The earthquake which took place in Gyeongju on September 12th, 2016 damaged some of the properties in this city and the area nearby and may have been recorded as the strongest ever in Korean history. More seriously, though, it caused most Koreans to be terrified with the idea that Korea is no more safe from earthquakes and that they may have their properties destroyed or lost all of sudden by natural disasters such as earthquakes they witnessed in Gyeongju area.

From this perspective, this study focuses on examining how housing market represented as apartments, in particular, living on high floors in this research reacted to Gyeongju Earthquake with transaction price data of those in Gyeongju, Ulsan, and Changwon from January 2016 to December 2017. To do so, it distinguishes periods such as before the earthquake and after the quake and then compares the results of each period.

The Hedonic function model uses a semi-logarithmic function model to take the logarithmic value for the dependent variable apartment price and the apartment price per area, and for other descriptive variables, the log value is not taken except for the distance variable. The structural characteristics of a house include the number of dwellings, the total number of floors, the number of years elapsed, the square of the years elapsed, the dedicated area, the seismic design pile, the cascaded pile, the city gas pile, the parking lot per generation, and the neighboring function, the urban pile and the area pile and the total number of households. As an accessibility variable, the distance from elementary school is used. In addition, if the analysis period is September 20, 2016, the distance from the epicenter was used as a variable.

The results of this empirical analysis :
The coefficients of living floor in both Gyeongju and Changwon before the quake show significantly positive signs, but insignificant for 3~6 months after the quake, still they recover the significantly positive sign after this period. This may imply that Gyeongju and Changwon might have overreacted to the quake. On the other hand, Ulsan which experienced a similar quake 2 months before Gyeongju Earthquake shows no significant impact of Gyeongju Earthquake on the preference of apartment height.

The proximity to the epicenter appears to have had a negative impact on apartment prices. In other words, in Gyeongju, the apartment price near the epicenter was higher because the place close to the epicenter was in the urban area and the remote area was in the rural area, and after the earthquake, the absolute value of the negative regression coefficient of the accessibility variable to the epicenter was reduced, which seems to have decreased relatively. In Ulsan, where the epicenter was close to the epicenter was also originally high in apartment prices, but the regression coefficient of distance variables to the epicenter temporarily changed to insignificant after the earthquake, so that proximity to the epicenter negatively affected apartment prices and then changed back to its original state six months after the earthquake.

It is necessary to recognize that the high occupancy level in the event of an earthquake does not necessarily result in a higher risk In other words, the risk of building collapse in the event of an earthquake is not great, so being in a higher level is not necessarily more dangerous. A lower occupancy level can escape faster than a higher floor from a building, but it is not safer to escape quickly. An earthquake while escaping from a building may cause a person to fall or be hit by a drop depending on the vibration. Rather, it may be safer to escape when an earthquake is temporarily calmed down.