Methodological remarks
The Használtautó.hu - HCSO experimental statistics were created in the framework of a cooperation agreement with Használtautó.hu Ltd. The database under study is the total supply data set of car advertisements managed by Használtautó.hu, for the 6th, 15th and 24th of each month. With the exception of Figure 4, which shows the variation over time, duplicates have been filtered out to produce statistics for the whole month. If an advertisement has a changed asking price within the month, the price in the statistics is the average of the three days of the month.
Cases where a statistically relevant data gap or data error occurred were removed from the dataset. In addition, vehicles less than one year old, vehicles with 0 kilometres and vehicles less than HUF 50,000 were filtered out.
The fuel types distinguished by Használtautó.hu were classified in simplified categories. The table below shows the matching:
Használtautó.hu category
Categories for the experimental statistics publication
Diesel
Diesel
Petrol
Petrol
Petrol/gas
Other
Diesel/gas
Other
Hybrid (petrol)
Hybrid
Hybrid (diesel)
Hybrid
Hybrid
Hybrid
Electric
Electric
LPG
Other
CNG
Other
Hydrogen/electric
Other
Ethanol
Other
Gas
Other
Biodiesel
Other
Other
Other
Empty
Empty
A regression model was used to examine the components of used car prices. This procedure quantifies the partial effects, assuming all other factors are held constant. The dependent variable used in the model is the natural logarithm of the supply price of cars, explained by the following independent variables:
-
age of the car in years interacted with the five fuel types (petrol, diesel, hybrid, electric and other)
-
mileage in 1,000 km
-
cylinder capacity in 100 cm3-ben
-
power in kW
-
condition of the car as dummy variables excellent, normal, spared, undamaged, nearly new, damaged/ faulty
-
brand groups
The clustering of brands was created by Ward clustering, which created 8 distinct categories based on the following variables: asking price, age, mileage, engine capacity, power. Peripheral brands with fewer than 10 cases were excluded from the process, accounting for approximately 0.1% of the dataset.
For the regression model, heteroscedasticity tests and normality tests showed satisfactory results.
Methodological remarks
The Használtautó.hu - HCSO experimental statistics were created in the framework of a cooperation agreement with Használtautó.hu Ltd. The database under study is the total supply data set of car advertisements managed by Használtautó.hu, for the 6th, 15th and 24th of each month. With the exception of Figure 4, which shows the variation over time, duplicates have been filtered out to produce statistics for the whole month. If an advertisement has a changed asking price within the month, the price in the statistics is the average of the three days of the month.
Cases where a statistically relevant data gap or data error occurred were removed from the dataset. In addition, vehicles less than one year old, vehicles with 0 kilometres and vehicles less than HUF 50,000 were filtered out.
The fuel types distinguished by Használtautó.hu were classified in simplified categories. The table below shows the matching:
| Használtautó.hu category | Categories for the experimental statistics publication |
|---|---|
| Diesel | Diesel |
| Petrol | Petrol |
| Petrol/gas | Other |
| Diesel/gas | Other |
| Hybrid (petrol) | Hybrid |
| Hybrid (diesel) | Hybrid |
| Hybrid | Hybrid |
| Electric | Electric |
| LPG | Other |
| CNG | Other |
| Hydrogen/electric | Other |
| Ethanol | Other |
| Gas | Other |
| Biodiesel | Other |
| Other | Other |
| Empty | Empty |
A regression model was used to examine the components of used car prices. This procedure quantifies the partial effects, assuming all other factors are held constant. The dependent variable used in the model is the natural logarithm of the supply price of cars, explained by the following independent variables:
-
age of the car in years interacted with the five fuel types (petrol, diesel, hybrid, electric and other)
-
mileage in 1,000 km
-
cylinder capacity in 100 cm3-ben
-
power in kW
-
condition of the car as dummy variables excellent, normal, spared, undamaged, nearly new, damaged/ faulty
-
brand groups
The clustering of brands was created by Ward clustering, which created 8 distinct categories based on the following variables: asking price, age, mileage, engine capacity, power. Peripheral brands with fewer than 10 cases were excluded from the process, accounting for approximately 0.1% of the dataset.
For the regression model, heteroscedasticity tests and normality tests showed satisfactory results.