Determinants of the underpricing of new shares during the subscription period: empirical evidence from the Spanish stock exchange

Abstract

This article reports on an issue hitherto unexplored in the literature, namely, the ‘new shares’ price setting during the subscription period. We report evidence of a spread between the old stock price and the value of new shares obtained through subscription. A framework is developed within which to analyse the explanatory factors involved in this spread in the Spanish Stock Exchange. The empirical evidence suggests that new share prices during the subscription period are influenced by a range of factors, such as difference in the amount of tax to which subscription rights and capital gains are subject, characteristics of the issuer and the issue, norms established between clients and banks and the microstructure of the subscription rights market.

Acknowledgement

The authors would like to thank the Spanish Ministry of Science and Technology (MCYT SEC2003−07808-C03) for financial support. The authors are pleased to acknowledge the very helpful comments of one anonymous referee and the Editor Mark Taylor. We are responsible for any errors.

Notes

¹ See Armitage (1998) for detailed reviews of the evidence from the USA and the UK on seasoned equity offers (SEOs) and rights issues.

² In clause 158 of the Inc. Corporations Law, Spanish legislation allows current shareholders and convertible bond holders pre-emptive subscription rights on new share issues.

³ An earlier article by Riaño et al . (2004) on the Spanish stock market is largely based on a univariate analysis in which there are no variables relating to the trading activity of old shares or subscription rights.

⁴ We define the ‘new share’ price during the subscription period as the value of rights required to subscribe to a new share plus the subscription price.

⁵ Note that, in a free issue, any investor not owning old shares or subscription rights may subscribe new shares by purchasing the necessary subscription rights in the subscription rights market.

⁶ For shares that paid out dividends during the subscription period, adjusted returns were calculated for the days prior to dividend payment, by subtracting the amount payable in dividends from the average price of an old or ex-right share, given that holders of subscription rights are not entitled to dividends.

⁷ To be exact, in Spain, the amount received on the sale of subscription rights is not considered capital gains, but simply a reduction in the cost of the portfolio. In other words, subscription rights are not liable to taxation until the shares to which they are attached change hands. This is tantamount to tax deferral and, therefore, to a lower rate of taxation on returns from subscription rights, than on capital gains.

⁸ This underpricing may be also be one of the reasons for the lack of liquidity in new shares until they are admitted to quotation.

⁹ A similar approach can be also seen in Mcguinness (2001).

¹⁰ In Spain, as in other European countries, the level of discount is higher than in the USA. See, for example, Eckbo and Masulis (1992) on USA, Tsangarakis (1996b) on Greece, Bohren et al . (1997) on Norway, Martín (2000) on Spain, Gajewski and Ginglinger (2002) on France, Burton and Power (2003) on UK. Burton et al . (2004) suggest that firms are fully aware of the potential cost savings associated with deeply-discounted rights issues, but concerns about investor reaction to any offer-induced earnings dilution, continue to mitigate against any significant increase in their popularity.

¹¹ This system enables broker-dealers to conduct stock lending, and make loans directly related to purchases and sales. It allows investors to buy or sell in cash shares that comprise the Ibex35 for a value up to three times the cash investment, through the granting of a loan, in the case of purchases, or stock lending for sales.

¹² Over the observation period 1992 to 1998, the subscription period tends to last approximately 1 month (24 market sessions), in accordance with the 1 month minimum legal limit.

¹³ The SIBE (Spanish Stock Market Interlinking System) or continuous market is chosen in order to avoid problems with different trading systems. Another important reason for this is the greater liquidity of stock trading on this market, which provides more opportunity for arbitrage. The continuous market represents ∼98.5% of all stock market trading in Spain.

¹⁴ Law 37/1998, of 16 November, Ammendment to Law 24/1988 of Stock Market legislation.

¹⁵ Several studies have confirmed the influence of ticks on markets (among others Harris, 1994 on the USA market; Blanco, 1999 and Abad, 2003 on the Spanish market).

¹⁶ The events in question were: (1) two subscription periods that fell between the end of 1998 and the beginning of 1999, thereby coinciding with the changeover of the asset trading system to euros; (2) nine subscription rights that were theoretically negative, which meant that the frequency of quotations on subscription rights reduced sharply and at times ceased entirely; (3) two ex-right shares that were not listed during the subscription period because they coincided with an operation to reduce social capital to zero; (4) three 2-week subscription periods when quotations on subscription rights were infrequent; (5) three seasoned equity offerings that became the object of an subsequent public offering or takeover bid; (6) a preferred share issue; (7) a seasoned equity offering coinciding with an issue of convertible bonds; (8) two cases in which old shares with full economic rights and old shares with rights varying over the subscription period drew equal, thus increasing the overall market for old shares by including those that had been listed separately because of the aforementioned difference and (9) two outliers with extreme subscription returns.

¹⁷ In the free issues, as the subscription price is equal to 0, the issue size will be nil.

¹⁸ In order to obtain the mean upper bound it was necessary to estimate real transaction costs. Spanish stock market Regulations were consulted to check restrictions on transaction costs based mainly on Royal Decree 949/1989, of 28 July on commissions applicable in operations involving stock admitted for trading on the stock market. Brokerage commissions were 0.25%. It is true that these costs form only part of the transaction costs (see, among others, Mackinlay and Ramaswamy, 1988). Nevertheless, average subscription returns are usually too high to warrant this assertion.

¹⁹ Confirmed by Friedman's test, in which the chi-square statistic takes a value of 123.279, thus rejecting the null hypothesis that average subscription returns remain the same throughout the subscription period.

²⁰ The null hypothesis that average values remain the same throughout the subscription period is rejected for all the variables, except those relating to shares: relative depth at the ask, relative depth at the bid and total relative depth.

²¹ Other specifications also proposed for the model, setting four levels of nonfree issues (nonfree issues with a subscription discount of 25% or less, between 25 and 50%, between 50 and 75% and over 75%) instead of two, yielded inferior results. We also checked for multicollinearity.

²² The remaining binary variables relating to the period of admission to quotation (ranging from 31 to 60 days, 61 to 90 days and over 90 days) were not significant.

²³ Note that these results are consistent with those obtained in the preliminary univariate analyses of these variables, except for the sector, which was relevant in nonfree issues. The admission period variable, meanwhile, which the preliminary univariate analysis indicated as being of little relevance in nonfree issues, is now seen to be of more importance.

²⁴ Specifically, on average, throughout the subscription period, the relative volume of trade in subscription rights from free issues accounts for 0.776% of the total vs. 2.134% in nonfree issues. Figures data for the last three days, however, are 0.826% in free issues and 6.755% in nonfree issues.