Scale economies, cost complementarities and technical progress in Indian banking: evidence from fourier flexible functional form1

¹ Opinions expressed in this article are the sole responsibility of the authors and do not necessarily reflect the views of the institutions with which they are affiliated.

Abstract

This article uses a multi-product Fourier flexible cost function specification to investigate scale economies, cost complementarities and technical progress of Indian banks during the post reform period 1992 to 2003. The empirical results indicate that there exist significant economies of scale for all size classes of banks and there is no evidence of diseconomies of scale, even for larger banks. In particular, for small and medium-size banks, there is enough opportunity to increase output by either increasing the scale or merging with other banks to improve the average cost curve. The evidence is fairly robust even after controlling the impact of asset quality and overall risk exposure in the cost function specification. In addition, the statistical test confirms that the industry cost function does not have a Translog form. The results do not find any empirical evidence of cost complementarities between outputs. In terms of technical progress, Indian banks have experienced significant cost reduction which is as high as 5% for the recent period. However, the effects due to the scale augmenting and nonneutral components are found to be negligible.

¹ Opinions expressed in this article are the sole responsibility of the authors and do not necessarily reflect the views of the institutions with which they are affiliated.

Notes

¹ Opinions expressed in this article are the sole responsibility of the authors and do not necessarily reflect the views of the institutions with which they are affiliated.

² The banking system in India consists of commercial and co-operative banks, of which the former account for around 98%of banking system assets. The public sector banks account for about 80% of commercial banking in India.

³ In terms of size and scale, the big Indian banks are relatively small. The combined assets of the five largest Indian banks: the State Bank of India, ICICI Bank, Punjab National Bank, Canara Bank and Bank of India, on 31 March 2003 were less than the assets of the largest Chinese bank: China Construction Bank, which is roughly 7.4 times the size of the State Bank of India.

⁴ The present work extends the work of Jagtiani et al. (1995) in several ways. First, the analysis is more dynamic in nature in the sense that it considers the cost complementarities and scale economies over several time periods, rather than a single time point. Second, it uses FF cost function as against conventional Translog function. Third, the present study employs the Seemingly Unrelated Regression (SUR) method. Such differential estimation approach is dictated by the consideration that the errors in the equation of the FF/Translog function are likely to be contemporaneously correlated.

⁵ One may argue that the ultimate objectives of Indian banks vary across various ownership patterns like state-owned banks, private and foreign banks and a common cost frontier may not be suitable to capture the extent of scale economies existing in the banking sector. However, as a part of the current ongoing reform process, a market driven level playing field is being created for all banks and profit maximization has been a uniform unwritten objective of all banks.

⁶ For a detailed description of these approaches, see Berger and Humphrey (1992).

⁷ Of course there is an exception when two private sector banks viz., ICICI Bank and Bank of Madura were merged in 2001. Now, ICICI Bank is the second largest bank in India.

⁸ Interaction terms of these items with outputs and input prices are not included in the model.

⁹ The estimated coefficients of NPL and RWA were −0.009 and −0.034 with SEs 0.008 and 0.029, respectively. When both were included in the model, the coefficients were −0.003 and −0.035 with SEs 0.006 and 0.028, respectively.