The post-offering performance of IPOs in the Indian banking industry

Abstract

In the literature, the underperformance of IPOs is a well-documented empirical anomaly. This study concentrates on IPOs from the banking sector of an emerging economy, India. In a developing country, the role of the banking sector for economic development is undisputed. In view of its importance in economic resource allocation and its distinction from other industries in general, this paper analyses the post offering performance of banking sector IPOs in detail. The performance evaluation on the basis of stock returns did not find significant evidences of underperformance for the IPOs from the banking sector. Moreover, the study, based on key accounting parameters, found improvement in the performance of the banks in the post-listing period. There were no significant differences across ownership groups (public sector banks vis-à-vis their private counterpart) in the IPO performance.

Acknowledgements

The author is a Research Officer in the Monetary Policy Department, Reserve Bank of India. The views expressed in this article are the author's own. The author is grateful to Dr Subrata Sarkar for his helpful comments on the earlier drafts of this paper and Ms Soma Mittra for her support.

Notes

For some banks, which haven’t completed three years, the latest available month was included for the calculation of BHR.

The PROWESS database is compiled by Centre for Monitoring the Indian Economy (CMIE). This dataset is similar to the COMPUSTAT database in USA.

This is done by regressing BHR _i for a particular aftermarket period j over a dummy variable DPSB (taking value one if the IPO is from a public sector bank, otherwise zero). The regression equation is as follows: BHR _i  = α + β(DPSB) + υ. However, none of the slope coefficient of DPSB dummy (for j = 1, 3, 6, 12, 18, 24, 30 and 36) was significantly different from 0. These results are not reported due to lack of space.

These crucial ratios were reported in Report on Trend and Progress of Banking from the year 1996–1997 and data up to 2001–2002 were available during the time of the study.

Notice that the fixed effect model has an overall constant as well as a “group” effect for each group and a “time” effect for each period. The problem of multicollinearity – the time and group dummy variables both sum to one – is avoided by imposing the restriction Σ _i  a(i) = Σ _t λ(t) = 0. A full set of estimates is produced for the two-factor model in the same fashion as for the one factor model. The model is described in standard textbooks such as Judge et al. (1985) or Greene (1997)’ – LIMDEP, Version 7, User manual, pp. 338–9.