Inflation Targeting, Flexible Exchange Rates and Inflation Convergence
One sentence summary: After adopting inflation-targeting, Turkish regional inflation rates have diverged from each other in terms of CPI groups with relatively tradable components, while they have converged to each other in terms of CPI groups with relatively non-tradable components.
The corresponding paper by Hakan Yilmazkuday has been published at Applied Economics.
Abstract
Using a disaggregated-level CPI data, this paper compares convergence properties of regional inflation rates in a small-open economy, Turkey, between pre-inflation-targeting and inflation-targeting periods, where the latter also corresponds to a flexible exchange rate regime. It is found that during the inflation-targeting period, Turkish regional inflation rates have converged to each other in terms of CPI groups with relatively non-tradable components, while they have diverged from each other in terms of CPI groups with relatively tradable components. Since a common and sound monetary policy among the regions of a country is supposed to have a convergence effect on regional inflation rates according to the conventional wisdom, the results for CPI groups with relatively tradable components require further attention and have significant policy implications, especially for inflation-targeting countries using flexible exchange rate regimes.
Non-technical Summary
There is a large body of evidence of the convergence of inflation rates among the countries under a common monetary policy and currency; however, there is less evidence on the convergence of regional inflation rates within a country. Such an analysis is important for policy makers, because persistent differences in (actual and expected) inflation among regions of a country may lead to disparities in regional real interest rates, given a common national monetary policy. These diversities may be exacerbated by cyclical considerations: A region where economic activity is relatively subdued is likely to have weak inflationary pressures and therefore experience a relatively high real interest rate; this, in turn, could further add to the divergence of inflation. On the other hand, sharing a common national exchange rate, inflation differentials may work as an adjustment mechanism: Regions with higher productivity or lower wage growth than others would experience a depreciation of the real exchange rates (i.e., a fall in relative prices) and thus a gain in production (and trade) competitiveness. Overall, whether the expansionary effects associated with a real-interest-rate reduction or the contractionary ones induced by real-exchange-rate appreciation, due to a positive inflation differential, would dominate, and the horizon at which this might happen is an empirical question. The answer will depend to a large extent on the magnitude of inflation differentials and on their persistence. However, part of the differences in inflation could also be due to regional heterogeneities in the relative productivity growth of the tradable versus the non-tradable sector (the so-called Balassa-Samuelson effect), and therefore they might last as long as these persist.
Within this context, we investigate whether there is a role for national monetary policy in explaining the convergence of regional inflation rates within a country. Although the convergence of regional inflation rates may be due to several external factors (e.g., the international great moderation), in addition to the discussion above, this paper investigates how national monetary policy may influence the convergence of regional inflation rates through the following mechanisms:
- Decreasing (respectively, increasing) inflation rates may lead to less (respectively, more) frequent price changes due to decreasing (respectively, increasing) opportunity or menu costs; depending on the degree of region-specific frequency of price changes, this mechanism may lead to diverging (respectively, converging) inflation rates during low (respectively, high) inflationary episodes.
- Depending on the degree of region-specific exchange-rate pass-through, which is affected by the contents of region-specific consumption baskets at the individual-good level (i.e., traded versus non-traded goods), a flexible (respectively, fixed) exchange rate regime may lead to higher volatilities in the prices of traded goods; this, in turn, affects the distribution and convergence of prices across regions.
- According to the Balassa-Samuelson effect, faster (respectively, slower) growing regions experience a rising relative price of non-traded (respectively, traded) goods and a real appreciation over time; thus, regions with similar (respectively, different) growth rates may experience convergent (respectively, divergent) prices of non-traded goods and divergent (respectively, convergent) prices of traded goods. Since the stability of growth in an economy is highly related to the effectiveness of monetary policy, convergence of regional prices can also be affected by the national monetary policy.
The investigation is achieved through comparing the convergence properties of inflation rates of CPI groups (i.e., combinations of good categories) among geographical regions of Turkey between the pre-inflation-targeting and inflation-targeting periods. The motivation (i.e., the selection of Turkey as the focus of investigation) mainly comes from two sources: (i) rapidly decreasing inflation rates (i.e., a successful monetary policy) in the inflation-targeting period, (ii) the estimated structural breaks in the first and second moments of inflation within Turkey. Specifically, the structural break analysis estimates that the national inflation rate has a break right after the beginning of inflation-targeting regime in January 2002; and the cross-sectional standard deviation of regional inflation rates has a break right before the financial crisis in February 2001, after which Turkey adopted a flexible exchange rate and started relevant reforms in the economy to begin conducting inflation targeting. These estimated break dates are used to analyze the possible effects of an inflation-targeting regime, together with a flexible exchange rate regime, on the convergence regional inflation rates.
It is found that the CPI groups with relatively tradable components have diverged from each other, while the CPI groups with relatively non-tradable components have converged to each other, during the inflation-targeting period. Since a common and sound monetary policy (during the inflation-targeting period) among the regions of a country is supposed to have a convergence effect on the individual inflation rates, the results, especially for CPI groups with relatively traded products, are surprising and require further attention. One explanation can be through the effects of the frequency of price changes on relatively traded products; i.e., decreasing inflation rates have led to less frequent price changes due to decreasing opportunity/menu costs, which, in turn, have led to diverging inflation rates during low inflationary episodes, depending on the degree of region-specific frequency of price changes. Since the inflation-targeting period also corresponds to a flexible exchange rate regime in Turkey, an alternative approach may suggest that the traded good shares differ across Turkish regions, so that a high volatility in the exchange rates (due to the flexible exchange rate regime) is reflected in the regional inflation rate differences, depending on the degree of region-specific exchange-rate pass-through. Finally, according to the Balassa-Samuelson effect, the results imply that the inflation-targeting period coincides with a productivity growth of non-tradable goods (e.g., services) and a productivity fall in tradable goods. Since improvements in the size or productivity of the service sector have recently become an indicator of development, the results may also correspond to clues regarding the development of Turkey during the inflation-targeting period.
The results of this paper have important policy implications, which may be related to either the inflation-targeting process or the flexible exchange rate regime: (i) If CPI groups in different regions are affected differently by monetary policy conducted under inflation-targeting regime (i.e., different regions or sectors are affected differently by the policy instruments such as money aggregates, exchange rate, and interbank rates), then inflation-targeting regime may be the reason for this result; (ii) alternatively, if CPI groups in different regions have different imported good shares, the volatile exchange rate (due to the flexible exchange rate regime) may be a potential reason for the results of this paper. In particular, the Central Bank of Turkey implemented its monetary policy with both interest rates and foreign exchanges until the end of 1999, with foreign exchanges in year 2000, and with short-term interest rates since 2001. These different monetary policy tools may affect prices of different products differently; e.g., exchange rates affect prices of tradable goods more than non-tradable goods. If each geographic region has a different weight on these two set of prices, it is plausible that the flexible exchange rate regime also has its influence on the convergence of regional inflation rates. Understanding these linkages across regions and sectors is the key to a thorough monetary policy, at both national and regional levels.