Özet:

Purpose- The aim of this study is to calculate the coefficient parameters of the factors affecting the pricing in the low, average, and overpricing intervals and points of the US Global dollar index and then to investigate the dynamic historical effects of these parameters. Methodology in the study, "Quantile Regression" to calculate parameter differences in pricing intervals and "Kalman Filtering" methods to calculate their historical dynamic effects. Findings- In the design of the study, The intervals are 0.5-0.95 incremental overpricing, 0.5 (Median) average, and 0.05-0.5 (Median) underpricing (low-pricing). The study's results show that model 0.4. quantile has respectively the highest value R2 and adjusted R2 values, approximately 53.2 percent and 49.1 percent. Additionally, the probability value of all 19 estimated models is statistically significant at the 0.05 level. While the coefficients of the Baltic Dry Index (at 5%), the Global gold prices (at 5%), and the US 10-year bond yields (10%) are negative, the coefficients of the Nasdaq (10%) and Vix (5% and 10%) have positive signs. These variables are significant in the underpricing quantiles that conducted interval of the US Dollar index (0.05-0.5) in the research design. The price point that represents the median value yields the same results. From that point of view, only the Vix index is significant and only at a 10% level. The Baltic Dry Index (5%), Bitcoin and Gold prices (5% and 10%), US 10-year interest rates-yields (5%), and CDS premium are among the factors that are relevant in the highquantile overpricing range of the US Dollar index (0.05-0.5). (5 percent and 10 percent ), Although the variables' coefficients are negative, the coefficients for inflation (at 5% and 10%), Nasdaq (at 5%), and the VIX index (10%) are positive. The dynamic coefficients determined historically and dynamically using the Kalman filtering technique in all quantiles have had the same values. Conclusion- Since Kalman analysis and quantile regression analysis have different theoretical background, parameter differences in underpricing and overpricing periods may be eliminated when historical dynamics are examined. As a result, even though the findings of quantile regression and the results of Kalman analysis were roughly parallel, the predicted parameters for some variables did not closely match the effects of either technique. The literature has noted that research utilizing both methodologies might run into such statements that can be encountered in the study’s findings under comparable circumstances (Bernardi v., 2016: 34). Additionally, as the geopolitical risk index conveys countercyclical hazards, the rising geopolitical risks in the historical coefficients raised the US dollar index, according to Kalman's study.

Anahtar Kelimeler: