Measuring Monetary Policy Uncertainty and Its Effects on the Economy: The Case of Turkey

ABSTRACT

In this study, we present new empirical insights for monetary policy uncertainty perception in Turkey by introducing a set of new measures. We construct a news-based measure based on the frequency counts of articles in newspapers that contain specific terms related to monetary policy uncertainty and two survey-based measures based on the interest rate expectations. Using these measures, we investigate the impact of monetary policy uncertainty on economic activity using vector autoregression models. Our findings indicate that monetary policy uncertainty remained relatively high and volatile during the implementation of unconventional monetary policy. The impulse response analysis suggests that monetary policy uncertainty shocks have negative effects on economic activity in Turkey.

KEYWORDS:

• monetary policy
• Turkey
• uncertainty

JEL Classification:

• D80
• E52

Disclosure Statement

No potential conflict of interest was reported by the authors.

Notes

1. The Reserve Options Mechanism (ROM) is a mechanism that allows banks and financial institutions to keep a certain ratio of their Turkish lira reserve requirements in foreign currency and/or gold. It is designed to limit the adverse effects of excess volatility of capital flows on financial stability.

2. Theoretically, the effects of a monetary policy uncertainty shock would be different from those of a monetary policy shock. A positive monetary policy shock would lead to an appreciation of the domestic currency and a fall in the inflation rate, contrary to a positive MPU shock which would lead to a depreciation and an increase in the inflation rate, while both would dampen economic growth.

3. Kuttner (2001) separates the “surprise” element of Fed policy actions from Fed futures contracts and shows that the unanticipated part of the policy rate changes affects market rates significantly, while the anticipated part has essentially no effect.

4. There is also a wide empirical literature on the impact of overall economic uncertainty on markets and macroeconomic fluctuations. See Bachmann, Elstner, and Sims (2013), Bloom (2009, 2014), Castelnuovo, Lim, and Pellegrino (2017), Jurado, Ludvigson, and Ng (2015) among others.

5. Regarding the effects of monetary policy uncertainty on markets, Kurov and Stan (2018) examine whether monetary policy uncertainty influences markets’ reaction to macroeconomic news in the U.S., using intraday futures data. They show that the response to macroeconomic news weakens in the stock and crude oil markets, while it strengthens in the Treasury, interest rate and foreign exchange markets when policy uncertainty is high.

6. Data on implied volatility of interest rate options is not available since there is no functioning market for this derivative.

7. More precisely, Money Market Rate Deviation Index, calculated as the difference between the CBRT’s average funding rate and Borsa Istanbul Repo and Reverse Repo Market overnight interest rate (See Talasli 2018).

8. The Herfindahl-Hirschmann Index of CBRT Funding is a measure of concentration of the channels that the CBRT uses to fund the market. A smaller index value indicates that the CBRT provides its funding through various channels, which implies higher uncertainty in liquidity management. A higher index value suggests a higher concentration of funding channels and thus lower uncertainty. If the CBRT provides all of its funding through one channel, the index value would be equal to one, implying no uncertainty associated to liquidity management.

9. Their results suggest that while a policy tightening flattens the yield curve for all uncertainty levels, it is more pronounced when monetary policy uncertainty is low; and that (contrary to expectations), positive policy surprise decreases long-term rates via anchoring inflation expectations.

10. We pay particular attention to take a balanced number of newspapers with different political tendencies, so that we can obtain an average perception of uncertainty as neutral as possible.

11. Interpress is a media monitoring company in Turkey.

12. We also conduct a detailed “human audit” to assess the relevance of selected articles, to make sure they discuss issues related to monetary policy uncertainty in Turkey.

13. The import content of production and the exchange-rate pass through are considerably high in Turkey, which makes fluctuations in the exchange rate more of a concern in terms of inflation. The volatility of the exchange rate is also a huge concern for financial stability in emerging markets due to high dependency on external financing.

14. The standard deviation of the scaled news for each newspaper is calculated for the period 1998:01–2018:12.

15. Banking supervision and regulation were extensively strengthened after the crisis.

16. The only available forecast horizon is the end of the current month.

17. We also examine the mean forecast error of overnight interest rate expectations. With this measure, we can distinguish between upside and downside uncertainty. The results suggest that forecast errors are generally negative when the interest rate is increasing, and positive when the interest target rate is decreasing. This finding supports the idea that the interest rate is higher than expected during tightening periods, and lower than expected during easing periods. Besides, it is also worth noting that upside uncertainty is typically higher in magnitude relative to downside, implying that participants are less certain during easing periods than they are during tightening periods.

18. The contemporaneous cross-correlations between the news-based and the survey-based measures are quite low; 0.12 for the MSFE and 0.11 for the Disagreement.

19. Volatility of bond yields is defined as the monthly standard deviation of daily rates.

20. Implied USD/TL exchange-rate volatility is taken from Bloomberg.

21. Calculated as the monthly standard deviation of differences.

22. The real exchange rate is the CPI-based real effective exchange-rate index, obtained from CBRT. An increase in the index implies an appreciation in the Turkish lira.

23. The real interest rate is calculated by using the 1-year bond rate and the 1-year ahead annual CPI inflation expectation from the CBRT Survey of Expectations, both obtained from the CBRT.

24. GDP data is obtained from the Turkish Statistical Institute (TurkStat).

25. We consider the fluctuations in the real exchange rate as the most exogenous variable, since global financial conditions and other domestic factors (besides monetary policy actions) can play important roles in exchange-rate developments in emerging markets. Still, changing the ordering of the variables in the VAR specification does not lead to significant change in results. Moreover, as a robustness check, we estimate an alternative VAR model using the variables in Husted, Rogers, and Sun (2017). The ordering of the variables in this VAR model are as follows: the industrial production (IP), the consumer price index, the MPU index, the government bond rate and the credit spread (EMBI Turkey). By this ordering, we assume that the innovations in the interest rate and credit spread do not affect MPU contemporaneously. The impulse response of IP to a one standard deviation shock of MPU shows that IP reacts negatively to the uncertainty shock, with a peak impact reached after three quarters at around −2.3%, and the response is significantly negative for up to six quarters.

26. The CBRT implements inflation targeting since 2006.

27. A one standard deviation shock to MPU is about 1/5 of the change in uncertainty in June 2006.

28. The impulse responses of the real exchange rate and real interest rate to a one standard deviation shock of (news-based) MPU in the VAR framework shows that the Turkish lira depreciates in real terms and the real interest rate rises, though the results are not statistically significant.