This study estimated the consumer price index bias using the Household Income and Expenditure Survey(MDIS) from 1990 to 2024 and Hamilton (2001). Hamilton (2001) suggested a method that can be used when Engel’s law is valid, and this study used the household price index to estimate this model. The main results are as follows. First, the Engel coefficient, defined as food expenditures compared to total expenditures, rapidly decreased from 1990 to 2010, moved sideways after 2010, and recently even increased slightly. Second, when dividing households into upper, middle, and lower groups based on total expenditures, the Engel coefficient of lower group households was the highest and that of upper group households was the lowest. This is consistent with Engel’s law. Third, according to the results of estimating the relative price of food compared to non-food items, the relative price of upper group households was high and that of lower group households was low. However, according to the results of estimating the total price index, the total price indexes of upper and lower households moved almost similarly. Fourth, the relative price of food compared to non-food items has increased significantly since 2010. Fifth, before 2010, there was a positive consumer price index bias every year, but after 2010, the consumer price index bias seems to have almost disappeared. Before 2010, the consumer price index seemed to have been overestimated, and the Engel coefficient had been decreasing significantly over time. However, after 2010, the decline in the Engel coefficient stopped, and the consumer price index bias seems to have disappeared. It seems that the background of this change was the increase in the relative price of food compared to non-food items.