The central bank pursuing price stability should endeavor to find appropriate measures and tools for identifying the exact causes of changes in and stabilizing of inflation expectations. In this regard, the central bank should continuously strive to understand the formation process and determinants of inflation expectations. To shed light on the formation of inflation expectations in Korea, we conduct micro-level analyses employing the microdata of the Consumer Survey, as well as macro-level analyses using the common aggregated measure of inflation expectations. Furthermore, via comparison with some advanced economies, we attempt to discern the commonalities and peculiarities of Korean inflation expectations.

 The macro-level analyses reveal that actual inflation and inflation perceptions, i.e., the perceived level of past inflation by consumers, mainly influence inflation expectations in Korea, which implies that inflation expectations are formed in a backward-looking way. In particular, considering the fact  expectations and perceptions of inflation are nearly the same, it can be inferred that subjective perceptions of actual inflation exert stronger influence on inflation expectations compared to actual inflation information. Moreover, inflation expectations are formed depending heavily on the price information about a narrow set of consumer goods that are frequently purchased or whose expenditure shares are higher. Turning to macroeconomic indicators, whereas price  information, such as the Consumer Price Index and wages, is estimated to be utilized in forming inflation expectations, information on economic and financial conditions, including industrial production and interest rates, is not. In the advanced economies such as the United States, the euro area, the UK, and Japan, inflation expectations are formed similarly but in a more forward-looking way than in Korea.
 
 In the micro-level analyses, it turns out that the expected level of inflation varies depending on the Consumer Survey respondents’ individual characteristics,  including age, sex, income, and economic sentiments. However, there exists a crucial distinction between Korea and the advanced economies: in Korea, the inflation expectations of respondents who are younger or earn larger income  exceed those of respondents who are older or earn less. Turning to economic sentiments, a respondent having a pessimistic economic outlook tends to form inflation expectations at a higher level relative to an optimistic one, which suggests the possibility that the discrepancy between inflation expectations and actual inflation could regularly widen in a low inflation environment induced by economic slowdown as the number of pessimistic respondents grows.

 The main finding of this paper that inflation expectations in Korea are formed in a backward-looking way while depending mainly on a confined range of information implies that the most effective measure for anchoring inflation expectations is to stabilize actual inflation. Additionally, communicating to the public a firm intention to pursue price stability alongside the monetary policy stance can contribute to anchoring inflation expectations to some extent.