What is My Car Worth? Hedonic Price Analysis of the German Used Car Market

• Název práce v češtině: Jaká je hodnota mého vozu? Hedonická metoda oceňování německého trhu ojetých vozů
• Název v anglickém jazyce: What is My Car Worth? Hedonic Price Analysis of the German Used Car Market
• Klíčová slova: trh ojetých vozů; cena ojetých vozů; hedonická metoda oceňování; Bayesovská metoda průměrování, Německo
• Klíčová slova anglicky: market for used cars; used car prices; hedonic price analysis; Bayesian model averaging; Germany
• Akademický rok vypsání: 2018/2019
• Typ práce: diplomová práce
• Jazyk práce: angličtina
• Ústav: Institut ekonomických studií (23-IES)
• Vedoucí / školitel: Mgr. Petr Polák, M.Sc., Ph.D.
• Řešitel: skrytý - zadáno vedoucím/školitelem
• Datum přihlášení: 15.05.2019
• Datum zadání: 15.05.2019
• Datum a čas obhajoby: 15.09.2020 09:00
• Místo konání obhajoby: Opletalova - Opletalova 26, O206, Opletalova - místn. č. 206
• Datum odevzdání elektronické podoby: 30.07.2020
• Datum proběhlé obhajoby: 15.09.2020
• Oponenti: RNDr. Michal Červinka, Ph.D.

Seznam odborné literatury

AKERLOF, G. A. (1978). The market for lemons: Quality uncertainty and the market mechanism. Uncertainty in Economics, Academic Press, 235 – 251.

Andrews, T. and Benzing, C. (2007). The determinants of price in internet auctions of used cars. Atlantic Economic Journal, 35:43–57.

Bauer, I., Zavolokina, L., and Schwabe, G. (2019). Is there a market for trusted car data? Electronic Markets.

Bond, E. W. (1982). A direct test of the ”lemons” model: The market for used pickup trucks. The American Economic Review, 72(4):836–840.

Gongqi, S., Yansong, W., and Qiang, Z. (2011). New model for residual value prediction of the used car based on BP neural network and nonlinear curve fit. Measuring Technology and Mechatronics Automation, International Conference on, 2:682–685.

Hoeting, Jennifer A., David Madigan, Adrian E. Raftery, and Chris T. Volinsky (1999). Bayesian Model Averaging: A Tutorial. Statistical Science 14, no. 4 (1999): 382-401. www.jstor.org/stable/2676803.

Kihm, A. and Vance, C. (2014). The determinants of equity transmission between the new and used car markets a hedonic analysis. SSRN Electronic Journal.

Kim, J.-C. (1985). The market for ”lemons” reconsidered: A model of the used car market with asymmetric information. American Economic Review, 75(4):836–43.

Madigan, D and Raftery, A. E (1994). Model Selection and Accounting for Model Uncertainty in Graphical Models Using Occam'sWindow. Journal of the American Statistical Association. 89: 1535 1546

Mishra, D. P., Heide, J. B., and Cort, S. G. (1998). Information asymmetry and levels of agency relationships. Journal of Marketing Research, 35(3):277–295.

Pal, N., Arora, P., Kohli, P., Sundararaman, D., and Palakurthy, S. S. (2019). How much is my car worth? A methodology for predicting used cars’ prices using random forest: Proceedings of the 2018 future of information and communication conference (ficc), vol. 1. Pages 413–422.

Pudaruth, S. (2014). Predicting the price of used cars using machine learning techniques. International Journal of Informatione and Computation Technology, 4:753–764.

Wasserman L. (2000). Bayesian Model Selection and Model Averaging, Journal of Mathematical Psychology, 44:92-107.

Wu, J.-D., Hsu, C.-C., and Chen, H.-C. (2009). An expert system of price forecasting for used carsusing adaptive neuro-fuzzy inference. Expert Syst. Appl., 36(4):7809–7817.

Předběžná náplň práce v anglickém jazyce

Motivation:
The automobile industry evinced significant changes in recent years. With the inception of the economic crisis in 2008, customers started with searching for alternative products. Therefore a relatively high retail price of a car steered their seek to the second-hand market.

In the consequence of the increasing used car demand arose a question which factors influence the price of the second-hand car most significantly. Individuals offering their cars on the market, professional merchants selling pre-owned cars, households intending to buy a new car and resale it in the future, lease companies, manufacturers and many other car market players wonder how to value a given car. This thesis aims to answer this question and to appraise car characteristics as value drivers.

In recent years, the hedonic price classification has attracted a lot of researches attention and therefore the related literature is relatively miscellaneous.
But some of the researches suffer from a limited number of observations or lesser number of car characteristics and from a missing regional effect. This thesis aims to evaluate a complete online database of used car comprised by hundreds of thousands advertisements and detail description of each automobile on the internet platform Autoscout24.de.

Moreover, the researches vary according to used methodology significantly. Refer to Madigan and Raftery (1994) who oppose unsubstantiated model selection in literature, the hedonic price model in this thesis will be performed on the basis of alternative approach to model selection: Frequentist Model Averaging (FMA) and Bayesian Model Averaging (BMA).

Hypothesis #1: What are the main car characteristics influencing a residual value of a used car?
Hypothesis #2: How the hedonic price factors differ in regions?

Methodology:
Cross-sectional data collected on the online platform Autoscout24.de will contain basic car characteristics (age, mileage, manufacturer, model, body type, transmission, fuel type, fuel consumption, etc.) as well as features related to a car equipment and geographical specification. The effect of variables to an offered price will be analysed using appropriate linear models and machine learning approaches. The most inspiring researches that will serve as examples are Gongqi, Yansong, and Qiang (2011) with their use of BP neural network and Pal, Arora, Kohli, Sundararaman, and Palakurth (2019) comparing linear regression and RF. The final model specification will be derived from Frequentist Model Averaging (FMA) and Bayesian Model Averaging (BMA).


Expected Contribution:
The target of this thesis is to complete a present literature concerning with the hedonic price analysis, to proffer a research based on a sufficient amount of observations and various car characteristics. Moreover, the accuracy of different used models in the context of BMA and FMA performing as appropriate model selections will be discussed. To conclude a practical application, both sellers and buyers (professional as well as individuals) can utilize expected results of this thesis. New information about the real value drivers could be able to mitigate an information asymmetry and an adverse selection problem.



Outline:
1. Introduction
2. The used car market description
This section will focus on the second-hand car market and its specificity. The integral part of this introduction will be a differentiation of selling channels, product definition and contemplation about policy interest in the used car market.
3. European specificity
To put a subsequent research into perspective of nowadays evolution of European used car market, this section will describe participants, market channels and second-hand market environment in general in European Union.
4. The used car market from the microeconomic point of view
This section will refer to a basic microeconomic identification of the market and market players. A deeper analysis will expose roles of sellers and buyers, an information problem and the last part will be concerned with adverse selection, quality level and market for lemons phenomenon.
5. Literature overview
The literature overview section focuses on two integral questions: what are the most important factors affecting the hedonic price and whether good products are driven out by bad quality cars. Both questions will be summarised by various recent studies differing in approaches. The second part will introduce key research papers concerning with BMA and FMA methods.
6. Methodology
Methodology used in the own research will be discussed in details referring to requirements, advantages and limitations.
7. Results
The overall results of this thesis will be evaluated and critically assessed. A potential future research questions will be defined.
8. Conclusion