Effects of accounting of characteristics heterogeneity in video game price indices

This paper attempts to construct hedonistic price indices for the computer game market from 2007 to 2023. The peculiarity of the work is that we use hedonistic regressions that consider additional factors reflecting the consumer's reaction to the object of study. In the previous study, we considered only the parameters of the objects set by the manufacturer. Previously, we modelled prices assigned by the manufacturer at the time of game release, and the study period did not cover 2022 and 2023. This paper uses final product prices recorded at the data collection time, and the period captures 2023. Hedonistic regression cleared the time-dependence of a price from the number of game characteristics and some game demand factors. However, there are unobservable hard-to-specify sets of characteristics, such as genre features or game mode. Calculating the outcomes of the described variables directly in the model equation is complex. At the same time, if we assume that they uncorrelation with the variables explicitly specified in the model equation, we can treat them as random effects of various kinds. It allows us to obtain efficient and perhaps even reasonable price index estimates. For building indexes, we use a standard hedonistic approach with temporary dummy variables, modified to allow for the random effects described above. We evaluate the constructed models using OLS and the simulation likelihood method. This approach identifies factors whose outcomes on prices differ significantly depending on which genres, modes and user audience scale groups the objects under consideration fall into. The final part of the study analyzes monthly data from 2020 to 2023 for a deeper examination of the effects following February 24, 2022.