The Impact of Charter Values on Bank Capital in Asia: A Threshold Regression Analysis

ABSTRACT

Recent theories suggest the relation between banks’ charter values and capital buffers is nonlinear and a function of the value of the charter. Using novel threshold estimation techniques, we investigate this for a cross section of 239 commercial banks in 24 Asian economies during the 2008 crisis and post crisis periods. For the latter period, a negative relation seems to dominate for high charter value banks in Asia (excluding Middle East region). During crises, our results for advanced Asian economies are consistent with capital buffer theory. Our findings raise distinct policy implications for the regulation and supervision of Asian banks.

KEYWORDS:

• bank regulation
• bank risk
• capital buffers
• charter values
• nonlinear relation

Supplemental Data

Supplemental data for this article can be accessed here.

Notes

^1. Bank capital buffers and risk taking are inversely related. Banks maintain buffers to cushion potential negative shocks and reduce costs associated with any regulatory violations (Elizalde and Repullo 2007; Milne and Whalley 2001).

^2. Niu (2012) and Jimenez, Lopez, and Saurina (2013) examine whether the relation between bank competition and risk taking is nonlinear for banks in the United States, and Spain, respectively.

^3. Examining the effects of charter values on bank capital in Asian developing economies is also very relevant as the publicly listed banks in our sample represent on average 65% of the total banking assets in these economies.

^4. The threshold is a function of a number of variables including risk and charter value as well as recapitalization costs and the relative costs of equity versus debt (Milne and Whalley 2001).

^5. Stolz (2007) mentions that the cost of recapitalization is difficult to observe, and that testing the effects of bank charter values on bank capital buffers are therefore conducted without controlling for this variable.

^6. The risk taking and solvency ratio coefficients obtained for 2008 however are less significant and relevant.

^7. Tobin’s q ratio is an ideal measure of market power given that the capitalized value of monopoly rents whether arising from dominance in the asset markets, or the deposit markets, or both, is reflected in the firm’s equity market value, and hence in their assets, but not in the costs of the acquired assets (Keeley 1990).

^8. As we are unable to observe the fixed cost of recapitalization, it is difficult to identify which of the banks in the sample have a charter value higher (or lower) than their cost of recapitalization (see Stolz 2007).

^9. As indicated in Table S1 available online, the $\mathrm{T}\mathrm{O}\mathrm{T}\mathrm{C}\mathrm{A}{\mathrm{P}}_i$ measure exhibits a weak association with $\mathrm{R}\mathrm{B}\mathrm{U}{\mathrm{F}}_i$.

^10. Of a total of 62 Islamic banks globally only 27 banks are publicly listed of which 15 figure in this study.

^11. We use this approach even though our earlier Hausman test results reveal $\mathrm{C}{\mathrm{V}}_{\mathrm{T}\mathrm{B}\mathrm{Q}}$ is exogenous.