Bekov, Dastan2024-12-192024-12-192024-12-12Bekov, D. (2024). The Predictive Power of Extremes in Inflation Expectations: A Study Of Kazakhstan's Inflation Anchoring. Nazarbayev University Graduate School of Businesshttps://nur.nu.edu.kz/handle/123456789/8349This thesis examines the predictive power of extreme inflation expectations on actual inflation in Kazakhstan. Using National Bank of Kazakhstan (NBK) survey data and key macroeconomic indicators from 2016 to 2024, the study assesses whether extreme expectations provide better forecasting accuracy than median expectations. A vector autoregressive (VAR) framework is used to model inflation using median, extreme, and combined values of inflation expectations. Out-of-sample forecasts show that extreme values outperform median values in predicting actual inflation, highlighting their relevance during periods of increased inflation volatility. The results highlight the limitations of relying solely on average expectations and underscore the value of including extreme values for a more nuanced perception of future inflation. By demonstrating the importance of extreme values in inflation expectations household survey, this study offers actionable insights for improving the NBK’s inflation forecasting methods and enhancing the effectiveness of inflation targeting policy in Kazakhstan.enAttribution-NonCommercial-ShareAlike 3.0 United StatesInflation expectationsExtreme valuesinflation anchoringType of access: EmbargoMaster`s thesis