Financial crisis, monetary policy reform and the monetary transmission mechanism in Turkey

Abstract

Turkey experienced a financial crisis in 2000–2001, which led to significant financial reforms. The reforms resulted in a switch to a floating exchange rate, granted greater central bank independence and pursuit of a more credible monetary policy. Investigation of the channels of monetary policy in both periods finds that monetary policy's output effects have been strengthened considerably by the reforms. In the pre-crisis period, monetary policy was highly inflationary, while in the post-crisis period, monetary policy targets low inflation and has become a tool for output stabilization. These results support the importance of central bank independence and a credible policy.

Keywords:

• financial crisis
• monetary policy
• monetary transmission mechanism
• Turkey

Notes

1. Research has established that policy reforms have important economic impacts. Lucas (1976) established that a regime switch changes behavior within an economy. Barro and Gordon (1983) emphasize that a regime change solves the time consistency problem (Kydland and Prescott, 1977). Alesina and Summers (1993) demonstrate that central bank independence results in less inflationary monetary policy.

2. The contraction is directly proportional to the interest elasticity of money demand (Bean et al., 2002).

3. The exchange rate was not publicized, but was known in advance by high level bureaucrats who sold the information to currency traders including bank owners. Currency traders bought the currency from the central bank when it was cheap, and sold it back when it was expensive.

4. Sahinbeyoglu (2001), using a small aggregate macroeconomic model for the period 1987 to 1999, concludes that the monetary transmission mechanism is weak when fiscal and monetary policies are uncoordinated. Ozturkler (2002) estimates a VAR for the period 1986 through 2001 and finds a strong interest rate channel but a weak bank lending channel.

5. Basci et al. (2007), based on studies conducted by the central bank's staff, report that the interest rate and credit channels have strengthened, while the exchange rate channel has weakened in the recent years. Cifter and Ozun (2007) use a vector-error-correction-model based causality test to analyze the data from 1997 through 2006. They find that the interest rate channel is relatively stronger in Turkey. This study does not investigate the possibility of a structural break in 2001.

6. Aslan and Korap (2007) analyze the monetary transmission mechanism in Turkey for the 1992–2004 period using Vector Autoregressions, while ignoring the possibility of a structural break.

7. Empirical studies examining the MTM before and after crisis periods in emerging economies similar to Turkey find that post-crisis reforms have changed the MTM in these economies. These studies examine the MTM in Mexico (Sidaoui & Ramos-Francia, 2008), Chile (Betancour, De Gregorio, & Medina, 2008), and Argentina (Gomez-Gonzalez & Grosz, 2007).

8. The data are from the International Financial Statistics Database of the International Monetary Fund.

9. Historically money developments have played an important role in Turkish monetary policy strategies.

10. Fung (2002) and Hesse (2007) conducted their analyses using the sub-sample data: pre-crisis and post-crisis.

11. The Likelihood Ratio Test is used to test whether the error covariance matrix is diagonal for the baseline model for both pre- and post-crisis periods. The test fails to reject the null hypothesis that the error covariance matrix is diagonal at the 5% level of significance for both periods. If the error covariance matrix is diagonal, which is the case for the baseline model, the estimated impulse response functions are not sensitive to the ordering of the endogenous variables in the VAR model.

12. Test results are available upon request. Both the Augmented Dickey-Fuller (ADF) test and the Augmented Dickey-Fuller-GLS test based on the Schwartz-Bayesian Information Criteria are used to test for unit roots. The 5% level of significance is used for all tests.

13. Test results are available upon request.

14. Test results are available upon request.

15. Since the VAR in levels includes non-stationary variables, the confidence intervals with impulse responses would be incorrect. Confidence intervals are not reported with the impulse responses in these graphs.

16. In Figure 1, the horizontal axis denotes the number of months elapsed after the shock, and the vertical axis denotes the deviation from the baseline level of the target variable in response to a one time unit shock to the policy variable (INTRATE).

17. Another method of analysis would be to include all four channel variables simultaneously, but data limitations preclude this approach.

18. The likelihood ratio test is used to assess the robustness of the results. The null hypothesis of an exogenous channel is tested against the alternative hypothesis of an endogenous channel. If the null is not rejected, the null of no effect of a channel (the baseline model) is tested against the alternative of a non-zero effect.

19. Data are from the International Financial Statistics database.

20. The lag length tests are discussed above.

21. Likelihood ratio tests indicate an exogenous exchange rate channel in the pre-crisis period and an endogenous channel in the post-crisis period.

22. Basci et al. (2007) also find that the supply side effect is dominant.

23. The data are from the Central Bank of Turkey.

24. Likelihood ratio tests indicate the asset price channel works exogenously in both periods.

25. Data are from the Central Bank of Turkey.

26. Likelihood ratio tests support the existence of an exogenous credit channel in the pre-crisis period and an endogenous channel following the crisis.

27. Deposit rate data are from the Central Bank of Turkey.

28. Ca'Zorzi, Hahn, and Shanchez (2007); Kara et al. (2005); Leigh and Rossi (2002) all find that pass through decreases with the adoption of floating exchange rates.

29. Investigation of the effects of monetary policy since the adoption of explicit inflation targeting in 2006 is a topic for future research.