Afanasieff, T., P. Villa Lhacer and M. Nakane, 2002, “The Determinants of Bank Interest Spreads in Brazil,” Central Bank of Brazil Working Paper Series No. 46, Banco Central do Brasil (August).
Biderman, G., C. Velázquez, S. Giménez and Z. Espinola, 2012, “Costos de la Información en el Sistema Paraguayo,” Central Bank of Paraguay Working Paper, No. 36.
Brock, P. and H. Franken, 2003, “Measuring the Determinants of Average and Marginal Bank Interest Rate Spreads in Chile, 1994–2001,” Economia Chilena, Vol. 6(3), pp. 45–65.
Espino, F. and C. Carrera, 2005, “Concentración Bancaria y Margen de las Tasas de Interés en Perú,” Estudios Económicos del Banco Central del Reserva del Perú.
Espinola, Z. and C. Velázquez, 2012, “Tras los Mitos de la Banca Extranjera en América Latina: Estudio del Caso Paraguayo,” Central Bank of Paraguay, mimeo.
Estrada, D. and I. Orozco, 2006, “Determinants of Interest Rate Margins in Colombia,” Borradores de Economía No. 393, Banco de la Republica de Colombia.
Gustale, J., 2011, “Hacia una Supervisión Basada en Riesgos: ¿Y Qué tal si Aprovechamos esta Coyuntura Económica para Fortalecer la Legislación Financiera?” Central Bank of Paraguay.
International Monetary Fund, 2004, “Bank Profitability and Competition,” Technical Note, Financial Sector Assessment Program Update, Republic of Kazakhstan, IMF Country Report No. 04/336.
Mlachila, M., 2009, “Recurrent Financial Crises: Causes, Costs, and Consequences” in Paraguay: Addressing the Stagnation and Instability Trap.
Morón, E., J. Tejada, and A. Villacorta, 2010, “Competencia y Concentración en el Sistema Financiero en el Perú,” Documentos de Discusión, DD/10/03, Universidad del Pacifico.
World Bank, Doing Business Database, 2012.
Prepared by Kevin Ross (WHD) and Viviana Garay (WHD-Asunción Office). The exact same analytical framework was used in Ross and Peschiera (2012) in their examination of Peruvian interest rate spreads—allowing for a comparison between these two Latin American banking systems.
The reduction reflected a generalized lack of confidence in the system rather than any particular “flight to quality” effect.
In 2010 Itau from Brazil merged with Interbanco, becoming the 2nd largest bank in the system. One small bank closed operations in 2011, while financieras (Vision, Itapúa, Familiar and Atlas) and one cooperative (Bancop) started operating as banks. CCB China (2005), Lloyds TBS (2007) and ABN Ambro (2009) left the system as HSBC (2007) began operations.
See IMF (2004) for a full derivation. In the analysis, we use annual income and balance sheet data from the BCP and from the SIB.
It can be difficult to discern the true level of interest bearing assets and liabilities to use in the calculation of the effective lending and deposit interest rates. This can result in large residuals.
Paraguayan banks report large valuation inflows and outflows on foreign exchange trading and other non-core banking activities. Given that non-interest income is scaled by deposits, large valuation inflow effects that do not take into consideration non-interest valuation expenses can sharply increased this ratio, implying a large non-interest return that could be used to help to cover costs. This would be reflected by a large positive residual. To avoid this distortion, we report the net non-interest income ratio. However, this implies a somewhat smaller non-interest income margin—which can result in negative residuals.
Variables not defined in the text include: (i) EBT, earnings before tax; (ii) TA, total assets; (iii) IR, interest revenue; (iv) IE, interest expense; and (v) EA, earning assets.
Looking from the bottom up in 2012, the return on earning asset was 11.7 percent. However, 4.4 pps (1.2 x 3.6) are subtracted given the cost of liabilities and their size relative to earning assets—leaving a 7.3 percent NIM. Since only 72 percent of all assets are utilized for earning activities, the return is further cut to 5.6 percent. Given the need to cover administrative burdens it is reduced by another 2.7pps to 2.6 RoA. This RoA, however, is levered up 8.7 times, increasing the return to 22.6 percent, whereby tax policies lower it to 20.7 percent RoE. Compared to 2011, lower leverage was a key factor behind the reduction in RoE.
All estimations are done in EVIEWS using unbalanced panel data regression techniques assuming fixed effects. Robust estimators were calculated using White cross sectional SUR corrections to ensure robust standard errors.
Banks in Paraguay hold few disposable investments (e.g., bonds) which are actively traded. Thus the market risk variable was dropped from the regression analysis.
Annex I. Description of Empirical Models
1. The empirical model follows Weber and Saborowski (2013) and implements an interaction panel VAR (IPVAR) framework introduced by Towbin and Weber (2013). The model includes the central bank policy rate and a weighted average of bank lending rates in a monthly frequency from 2000–12, for all countries for which data from the IMF’s International Financial Statistics is available (120 countries). The dynamic interaction between these endogenous variables is allowed to vary deterministically with the share of dollar denominated loans as a share of total loans. The relationship between the variables in this model is assumed to be governed by a system of “structural” equations. Ignoring the constant term, the system can be written as:
The reduced form of the structural model can be written as:
The model can be estimated by ordinary least squares, imposing restrictions on A0 to identify the coefficients in the structural form. Identification is achieved through a Choleski decomposition of the variance-covariance matrix Σe of reduced-form errorst. What is more, conditional impulse response functions are constructed and evaluated at different points of the sample distribution of the dollarization ratio.
2. Paraguay specific VARs are also estimated to quantify the effects of dollarization on interest rate transmission and the exchange-rate pass-through to inflation. The VAR includes the central bank policy rate, credit dollarization ratio, the weighted average of bank lending rates with monthly data from 2000–2013 (all in first differences). The pass-through of exchange rate to inflation is calculated using a VAR with the nominal effective exchange rate, deposit dollarization ratio, and CPI monthly inflation. In both estimations a recursive ordering is assumed, characterized by the idea that the more exogenous variables of the model precede the endogenous ones.2 To quantify the effects of dollarization on the interest rate transmission and in amplifying the exchange-rate pass-through, counterfactual scenarios are constructed by holding impulse responses of dollarization ratios fixed at zero at all forecast horizons.3 This hypothetical impulse response is then compared with the actual response to quantify the relative importance dollarization in hindering the interest transmission and amplifying the exchange-rate pass-through. To see this more clearly, consider the impulse response of the dollarization ratio (DOLLAR) to a shock to the policy rate (LRM):
The counterfactual analysis looks at a sequence of shocks such that (1) is equal to zero. Therefore, in the first month after impact this entails:
The required innovation
The required values for subsequent innovations can be recursively estimated as:
where h= 1, …, H.
The impulse response to the a shock to the central bank policy rate holding dollarization equal to zero at all horizons h is computed as:
with i being an indicator for each variable in the system.
Bachmann, R. and E. Sims, 2011, “Confidence and the Transmission of Government Spending Shocks,” NBER Working Papers 17063, National Bureau of Economic Research, Inc.
Bernanke, B., M. Gertler, and M. Watson, 1997, “Systematic monetary policy and the effects of oil price shocks,” Brookings Papers on Economic Activity, 28(1): 91–157.
Daban Sanchez, T., 2011, “Bank Excess Reserves in Paraguay—Determinants and Implications,” Chapter II, IMF Country Report No. 11/239.
García-Escribano, M. and S. Sosa, 2011. “What is Driving Financial De-dollarization in Latin America?” IMF Working Paper No. 11/10
Reinhart, C., R. Rogoff, and M. Savastano, 2003, “Addicted to Dollars,” NBER Working Paper 10015 (Cambridge, MA: National Bureau of Economic Research, Inc.).
Saborowski, C. and S. Weber, 2013, “Assessing the Determinants of Interest Rate Transmission Through Conditional Impulse Response Functions,” IMF Working Paper Series, WP/13/23.
Saborowski, C., S. Sanya, H. Weisfeld and J. Yepez, 2014, “Effectiveness of Capital Outflow Restrictions,” IMF Working Paper Series, WP/14/8.
Towbin, P., and S. Weber, 2013. “Limits of floating exchange rates: The role of foreign currency debt and import structure,” Journal of Development Economics, Elsevier, vol. 101(C), pages 179–194.
Prepared by Juan F. Yépez.
See Appendix I for description of the empirical models used in this section.
The empirical model follows the interaction panel VAR framework proposed by Towbin and Weber (2013).
The lag length was chosen according to the Akaike Information Criterion (AIC).
The model is qualitatively robust to different orderings.
This counterfactual analysis follows closely from the analysis of Sims and Zha (1998), Bernanke, Gertler, and Watson (2003), and Bachman and Sims (2012).