Comment on “Flexible Inflation Targeting and Macroeconomic Performance: Evidence from ASEAN”

Abstract

My research interests include: (i) why did Thailand, the Philippines, and Indonesia adopt flexible inflation targeting (FIT)?; (ii) how do the institutional details of FIT differ among the three countries?; and (iii) how do we evaluate the performance of the inflation targeting? Has the macroeconomic performance improved due to the adoption of FIT?

Nookhwun and Waiyawatjakorn (2024)—N&W hereafter—is a welcome addition to the literature on FIT in Asia and answer most of my research questions. It provides a comprehensive survey of the monetary policy framework of the three FIT adopters and two non-adopters among the original five ASEAN countries. N&W (table 1) summarizes the monetary policy framework including objectives, FIT or not, policy instruments, and communication strategy. N&W (figure 2) is a comprehensive time-series (2000–2023) summary of actual inflation rates, inflation targets, and inflation expectations.

The Thailand case (N&W (figure 2(a))), which I am most familiar with, is interesting. Thailand adopted FIT in 2000 with a 0–3.5% range of the core inflation rate. It is rather a wide range compared to other FIT countries' ranges. Targeting core inflation, which is less volatile, seems to be easier than targeting headline inflation. Why did the Bank of Thailand set up a “low bar” for itself?

The answer for adopting FIT with a “low bar” was given to me by the person who was responsible for the adoption of FIT in Thailand. He said that it would be important to keep the inflation rate in the range to restore the credibility of the Bank of Thailand. So the low bar was intentional. He added that the range can be narrowed once credibility is established by hitting the target for some time. The reputation of the Bank of Thailand had been tarnished in the financial crisis of 1997–1998, by allowing foreign reserves to be used in futile efforts to defend the fixed exchange rate.

After the adoption of FIT in Thailand in 2000, the core inflation rate would stay for more than 90% of the time in the 0.0–3.5 range. The range was narrowed to 0.5 to 3.0 in 2009. This showed the confidence gained by the Bank. Grenville and Ito (2010) recommended that the indicator inflation should be headline inflation instead of core inflation because headline inflation is more similar to the cost of living of the ordinary people. They also recommended that a point target with a tolerance band would be better than the range because it would have an expectation anchoring effect. These recommendations were adopted in 2015, with a point target of 2.5% plus/minus 1.5%. But the timing turned out to be extremely poor in retrospect. Energy prices fell sharply in 2015 and the global inflation rate came down. The inflation rate in Thailand undershot the lower bound of the tolerance band (namely, below 1.0%) for almost the entire period from 2015 to 2020. The new framework adopted in 2020 re-specified the range to be 1 to 3%. Headline inflation became very volatile from 2020 – first becoming negative due to COVID and then showing very high inflation in 2022. Inflation in Thailand was strongly influenced by global factors from 2015 on.

In evaluating the performance of inflation targeting, a researcher would like to construct a counterfactual economy, that is, the same economy without FIT. The comparative evaluation of FIT should be done against the counterfactual. But it would be very difficult to construct such a counterfactual.

An alternative approach is (a) a time-series comparison, namely a comparison of the economic performances before FIT adoption and after FIT adoption in the same country. This is done in N&W (figure 3). However, the time periods with and without FIT are subject to different global shocks and different domestic shocks, so the time-series comparison is not perfect. Another problem is causality. A very high inflation, regardless of the reason for it, may prompt the government (or the International Monetary Fund to give advice) to adopt the FIT. The correction of the high inflation rate would have happened anyway. This was the case in Indonesia where the inflation rate came down to a much lower level after the FIT was adopted, as noted in N&W (section 3.1).

Another alternative is (b) a cross-section approach. A comparison of similar countries with FIT and without FIT in the same time periods. This is done in N&W (section 3.4) where a difference-in-difference approach is adopted and a comparison of FIT adopters and non-adopters for a large number of countries (52 developing countries, of which 30 are FIT adopters and 22 are non-adopters) is shown. Global factors are controlled, but of course, there is a limit to how similar the two countries are. N&W (table 3) shows the results using three different estimation methods. N&W concludes that their “findings suggest that IT benefits adopters mainly in terms of lower inflation.” Although the results were in favor of FIT adoption, the difference was not decisive.