European Journal of Scientific Research ISSN 1450-216X Vol.36 No.2 (2009), pp.172-183 EuroJournals ublishing, Inc. 2009 http://www.eurojournals.com/ejsr.htm Volatility of onetary olicy in a Developing Economy: In Context of akistan Syed Khurram Arslan Wasti -hil Fellow, Applied Economics Research Centre University of Karachi, Karachi-akistan E-mail: arslan_bzu@hotmail.com Sulaiman D. uhammad Associate rofessor, Department of Economics Federal Urdu University, Karachi-akistan E-mail: sulaiman1959@gmail.com Narjis Khatoon - hil Fellow, Applied Economics Research Centre University of Karachi, Karachi-akistan Email: rizvi.narjis@yahoo.com Irfan Lal -hil Fellow, Applied Economics Research Centre University of Karachi, Karachi-akistan E-mail: irf_yoch@yahoo.com Abstract This study is to analyze the volatility of the monetary policy in less developed countries like akistan is carried out for the period of 37 years from 1971 to 2007. Different econometrics techniques are used i.e. unit root analysis, Johnson co integration technique & error correction mechanism for long run and short run association respectively. Further objective of this study is to examine some hypothesis of the ackinnon model with reference to akistan s economy and scrutinize the instruments of monetary policy and determinants of money supply in akistan. The estimated results show that there is a long-run equilibrium association among variables but we didn t able to predict short run equilibrium in first three models except the last model which reflects that real rate of interest and real private investment significantly effect the real money demand in the long run. The empirical results show that time deposits and broad money have a less volatile behavior in case of akistan s economy and also explain that stock of money reserve causes a large portion of variation in money supply. Keywords: onetary olicy, Volatility, Co-Integration Jel Classification Codes: E42, C44 Electronic copy available at: http://ssrn.com/abstract=1577033
Volatility of onetary olicy in a Developing Economy: In Context of akistan 173 I. Introduction It is important in monetary policy to take precise measure of size and direction of change. A number of variables and indicators are used as a measure of the position towards the monetary policy in across the world. In case of open economy, it is assumed that monetary policy affects the economy and prime objective is to evaluate the monetary policy, rate of inflation through two important transmissions. These transmissions are the interest rate and exchange rate channel. Two apparent objectives have been given to State Bank of akistan first: to stabilize monetary system. Secondly: to take initiative towards the perfect utilization of country s productive resources. Target rate of growth and inflation rates set by the federal government as the targets of monetary policy stance. In the context of the developed economies the Keynesian-monetarist debate has concentrated on the efficiency of stabilization policy. The effectiveness of fiscal and monetary policies depend upon the size and significance of macroeconomics variables such as marginal propensity to save, the interest elasticity of money demand and interest elasticity of investment (slope of IS and L Curve). One of the objectives of this paper is to test some hypothesis of the ackinnon model with reference to akistan economy. Another objective is to examine the instruments of monetary policy and determinants of money supply in akistan. The structure of this paper is as follow, Section II analyze the review literature. Section III represents brief description about ackinnon model, section IV describes the empirical evidence and instruments for the monetary policy in akistan and determinant of money supply will be discussed in section V. Summary and conclusion will be described in Section VI. II. Literature Review Rehman & Ahmed (2002) found that monetary base remained an important determinant of monetary stock (2) and the monetary policy remained an effective instrument in controlling and regulating monetary assets in akistan. Khalid (2005) suggests that imported inflation, seinorage and openness are key factors which cause to inflation in akistan. As well as result indicates that imported inflation, deficit-gd ratio, seinorage, money depth, exchange rate depreciation and domestic credit can be important determinants of inflation in akistan. Tehran & Lausing (2003) suggest that discretionary monetary policy in a model that allow for varying degrees of forward-looking behaviors. Forward- Looking behaviors, the optimal monetary policy can response either satisfy or expand to output gap; the inflation depends upon the portion which the economic model is affected by expectation. Qayyum (2002) recommends that akistani government should use monetary condition index to check the stance of the monetary policy. CI is designed to suggest as an indicator of monetary policy which helps the monetary policy stance. III. ackinnon odel onetary approach often prescribes low interest rate policies in order to raise investment and thereby aggregate demand. In the ackinnon model, the traditional policies is identified as one of financial repression that leads to disintermediation, follow low investment and affected the market segmentation in developing economies. The ckinnon Shaw hypothesis suggests that the level of financial intermediation should be closely related to the prevailing level of the real interest rate, the reason being that the level of real interest rate when held down their normal competitive levels which indicates the extent of financial repressions [Gregorio & Guidotti (1995)]. The main features of the ckinnon model are summarized as follows: the main variables which lead to retardation of growth in lower developing countries (LDC) are the fragmentation of capital and capital and money market. The main objective of the monetary policy should be promoted the development of capital and financial market rather then short run stabilization. As we know money and capital are the substitutes to each other. Electronic copy available at: http://ssrn.com/abstract=1577033
174 Syed Khurram Arslan Wasti, Sulaiman D. uhammad, Narjis Khatoon and Irfan Lal ckinnon argues that because of limited availability of alternative financial assets in the developing countries, money is the major channel for debt intermediation process associated with capital accumulation. oney and capital should be regards as compliment rather then substitute. The demand for money function according to ckinnon is written as: d = β Y I 2 3 i Y 4 1 β 5 e Where d is the demand for money balances including savings and time deposits, Y is the real income, i is the nominal interest rate of saving and time deposit, e is the expected rate of inflation I/Y is the ratio of real investment to real income. Thus, as to overcome at above fluctuations, it is necessarily to provide high rate of interest on monetary assets, in order to shift the portfolio of saving from non productive physical goods to monetary assets. In the process of capital formation is the insufficient flow of monetary saving to finance investment projects rather than the shortage of opportunities to profitable investment. The key policy object of the ckinnon Shaw school of thought is that the government constraint on banking such as (such as interest rate ceilings, high cash reserve requirement and directed credit programs) impede the process of financial development and consequently reduce economic growth [Husseun & Demetriades (1996)]. III. Data Analysis Following doctoring of ackinnon models for akistan is estimated. d Log( ) = β 1 2 log( r) 3 log( Y ) + µ (1) d Log( ) = β 1 2 log( i) 3 log( & ) 4 log( Y ) + µ (2) d Log( ) = β 1 2 log( Y ) + µ (3) d Log( ) = β 1 2r 3 log( RI) + µ (4) Where d : Demand for Re al oney Balances. r : The Re al Rate of Interest on Saving & Time Deposits. i: No min al Interest Rate. Y : The Re al Gross Domestic r oduct. & : The Rate of Change in Consumer r ice Index. with log. RI : Re al r ivate Investment with log. any empirical studies have considered log linear form in the analysis. The coefficient in the form of log linear calculates the elasticity directly. oreover log linear form reduces the problem of hetrosecdasticity in empirical analysis. Thus empirical results show that this model is suitable for policy inference. Data is taken annually from 1970-2007 various sources i.e. Annual Reports of SB, Economic Survey of akistan, World Development Indicators (WDI) & International Financial Statistics (IFS). This study using Johnson ultivariate cointegaration technique to examine the long run equilibrium relationship. For co integration test it is necessary to access whether a variable series is stationary or not, for this purpose we have to use Unit Root Analysis before testing of co integration.
Volatility of onetary olicy in a Developing Economy: In Context of akistan 175 Since it is essential, but not suitable condition for a set of variables to be co integrated. Augmented Dickey Fuller test (ADF) is employed for a unit root analysis. This test is for stochastic non stationary. In this phase we examine the matter that proposed by ckinnon, the demand function for money as discussed by ckinnon that either akistan economy shows the evidence of financially repressed economy or not. ckinnon incorporates two elements; first is the demand function of money in the developing countries additionally the relevant rate of interest the rate of saving and time deposits. Secondly the accurate measure of inflation relevant with the demand function for money in developing countries is the rate of change in consumer price index. Table: 1 Unit Root Test Result Variables Augmented Dicky Fuller (ADF) LEVEL 1 st Difference Inflation -0.44-5.61* Nominal Rate of Interest -0.49-7.13* Real GD 0.210-8.33* Real oney Demand 0.56-2.09** Real rivate Investment 2.25 9.61* Real Rate of Interest -3.13-5.80* Budget Deficit -0.62-3.39* Reserve oney 0.65-9.50* * & ** shows the level of significance at 1% and 5% respectively. Schwarz Information Criteria, Akaiken Information Criteria are used to select of optimal lag length. Almost all results show that all series are unit root at level but stationary at first difference. The Johnson ultivariate trace and maximum Eigen values cointegration is applied to the variables in model 1,2,3 and 4 is presented in following given tables: Demand for oney in akistan, 1971-2007: Regression Results. Table-2: Johnson Test for Cointegration Trace Test r 1 33.68* 29.79 r 2 5.07 15.49 r 2 r = 3 0.40 3.84 aximum Eigen Values r = 1 28.60* 21.13 r = 2 4.66 14.26 r 2 r = 3 0.40 3.84 * shows the 5% Level of significance Variables included in the vectors are: RD, R & RGD
176 Syed Khurram Arslan Wasti, Sulaiman D. uhammad, Narjis Khatoon and Irfan Lal Error Correction odel Estimates Dependent Variable RD Variables Coefficient Stand Error T-Values R 0.008 0.004 1.93 RGD -0.826 0.137-6.01 µ 1( 1) 0.07116 0.074 0.96 Constant 0.083 0.020 4.00 Table 3: Johnson Test for Cointegration Trace Test r 1 77.80* 47.85 r 2 24.88 29.79 r 2 r 3 6.58 15.49 r 3 r = 4 0.45 3.84 aximum Eigen Values r = 1 52.92* 27.58 r = 2 18.29 21.13 r 2 r = 3 6.13 14.26 r 3 r = 4 0.45 3.84 * shows the 5% Level of significance Variables included in the vectors are: RD, I, INF & RGD Error Correction odel Estimates Dependent Variable RD Variables Coefficient Stand Error T-Values I -0.14 0.062-2.39 INF -0.12 0.039-3.27 RGD -0.58 0.126-4.59 µ 2 ( 1) -0.0069 0.067 0.088 Constant 0.069 0.017 3.977 Table 4: Johnson Test for Cointegration Trace Test r 1 35.38* 25.87 r = 2 3.85 12.51 aximum Eigen Values r = 1 31.53* 19.38 r = 2 3.85 12.15 * shows the 5% Level of significance Variables included in the vectors are: RD & RGD
Volatility of onetary olicy in a Developing Economy: In Context of akistan 177 Error Correction odel Estimates Dependent Variable RD Variables Coefficient Stand Error T-Values RGD -0.80 0.14-5.56 µ ( 1) 0.075 0.07 1.01 3 Constant 0.082 0.021 3.75 Table-5: Johnson Test for Cointegration Trace Test r 1 38.60* 29.79 r 2 15.06 15.49 r 2 r = 3 2.30 3.84 aximum Eigen Values r = 1 23.54* 21.13 r = 2 12.76 14.26 r 2 r = 3 2.30 3.84 * shows the 5% Level of significance Variables included in the vectors are: RD, R &RI. Error Correction odel Estimates Dependent Variable RD Variables Coefficient Stand Error T-Values R 0.007 0.005 1.29 RI 0.028 0.032 0.86 µ 4 ( 1) -0.33 0.132-2.47* Constant 0.027 0.025 1.06 * shows the 5% Level of significance Table 2, 3, 4 & 5 represent both Trace and Eigen values give the same result. The alternative hypothesis of co integration is accepted for r = 0 for both Trace and Eigen tests at 5 % level of significance. This implies that there is one cointegration vector. Thus it is concluded that even though the individual variables series are non stationary but their linear combination is stationary. Thus, the level of estimates for policy implications, we apply the Error Correction odel estimates for short-run equilibrium but for model 1,2 and 3 can not be predicted, because their t-values show the insignificant results. On the other hand, model 4 shows the adjustment coefficient is negative and significant (-0.33) suggests that 33% of the disequilibrium is corrected immediately in next period. ckinnon emphasizes on financial repressed in developing countries because of monetary policy maintains high negative real interest rates in such countries like Chile & Brazil in 1960 s, the proportion of the time deposits in total deposits and the ratio of broad money (2) to GN is declined. This phenomenon is opposed to financial deepening which is characterized by a financial intermediation ratio, the ratio of financial assets to GN. It is valuable to analyze that whether any symptoms of financial repression exists in akistan or not. Table-7 provides informations about commercial banks time deposits as % of total deposits and the broad money 2 as % of current GN over the 1971 to 2007. Both ratios tend to show that monetary policy in akistan is not much volatile.
178 Syed Khurram Arslan Wasti, Sulaiman D. uhammad, Narjis Khatoon and Irfan Lal Table 6: Time Deposits & 2 as % of Total Deposit & GN (1970-2007) Years Commercial bank Time Deposits as % of Total Deposits 2 as % of GD 1970 53.88467 43.934 1971 48.4035 47.49 1972 45.26489 44.335 1973 46.94351 36.121 1974 44.94712 30.722 1975 48.16006 33.231 1976 44.9276 36.362 1977 46.38523 36.721 1978 47.51525 39.657 1979 47.64873 38.682 1980 47.63157 36.999 1981 49.33709 37.138 1982 49.58623 40.083 1983 52.49154 38.97 1984 53.70828 38.049 1985 51.98238 40.311 1986 51.59836 42.118 1987 48.86208 39.886 1988 48.11401 37.638 1989 42.81169 37.096 1990 38.68655 36.069 1991 37.12911 37.905 1992 41.32035 42.153 1993 54.60784 42.373 1994 54.24863 40.927 1995 55.98443 42.194 1996 62.40769 44.201 1997 52.7983 45.432 1998 55.61925 43.893 1999 53.80078 36.822 2000 56.75001 37.558 2001 56.64565 40.645 2002 56.29505 43.482 2003 51.82804 45.161 2004 50.41371 36.822 2005 41.4355 46.00 2006 54.0987 45.893 2007 42.6546 47.822 Data source: State Bank of akistan Annual Report 2008
Volatility of onetary olicy in a Developing Economy: In Context of akistan 179 V. Instruments of onetary olicy & Determinant of oney Supply The main instruments of monetary policy in akistan are the central banks credit to commercial banks, interest rate channel, margin requirement, selective credit controls and cash reserve ratios. Table 7: Trends of Banks Borrowing From SB & Interest Rates (1970-2007) Years Commercial Bank Borrowing from SB as % of Total deposit Bank Rate Weighted Average rate of Interest Deposit (%) Loans (%) Inflation Rate 1970 8.814801 5 ---- 5 4.613095 1971 8.121848 5 ---- 8.12 5.263158 1972 8.724684 6 ---- 8.65 23.10811 1973 7.725034 8 ---- 8.54 26.56422 1974 20.68354 9 4.03 9.48 20.90199 1975 22.57252 9 4.09 10.18 7.173601 1976 13.77995 9 6.22 10.92 10.17403 1977 16.68431 10 4.78 11.68 6.136087 1978 10.98719 10 5.02 11.8 8.242702 1979 13.57416 10 4.99 11.62 11.95135 1980 13.53443 10 5.03 11.09 11.8564 1981 18.24664 10 5.08 11.2 5.912162 1982 16.99561 10 4.96 10.72 6.379585 1983 14.12908 10 5.32 11.5 6.071964 1984 14.42316 10 6.05 10.83 5.618375 1985 13.56806 10 6.24 10.49 3.512881 1986 15.17402 10 6.39 10.91 4.68649 1987 15.99101 10 5.76 11 8.829886 1988 17.91457 10 8.80 10.7 7.829787 1989 18.84519 10 7.07 10.89 9.076559 1990 19.39105 10 7.67 10.59 11.77038 1991 19.58651 10 7.95 10.73 9.516616 1992 16.50065 10 8.23 12.54 9.970443 1993 15.44569 10 6.09 13.09 12.36338 1994 13.71861 15 6.17 12.78 12.35848 1995 14.37654 17 6.25 13.19 10.36049 1996 9.797463 20 6.42 13.04 11.38117 1997 9.080548 18 6.8 12.76 6.2233 1998 9.791508 16.5 6.81 15.64 4.152174 1999 12.45858 13 6.49 14.47 4.362346 2000 13.37966 13 5.47 13.25 3.15 2001 11.68122 10 5.27 13.07 3.286476 2002 12.04348 7.5 3.61 13 2.919091 2003 11.19075 7.5 1.61 11.87 3.143 2004 9.675067 7.5 0.95 8.41 4.635 2005 19.39105 10 8.34 13.09 9.326 2006 19.58651 15 6.69 14.37 7.946 2007 16.50065 15 6.54 13.08 7.866 Data source: State Bank of akistan Annual Report 2008 Table-8 exhibits the starting era of commercial banks borrowing from SB shows an increasing trend but with the passage of time it seems to be declined caused by an increasing trend in bank rates. Table also shows that inflation rate is quite volatile even now.
180 Syed Khurram Arslan Wasti, Sulaiman D. uhammad, Narjis Khatoon and Irfan Lal Table 8: oney ultiplier & Their Determinant in akistan Years Currency Outside Bank as % Demand Deposits Bank Reserve as % of Demand Deposits Narrow oney (1) ultiplier 1970 0 164.9552 ---- 1971 0 138.893 ---- 1972 0 131.3312 ---- 1973 19.95674 130.9136 ---- 1974 19.04143 131.4867 ----- 1975 5.111811 116.9277 ---- 1976 11.93696 94.05252 ---- 1977 13.31777 102.7347 ---- 1978 10.76602 102.3023 ---- 1979 18.45162 111.9499 ---- 1980 11.82255 116.1987 ---- 1981 20.36721 113.8152 ---- 1982 6.353936 109.2415 ---- 1983 15.11976 105.6524 ---- 1984 11.82606 124.4486 ---- 1985 6.923415 112.5996 ---- 1986 9.382492 118.0104 ---- 1987 12.64004 113.5104 3.76 1988 13.54763 117.7811 3.12 1989 8.766647 120.0498 3.33 1990 9.890955 121.1832 3.36 1991 13.98575 125.523 3.10 1992 134.0132 104.9908 3.22 1993 150.1547 127.4313 3.26 1994 78.51062 120.0948 3.40 1995 85.14548 131.6041 3.36 1996 86.72443 118.9337 3.21 1997 58.01513 87.22955 3.28 1998 64.50669 97.99948 3.19 1999 64.23394 104.5928 3.04 2000 77.69724 100.2337 3.10 2001 71.5912 112.251 3.08 2002 69.02062 101.9097 3.17 2003 60.56917 91.41831 3.17 2004 56.31374 110.133 3.29 2005 57.06573 97.8933 3.32 2006 55.22076 95.7598 3.25 2007 55.74206 92.22 3.36 ean 38.51 113.638 3.255 Coefficient of Variation in % Data source: State Bank of akistan Annual Report 2008 40.1875 13.6143 4.7876 The supply of money is the product of reserve money (currency and bank reserve with the SB). Table 9 shows the behaviors of these monetary variables in akistan. The narrow money multiplier does not seems to be fluctuated. The growth of money supply can be expressed in term of growth of reserve money (R) and the money multiplier. Table 10 represents the growth rates of monetary variables in term of percentage.
Volatility of onetary olicy in a Developing Economy: In Context of akistan 181 Table 9: Growth Rates of onetary Variables (%) Years g1 g2 gr 1970 ---- ---- ---- 1971 5.719587 13.77114 10.87487 1972 9.462495 17.46852 17.41242 1973 20.16881 13.80245 11.96103 1974-7.5-1.20366 0.68249 1975 40.51861 21.21492 9.816848 1976 30.77321 32.19702 19.78846 1977 19.1257 17.94754 22.69764 1978 23.54777 19.84247 17.57054 1979 16.43675 19.097 26.73715 1980 12.32117 15.69721 16.66438 1981 18.27753 11.52509 8.203897 1982 23.83159 21.83014 17.67667 1983 32.13014 20.91022 13.75943 1984 3.737297 4.647177 16.36589 1985 12.01416 14.7467 8.618098 1986 12.57396 16.07261 19.81227 1987 11.32635 16.40355 19.47032 1988 3.661963 7.70328 10.83251 1989-7.2765 7.400726 17.06892 1990-3.19894 11.64357 15.93439 1991 15.05919 18.94439 27.33315 1992 55.07576 29.30057 8.782828 1993 60.74293 18.13347 14.19812 1994 21.01599 17.3706 15.7125 1995 15.442 13.8098 17.93226 1996 39.15916 20.06856-3.64608 1997 5.037443 19.91231 14.3351 1998 12.64257 7.858833 12.95215 1999-1.62401 4.314387 12.99095 2000 15.1534 12.12538-2.05938 2001 14.07497 11.74594 28.29595 2002 18.15325 16.83867 8.809075 2003 8.519229 17.54003 16.54495 2004 18.79626 20.51444-91.1543 2005 55.0378 8.4051 8.6180 2006 13.1294 15.06451 19.8127 2007 90.66084 19.3330-9.5887 Data source: State Bank of akistan Annual Report 2008 Where g1: Represent the Growth Rate of oney Supply (1) g2: Represent the Growth Rate of Broad oney Supply (2) gr: Represent the Growth Rate of Reserve oney. Table 10 represents that the reserve money is the major determinant of money supply. This is because of stock of reserve money may be passively responsive to the demand for money and credit, especially for the financial requirements of the government. The level of budget deficit in akistan seems to be statistically significant variable in explaining the variation in the stock of reserve money.
182 Syed Khurram Arslan Wasti, Sulaiman D. uhammad, Narjis Khatoon and Irfan Lal Table 10: Reserve oney 1970-2007, Empirical Evidence Equations G Dependent Variable R Constant -14892.72 (1.32) BD 0.058 (-2.55) Regression Statistics R^2 DW 0.99 1.95 Where R: Represent Reserve oney. BD: Represent Budget Deficit. VI. Results & Discussions This study examines the different mechanism of monetary policy in akistan. The demand for money function formulated according to the ckinnon model, instruments of monetary policy and determinants by employing Johnson Cointegration test for the long run equilibrium relationship among variables using annual data for the period 1971-2007. For estimating short run equilibrium the Error Correction odel is applied. The results show that there is a long-run equilibrium relationship among variables, but we didn t able to predict short run equilibrium in first three models except the last model which shows that real rate of interest and real private investment significantly effect the real money demand in the long run. The empirical results show that time deposits and broad money have a less volatile behavior in case of akistan economy and also explain that stock of reserve money causes a large portion of variation in money supply. However, results also suggest that reserve money itself influenced by such factors as the level of budget deficit. olicy Implication & Suggestions The evidence shows that stock of reserve money explains the less variation in money supply in case of akistan. Further shows that reserve money itself is influenced by level of budget deficit. In term of policy implication it is suggested that in order to gain fruitful results from monetary policy in akistan, the policy makers should also consider fiscal deficit in their policies because this study has clearly shown that stock of reserve money has significantly affected the money supply in the country that stock of reserve money itself effected by budget deficit. In order to control money supply it is suggested that government should reduces its expenditures.
Volatility of onetary olicy in a Developing Economy: In Context of akistan 183 References [1] Aghevi, B, B. et al (1979) onetary olicy in Selected Asian Countries, IF staff papers 26: 775-824. [2] Ahmed. I & Rehman, H. Impact of onetary olicy on Assets in akistan. akistan Economic & Social Review. Volume XL, No. 2 (Winter 2002), pp, 199-208. [3] Economic Survey of akistan 2008-2009. [4] Galbiz, V. (1979) oney Investment & Growth in Latin America, 1961-1973, Economic Development & Cultural Change 15: 257-268. [5] Gregorio, J, D & Guidotti,, E. (1995). Financial Development & Economic Growth. World Development, Vol.23, No. 3, pp. 433-448. [6] International Financial Statistics Data set 2008-09. [7] Hussein, K, A & Demetraide,, O. (1996). Does Financial Development Cause Economic Growth? Time Series Evidence from 16 Countries. Journal of Economic Development. Vol, 51, pp, 387-411. [8] Islam, S. (1984). onterey olicy in Developing Economy: A case for Bangladesh. The Indian Economic Journal. Vol 35. No.2. [9] Jung, W, S. (1986). Financial Development & Economic Growth. Washington Institution. [10] Khalid, A,. (2005). Economic Growth, Inflation, & onetary olicy in akistan. The akistan Development Review 44: 4 art II (Winter 2005).pp, 961-974. [11] ckinnon, R. (1973). oney & Capital in Economic Development. Washington Institution. [12] oore, B, J. (1979). The Endogenous oney Stock, Journal of ost Keynesian Economics 2: 49-70. [13] World Bank Development Indicator data set 2008-09. [14] akistan State Bank of akistan Annual Reports 2008-09. [15] Qayyum, A. (Winter 2002). onetary Condition Index: A Composite easure of onetary olicy in akistan. The akistan Development Review.41: 4 art II. pp 551-566. [16] Shaw, E, S. (1973). Financial Deepening in Economic Development. New York: Oxford University ress. [17] Townsend, R,. (1983). Financial Structure & Economic Activity. American Economic Review 73: 895-911. [18] Trehan, B & Lansing, K, J.Forward-Looking Behavior & Optimal Discretionary onetary olicy. Economics Letters81 (2003). p 249-256.