TY - JOUR AU - Hoa Le AU - Uyen Vo AU - Dat Nguyen AU - Uyen Pham PY - 2022/09/30 Y2 - 2024/03/29 TI - Quantile regression with time-series data and applications in modeling of some big bank stock markets JF - Science & Technology Development Journal: Economics- Law & Management JA - STDJELM VL - 6 IS - 3 SE - Research article DO - https://doi.org/10.32508/stdjelm.v6i3.948 UR - http://stdjelm.scienceandtechnology.com.vn/index.php/stdjelm/article/view/948 AB - In the OLS regression model, the mean of the dependent variable is estimated based on the mean of the independent variables. The relationship between the independent variable and the dependent variable needs to be considered in many values instead of just through the mean of the dependent variable, then the quantile regression model is the optimal choice. In this paper, we study a quantile regression model in which the percentiles of the dependent variable are spread from 10% to 90%, with step by 10%, of the time-series data through autoregression model, to compare the fit of the models as well as the errors of the model. To demonstrate the results, we applied it to the time-series data on the closing stock prices of the four largest bank codes (on August 2, 2021) namely VCB, VPB, TCB, and BID. In which, percentile regression models have a very high fit, corresponding to about 80% or more. Besides, the estimated parameters are all statistically significant, different from the OLS regression model where some estimated parameters are not statistically significant. Furthermore, the OLS regression model is based on the mean, so it is susceptible to the outlier values, while the quantile is not affected by the outliers. In other words, the quantile regression model overcomes this weakness, so the quantile regression model is not affected by the outliers. ER -