares?
Although a lot of work has already been done, further investigation remains of importance to IPO research. Such investigation will no doubt, however, be very dif- cult because of the con dentiality of IPO allocation data. The uncovered sample of IPOs up to now is arguably biased given that the samples of IPOs were not randomly selected, because in the existing literature IPOs were usually conducted by one, or a few, underwriters.