Banking regulation: does compliance pay in emerging economies? Evidence from Trinidad & Tobago/Regulacion bancaria: ?Vale la pena el cumplimiento de normas en economias emergentes? evidencias de Trinidad y Tobago/Le respect de la regulation bancaire paie-t-il dans les economies emergentes? Le cas de Trinidad et Tobago.
This paper seeks to establish whether compliance with the Basel Prudential Regulations and Requirements (BPRR) by commercial banks in the emerging economy of Trinidad & Tobago brings about any measurable improvement in their performance. A model is constructed in which bank performance is explained by macroeconomic variables and bank-specific variables, in particular a variable representing the BPRR compliance rating constructed from confidential information provided by the Central Bank of Trinidad & Tobago. There is sufficient evidence that BPRR compliance affects performance, though not always in a positive way. Other bank specific and macroeconomic variables also have some influence but they are not as convincing as the influence of BPRR compliance. This is contrary to other findings in the literature.
El presente trabajo trata de establecer si el cumplimiento de las Regulaciones y Requerimientos de Basel Prudential (BPRR) por parte de los bancos comerciales en la economia emergente de Trinidad y Tobago trae como resultado alguna mejoria medible en su desempeno. Se ha disenado un modelo para explicar el desempeno del banco por medio de variables macroeconomicas y variables especificas de los bancos, en particular una variable que representa el nivel de cumplimiento con las Regulaciones y Requerimientos de Basel Prudential (BPRR) creada a partir de informacion confidencial aportada por el Banco Central de Trinidad y Tobago. Existe evidencia suficiente de que el cumplimiento con las BPRR afecta el desempeno, pero no siempre en una manera positiva. Otras variables especificas y macroeconomicas de los bancos tienen tam-bien alguna influencia pero no resultan tan convincentes como la influencia del cumplimiento con las BPRR. Esto contradice otros hallazgos que aparecen en la literatura.
Cet article cherche a etablir dans quelle mesure la conformite aux exigences et a la regulation prudentielles de Bale (BPRR) par les banques commerciales de l'economie emergente de Trinidad et Tobago, ameliore leur performance. A cette fin, il a ete construit un mode1e au sein duquel la performance bancaire est expliquee par des variables macroeconomiques et des variables specifiques aux banques, en particulier une variable representant une cote de conformite au BPRR etablie a partir d'informations confidentielles fournies par la Banque Centrale de Trinidad et Tobago. Il ressort clairement de cette analyse que la conformite au BPRR affecte la performance, mais pas toujours de maniere positive. D'autres variables macroeconomiques et specifiques aux banques ont egalement une influence, mais elles ne sont pas aussi determinantes que l'impact de la conformite au BPRR. Ces resultats contredisent les autres etudes menees dans la litterature specialisee.
Regulation not a one-size-fits-all strategy
This paper seeks to establish empirically whether compliance with the Basel Prudential Regulations and Requirements (1) (BPRR) by commercial banks in the emerging economy of Trinidad & Tobago brings about any significant improvement in their performance. Emerging economies, especially small ones like Trinidad & Tobago, face challenges that are different from those faced by their counterparts in the developed economies and regulation may not be a one-size-fits-all strategy. Many have argued that the Basel Accord was developed mainly to seek out the interests of the developed world, not those of emerging economies, especially small ones (Grynberg and Silva 2006). Since the process of compliance is a costly, time consuming one, performance enhancement ought to be a major incentive for a bank to comply.
A model is constructed in which bank performance is explained by bank-specific variables, macroeconomic variables and, most importantly, a variable representing the BPRR compliance rating of each of the four major banks in Trinidad and Tobago. The ratings are constructed from confidential information provided by the Central Bank of Trinidad & Tobago, the very first time that such information has been used in any study. (2) Panel data methods are applied to four specifications of the model, each one involving a different indicator of banking performance as the endogenous variable: net interest margin, return on assets, return on equity and non-performing loan ratio.
This study distinguishes itself from other studies that examine the nature of the relationship between the level of compliance with the Basel Core Principles (BCP) and the performance of the banking or financial system as a whole. These include Sundararajan et al. (2001), who look at the relationship between BCP compliance and the soundness of the financial system as a whole (and not specifically at commercial banks); Montgomery (2005), who investigates the relationship between Basel capital adequacy requirements and portfolio choice of Japanese banks; Podpiera (2006), who studies the relationship between banking sector performance and the quality of regulation and supervision as measured by compliance with the BCP; and Demirguc-Kunt et al (2006), who investigate whether compliance with the BCP improves bank soundness. The main distinguishing factor in this paper is the use of pristine, highly confidential bank information from individual commercial banks to measure Basel compliance by each bank, which is then used to look at the effect of Basel compliance on the commercial banks. As far as the authors know, this has never been done before.
In the following section, there is a brief survey of the literature on the evidence of Basel compliance and its effectiveness. The data and methodology employed in the analysis of the Trinidad & Tobago case are then discussed and reviewed, after which follow the results of the estimation exercise. The paper then concludes.
Effectiveness of Basel compliance: evidence from the literature
There are conflicting views about the effectiveness of compliance with Basel principles, especially in emerging economies. Majnoni and Powell (2005) discuss whether Latin American and other emerging economies would want to implement Basel II, stick to Basel I or adopt a mixed approach. They conclude that bank capital should be regulated but how to achieve this is not obvious. They opine that regulators should not, because of peer pressure or pressure of large international banks, adopt the complex rules of Basel II too quickly since banks in emerging countries are less advanced in using internal rating methodologies.
The International Monetary Fund and World Bank (2002) show that many countries are far from being fully compliant with the Core Principles and emerging economies lag behind their G10 counterparts. They suggest that full compliance is too strict a condition for moving to Basel II because not even all G10 countries are fully compliant. Banks should just comply to the degree required to implement the appropriate alternative chosen within the Basel II framework. They believe that many developing countries would opt for the simpler Basel II approaches. However, their advice is that emerging economies should stay with Basel I, or only adopt Basel II on a subset of banks for several years beyond 2007.
Sundararajan et al. (2001) examine the nature of the relationship between the level of technical compliance with the BCP and its relation to the financial system stability and soundness. They construct a BCP non-compliance indicator based on the Core Principle Assessment (CPA) for 35 countries in 1999-2000. The index is constructed as an unweighted aggregation of the number of Core Principles for which a country is assessed as being compliant or non-compliant. Correlation coefficients are calculated for indicators representing BCP noncompliance, credit risk and bank soundness, macroeconomic stability and banking developments. The observed correlations suggest, among other things, that credit risk and banking soundness indicators have significant association with selected macroeconomic and banking indicators, but not with BCP compliance.
In yet another study, Montgomery (2005) investigates whether stricter capital adequacy requirements introduced under the Basel Accord cause Japanese banks to alter their portfolios from heavily weighted risk assets, such as loans and corporate bonds, to unweighted assets such as government bonds. Panel data from 72 Japanese banks' balance sheets and income statements for fiscal years 1982-1999 are used to estimate the empirical relationship between the growth of various classes of the banks' assets and the banks' capital ratio, both before and after the Basel Accord of 1988. The study looks at the effect of capitalisation on growth rates of three types of assets--total assets, loans, and government bonds--as well as subordinated debt, which counts toward tier II capital.
The results show that capital adequacy requirements introduced under the 1988 Basel Accord affect the behaviour of international and domestic banks in Japan very differently. International bank portfolios are marginally sensitive to total regulatory capital in the post-Basel period. There is clear evidence that in the post-Basel period international banks with relatively low core capital ratios tend to shrink their overall assets and to shift their asset portfolio out of heavily-weighted risky ones into zero-weighted riskless assets. The sensitivity of the international bank portfolio is a post-Basel phenomenon, which suggests that the Accord brought about significant changes in the behaviour of international banks but did not affect the portfolio allocation of domestic banks.
Koehn and Santomero (1980) examine the issue of portfolio reaction to capital requirements by analysing the effect of capital ratio regulation on portfolio behaviour of commercial banks and the effects on bank portfolio risk of a regulatory increase in minimum capital asset ratio that is acceptable to the supervisory agency. Their results show that, for the system as a whole, the results of a higher required capital-asset ratio in terms of the average probability of failure is ambiguous. They then question the viability of regulating commercial banks in terms of capital requirements. They suggest that serious consideration should be given to the discontinuance of regulating banks via ratio constraints.
Alexander and Baptista (2006) attempt to show that holding minimum capital levels might not decrease fragility but rather encourage banks to engage in 'regulatory capital arbitrage' which occurs where the bank reduces the amount of capital required by increasing the risk of its trading portfolio. They assume a bank uses a mean variance model when managing its trading portfolio and, as a consequence of the Basel Capital Accord, bank management decides to control the portfolio's risk by imposing a constraint on the size of its value at risk. They find that, under certain circumstances, there would be an increase in the risk of the bank's portfolio and therefore bank fragility. The BCP allow a bank to engage in arbitrage by being able to reduce the amount of capital required while increasing the risk of the portfolio.
In contrast to the studies discussed above, there are some that support the need for regulation of banks. Podpiera (2006) studies the relationship between banking sector performance and the quality of regulation and supervision as measured by compliance with the BCP. The main question is whether BCP compliance creates a regulatory and supervisory environment that helps improve banking sector performance. He uses non-performing loans and net interest margins as performance indicators. The level of nonperforming loans reflects the degree to which banks are able to perform one of its basic functions: collecting the money it lends. Net interest margin is the measure of the efficiency of banking sector performance since it indicates the cost of banking intermediation. He finds a direct positive effect of compliance with the Basel Core Principles on banking sector performance, as measured by the share of nonperforming loans and net interest margins. These results suggest that being Basel compliant creates a regulatory environment that helps improve banking sector performance: higher compliance is associated with lower non-performing loans and a narrower net interest margin.
Demirguc-Kunt, Detragiache, and Tressel (2008) provide yet another study that establishes benefits of regulation of the banking industry. They investigate whether compliance with the BCP improves bank soundness. An important aspect of their study is that they differentiate among different elements of the regulatory framework to help prioritise reform efforts. They find a significant and positive relationship between compliance with information provision and bank soundness. Their results suggest that countries aiming to upgrade banking regulation and supervision should consider giving priority to information provision over other elements of the core principles. Because information provision is a necessary condition for effective discipline, this policy recommendation is consistent with the approach to regulation and supervision recommended by Barth et al. (2004), who stress the importance of mechanisms to empower market discipline and are skeptical about structures that assign too much power to regulators. Bongini et al. (2002) also find that market disclosure is very important. They compare the forecasting ability of traditional early warning indicators of bank fragility (CAMEL type balance sheet variables) with market indicators (default probabilities estimated from equity data using a Merton model and credit ratings) for banks active in the South East Asian countries. They conclude that the indicators they investigated have little predictive power although balance sheet indicators provide information about the cross sectional pattern of failures and non-failures across banks. Their results indicate that there is a need for greater focus on Pillar 3 of the Basel Accord regarding market disclosure on the quantity and quality of data released by banks in emerging markets.
Data and Methodology
In order to test whether compliance affects commercial bank performance in Trinidad & Tobago, econometric models are specified based on the following general formulation:
Y =f(bcp, op, pll, gdp, irate, erate, lrate) (1)
Y is a measure of bank performance, which will be explained by bank-specific explanatory variables consisting of a bank's BPRR compliance variable, bcp, its operating expense ratio op, and its provision for loan loss to total assets ratio, pll. Explanatory aggregate macroeconomic variables are the growth rate in real GDP, gdp, the inflation rate, irate, the exchange rate, erate and the prime lending rate, lrate.
All data used are in annual format and cover the 10-year period 1997-2006. Balance sheet data from the annual reports of all the banks are used to obtain measures of the bank-specific variables. Four alternative measures of Y are used initially: net interest margin (him), return on assets (roa), return on equity (roe) and nonperforming loan ratio (nplr). Of the measures, nim and nplr are measures of efficiency (Podpiera 2006): lower values imply greater efficiency. The roa is more a measure of profitability than efficiency while the roe variable may be considered a bit of both since it measures the return on shareholders' equity. Improved performance is indicated therefore by lower values of nim and nplr, and higher values of roa and roe.
These four indicators are used as measures of bank performance in preference to traditional indicators such as total assets, performing loans and non-performing loans. Koch and Mac Donald (2003) argue that the latter suffer from three basic flaws: (1) they ignore the wide diversity of strategies pursued by different institutions; (2) a bank's total assets no longer serve as a meaningful yardstick when banks engage in off-balance sheet activities and (3) analysis using such variables provides no direct information concerning how or which of the banks contribute to the creation of shareholder value. Ikhide (2000) also argues against the use of 'aggregative indices' such as total assets, loans or deposits as an index of bank performance since none of these can sufficiently capture bank performance. Faced with these problems, more and more studies have resorted to the use of accounting data on bank margins, costs and profits as measures of bank performance (Podpiera 2006), as is done in this paper.
The data for measuring bcp were obtained from the Central Bank of Trinidad and Tobago (CBTT) for the four major banks in Trinidad & Tobago. (3) The CBTT was asked to rate the individual banks on the level of compliance with the BPRR over the period 1997-2006 on a scale of 0 to 10. A score of 0 indicates that the bank is non-compliant, between 1-4 that it is materially non-compliant, between 5-9 that it is largely compliant, and 10 that it is fully compliant. The CBTT was also asked to provide the relative weight it attached to each of the BPRR. The overall summary measure of compliance, bcp, is calculated as a weighted average of ratings for each principle:
[bcp.sub.it] = [m.summation over (j=1)] [w.sub.j][r.sub.ijt]
where [w.sub.j] is the principle specific weight as provided by the Central Bank, and r are the ratings of the principles, which are bank-specific and varying over time. Note that this measure is necessarily bounded between 0 and 10.
The trend in the bcp index over time for each of the four banks is depicted in Figure 1. BPRR compliance appears to be steadily increasing over time for Banks 1, 3 and 4 while it is showing a definite increase in the case of Bank no. 2 after 2002, despite a somewhat downward trend before that.
[FIGURE 1 OMITTED]
The macroeconomic data were obtained from the CBTT's online database. The variables chosen are similar to those employed by Sudararajan et al. (2001) and are reflective of how the economy is performing and are expected therefore to affect a bank's performance.
Summary statistics of all variables used in the analysis are shown in Table 1. In the case of the explanatory variable of main interest, bcp, note that on average a bank's summary rating is a little above 7.4. However, the sizeable standard deviations show that there are considerable differences across the different banks and over time.
A possible specification of (1) is
[Y.sub.it] = [alpha] + [[beta].sub.1][bcp.sub.it] + [[beta.2]][op.sub.it] + [[beta].sub.3]][pll.sub.it] + [[beta].sub.4][gdp.sub.t] + [[beta].sub.5][irate.sub.t] + [[beta]sub.6][erate.sub.t] + [[beta].sub.7] [lrate.sub.t] + [[epsilon].sub.it]
where i denotes the bank's subscript and t is subscript indicating the year. An appropriate estimating procedure is Pooled Ordinary Least Squares (POLS). However, one possible problem with this specification is that the number of explanatory variables is limited and there may be other determinants of bank performance that cannot be controlled for, the effect of which would be captured in the error term. If these were correlated with one (or several) of the other explanatory variables, then the estimated coefficients would be biased. One possibility is to assume that all such unobservable factors are time invariant:
[[epsilon].sub.it] = [[mu].sub.i] + [v.sub.it]
where [mu] represents the time invariant, bank-specific, unobservables and n is a standard i.i.d, error term. Under this assumption, a panel fixed effects (FE) estimator may be employed, in preference to a POLS so as to purge these time invariant unobservables, where all variables are transformed into deviations from their mean.
By using the FE estimator, one is essentially subtracting the (cross-bank) mean from all variables. This removes cross-bank (time invariant) variation, leaving only the variation of variables over time within banks. The variation to be explained will primarily come over time from within rather than across banks. A classic F-test may be conducted to establish whether the assumptions underlying the use of the FE estimator, as opposed to those of the POLS estimator, are satisfied. The POLS may then used in cases where the FE estimator proves to be unsuitable.
Another possibility to both the FE and the POLS arises, however, if individual bank heterogeneity is uncorrelated with included variables, in which case [mu] may be taken to represent a bank-specific random element and the Random Effects (RE) estimator becomes the appropriate one to use. The well known Hausman test may be used to rule in favour of either the FE or RE estimator.
An increase in the value of bcp, the variable that concerns us the most, may result in improved performance or in performance deterioration. An increase in op, however, ought to cause nim and nplr to fall while causing roa and roe to rise. An increase in the value of pll may cause performance to increase or decrease, depending on the 'comfort' an individual bank takes in providing contingencies for losses. Increases in gdp may also have ambiguous consequences: rising gdp may result, for instance, as a consequence of an increase in the price of oil, which is often accompanied by excess liquidity in the banking system, which in turn may encourage banks to be more lax in the granting and management of loans. Increases in erate and irate are likely to result in performance deterioration, while an increase in Irate may have either a positive or a negative effect.
The possible direction of response of the Y variables (nim, roa, roe and nplr) to the seven explanatory variables is summarised in Table 2:
Table 2: Expected Direction of Response of Performance Variables nim roa roe nplr bcp [+ or -] [+ or -] [+ or -] [+ or -] op - + + - pll [+ or -] [+ or -] [+ or -] [+ or -] gdp [+ or -] [+ or -] [+ or -] [+ or -] erate + - - + irate + - - + lrate [+ or -] [+ or -] [+ or -] [+ or -]
Panel FE and RE estimators are obtained using the four alternative measures of Y: nim, roa, roe and nplr. The results obtained for the roe specification in both the FE and RE versions of the model are very poor since none of the variables is significant and the overall fit of the model is poor. This is so even when the model is specified using only the best performing explanatory variable (the one with the lowest p-value, which was bcp). To a large extent, these poor results were anticipated given the high variation of roe shown in Table 2, where the standard deviation is almost four times as large as the mean. This roe specification shall therefore be considered no further in this paper.
Two of the macroeconomic variables, gdp and irate, are very insignificant in all specifications of the model, a result that is at odds with those reported by Sundararajan et al (2001). Because of poor performance, these two variables are eliminated from further consideration. Other variables are progressively eliminated from the various specifications based on having p-values larger than 10% and the FE and RE versions compared using a Hausman test, which rules in favour of the RE estimation in all three cases. A summary of the results obtained is given in Table 3:
The overall fit for each retained specification is acceptable as measured by the classic [chi square] and [R.sup.2] statistics. The main concern of this paper is whether compliance with the BPRR affects bank performance. Here, bcp, the measure of Basel compliance, is significant at the 5% level in the nim specification and at 1% in the remaining two retained specifications: greater compliance results in an increase in the value of the net interest margins (performance deterioration), an increase in the return on assets (performance enhancement) and a decrease in the value of the non-performing loan ratio (performance enhancement).
That Basel compliance is the only significant variable in all three specifications is an extremely important finding: Basel compliance is more important, or at least as important as, other bank-specific and macroeconomic variables in explaining the performance of commercial banks in Trinidad & Tobago. Noteworthy, however, is that it does not result in improved net interest margins. On the other hand, it does result in better loan performance which, together with rising interest margins, may, in turn, explain why the return on assets (and on equity) improves.
The exchange rate and the prime lending rate have a significant impact on the net interest margin: as expected, when the exchange rate (measured in Trinidad and Tobago dollars per US dollar) increases, net interest margins also increase and, as the prime lending rate increases, the net interest margin decreases. The latter result is consistent with the view that the interest margin is elastic to the loan rate so that, as the loan rate increases, borrowing decreases sufficiently to allow for an overall fall in loan revenues.
Of the bank-specific variables, the operating cost ratio is significant only in the explanation of nim, in which case it is highly significant (1%) and negatively signed: the net interest margin, as is to be expected, falls as operating costs increase. The provision for loan loss positively affects non-performing loan ratio (5% significance), which may be interpretated to mean that, as banks make more provisions for bad loans, lending policies tend to become more lax (something of a moral hazard problem) and the value of non performing loans increases. The provision for loan loss has a positive effect on the return on assets and is strongly significant in this case (1%): as the provision is increased, the return on assets is higher.
Compliance improves performance
This paper investigates how compliance with BPRR affects the performance of an individual bank in Trinidad & Tobago. To this end, a unique data set of compliance ratings, macroeconomic and balance sheet variables was compiled for four major banks in Trinidad and Tobago and analysed using panel data estimators.
The results show that BPRR compliance significantly affects the net interest margin, the return on assets and the non-performing loan ratio. Among the macroeconomic variables, the growth rate and the inflation rate affects none of the four performance indicators while the exchange rate and the prime lending rate have a significant impact only on net interest margin. In terms of the bank-specific variables, the operating cost ratio is significant in one specification only while the provision for loan loss ratio positively affects return on assets and the non-performing loss ratio.
There is ample evidence that BPRR compliance brings about performance improvement, although there is some exception to this rule. This is good news for banks in emerging economies since the Bank of International Settlements expects all banks, whatever the jurisdiction in which they operate, to become compliant in a not too distant future. There is ample evidence that if banks, particularly in emerging economies, want to remain or become competitive, they must invest time and resources to implement these principles.
Alexander, O., and A. Baptista. 2006. "Does the Basle Capital Accord reduce bank fragility? An assessment of the value-at-risk approach." Journal of Monetary Economics 53:1631-660.
Barth, J., G. Caprio, and R. Levine. 2001. "The Regulation and Supervision of Banks Around the World: A New Database." In Integrating Emerging Market Countries into the Global Financial System, edited by Robert E. Litan and Richard Herring. Brookings-Wharton Papers on Financial Services. Washington, DC: Brookings Institution Press.
--. 2004. "Bank Regulation and Supervision: What Works Best?" Journal of Financial Intermediation 13(2): 205-48.
Bongini, P., L. Laeven, and G. Majnoni. 2002. "How good is the market at assessing bank fragility? A horse race between different indicators." Journal of Banking and Finance 26(5): 1011-28.
Caprio, G., and D. Klingbiel. 2002. "Episodes of Systemic and Borderline Financial Crises." In Managing the Real and Fiscal Effects of Banking Crises, edited by D. Klingebiel and L. Laeven, 31-49. World Bank Discussion Paper No. 428. Washington, DC: World Bank.
Demirguc-Kunt, A., E. Detragiache, and T. Tressel. 2008. "Banking on the Principles: Compliance with Basel Core Principles and Bank Soundness." Journal of Financial Intermediation 17(4): 511-42.
Diamond, D.W., and EH. Dybvig. 1986. "Banking Theory, Deposit Insurance, and Bank Regulation." Journal of Business 59: 53-68.
--. 1983. "Bank Runs, Deposit Insurance, and Liquidity." Journal of Political Economy 91: 401-19.
Grynberg, R., and S. Silva. 2006. "Harmonization without Representation: Small States, the Basel Committee, and the WTO." World Development 34(7): 1223-36.
Ikhide, S. 2000. "Efficiency of Commercial Banks in Nambia." Bank of Namibia Occasional Paper No. 4.
International Monetary Fund, and World Bank. 2002. "Implementation of the Basel Core Principles for Effective Banking Supervision: Experiences, Influences, and Perspectives." Washington, DC: IMF. www.imf.org/external/np/mae/bcore/2002/092302.htm.
Jackson, P., and W. Perraudin. 2002. "Introduction: Banks and Systemic Risks." Journal of Banking and Finance 26: 819-25.
Koch, T., and S. MacDonald. 2003. Bank Management. 5th edition. Cincinnati, OH: South-Western.
Koehn, M., and A. Santemero. 1980. "Regulation of Bank Capital and Portfolio Risk." Journal of Finance 35(5): 1235-44.
Majnoni, G., and A. Powell. 2005. "Reforming Bank Capital Requirements: Implications of Basel II for Latin American Countries." Economia 5(2): 105-49.
Montgomery, H. 2005. "The effect of the Basel Accord on bank portfolios in Japan." Journal of the Japanese and International Economies 19: 24-36.
Podpiera, R. 2006. "Does Compliance with Basel Core Principles Bring Any Measurable Benefits?" IMF Staff Papers 53(2): 306-25.
Sundararajan, V., D. Marston, and R. Basu. 2001. "Financial System Standards and Financial Stability: the Case of the Basel Core Principles." IMF Working Paper WP/01/62. Washington, DC: International Monetary Fund.
* The authors gratefully acknowledge the kind assistance of the Bankers' Association of Trinidad & Tobago and the Central Bank of Trinidad & Tobago, both of whom collaborated in providing information on Basel compliance. We would particularly like to thank Mr David Dulal-Whiteway, President of the Bankers' Association at the time that the research was being undertaken, and Mr Carl Hiralal, Inspector of Banks at the Central Bank. The usual caveat about the authors' sole responsibility for errors applies.
(1) The BPRR correspond to principles 9-16 of the Basel Core Principles for Effective Banking Supervision.
(2) This information was provided by the Central Bank of Trinidad & Tobago with the pre-approval of the Bankers' Association of Trinidad & Tobago, the umbrella body of all commercial banks in that country. This information is not normally available for studies like this one and a pre-condition for the release of the data was that the individual banks should not be named.
(3) The questionnaire used to obtain this information is available from the authors on request.
Table 1: Summary Statistics of Model Variables Standard Variable Mean Deviation BPRR index 7.465 0.6437272 Operating expense ratio 0.3526672 0.23892 Net Interest Margin 2.2204 0.76516 Return on Assets 0.0200588 0.0245034 Return on equity 0.1413488 0.522823 Provision for loan loss to Average Asset 0.118123 0.013347 Non performing loan ratio 0.0479562 0.0396074 Non performing loan coverage ratio 0.812559 0.8291881 GDP growth 8.635 2.794237 Exchange rate 6.22745 0.319152 Inflation rate 4.995 1.986702 Prime lending rate 13.463 3.215193 Table 3: RE Estimation of Models of Bank Performance nim roa nplr bcp 0.273 ** 0.017 *** -0.030 *** (0.115) (0.004) (0.008) OP -1.05 *** 0.027 *** (0.290) (0.011) p11 1.715 *** 0.999 ** (0.198) (0.420) erate 3.80 * 0.307 ** (2.02) (0.158) Irate -0.158 *** (0.022) Constant -21.03 * -0.136 *** -1.65 * (12.48) (0.035) (0.992) Observations 38 38 39 [chi square] 0.00 0.00 0.00 [R.sup.2] 0.78 0.70 0.51 Hausman 0.507 0.526 0.132 Notes: (1) [chi square]: p-value of [chi square]-test for overall goodness of fit of RE model; (2) Hausman: p-value of Hausman test of FE vs RE; (3) Standard errors in parentheses; (4) ***, **, and * 1%, 5%, and 10% significance levels respectively.