[Financial Insider] "AI Reveals 'Marketable Prices' for Every Apartment Unit Nationwide"

Interview with Lee Jong-ah, Head of KB Real Estate Big Data Center
"From next year, we will also predict prices 6 months and 1 year ahead"

Lee Jong-ah, Head of KB Real Estate Big Data Center at KB Kookmin Bank. Photo by Younghan Heo younghan@

This year, in the second half, apartment prices measured by artificial intelligence (AI) will be released on the KB Real Estate app. You will be able to see the 'transactionable price' for every individual unit of all apartments nationwide. Next year, AI plans to predict and announce apartment prices 6 months and 1 year in advance. The KB Real Estate Big Data Center, under KB Kookmin Bank, which was launched last August, has developed a real estate value estimation model (AVM).

On the 19th, at KB Kookmin Bank’s headquarters in Yeouido, Lee Jong-ah, head of the Big Data Center, said, "When deciding on real estate transactions and when banks assess risk, the accuracy of price trends is extremely important," adding, "We will expand information based on big data so that when you enter KB Real Estate, you can see not only listings and market prices but also transactionable prices and estimated prices at a glance."

▲You developed an AVM based on big data and AI. What does it show?

-You will be able to check the transactionable price of individual apartment units on the KB Real Estate app. Currently, KB announces prices by exclusive area. It shows upper average price, general price, and lower average price. Going forward, it will become much more detailed. We will provide information like 'Unit ○○○ in Building 11 of Daechi Eunma Apartment costs this much, Unit ○○○ in Building 201 of Mapo Prugio costs that much.' By entering individual apartment units nationwide, you can see the appropriate price. We plan to release a beta service within this year. Next year, we plan to estimate and publish apartment prices 6 months and 1 year ahead every month.

Lee Jong-ah, Head of KB Real Estate Big Data Center at KB Kookmin Bank. Photo by Younghan Heo younghan@

▲How is the apartment price measurement conducted?

-We utilize both public and private data. Public data includes actual transaction prices, building register information, interest rates, GDP (Gross Domestic Product), and government policy data. KB Real Estate holds data on market prices, development, school districts, floor utility, past transaction prices, and prices of surrounding apartments. AI processes hundreds of types of data. This is to estimate "what the price is at this point in time." There are concerns that real estate policies or interest rates might be difficult to predict. However, AI learns how prices moved when similar events occurred in the past and can predict prices accordingly. The opening of a subway line or the construction of a shopping mall is also reflected in the same way. Although there may be deviations depending on market conditions, it is meaningful to show the price trend.

▲How accurate is the prediction?

-There are similar places in the U.S. that predict real estate prices. I heard their prediction accuracy was high but has recently declined somewhat. They have many single-family homes. Because of irregular factors, it may be more difficult to predict. In the case of apartments, even within a complex, factors like which building you live in, the direction it faces, and the floor are close to fixed values. We will improve accuracy by learning other factors with AVM. Previously, prices were surveyed centered on people. Now, by combining AI, we can understand prices more accurately.

Lee Jong-ah, Head of KB Real Estate Big Data Center at KB Kookmin Bank. Photo by Huh Young-han younghan@

▲What impact will this have on the real estate and financial markets?

-When deciding whether to make a transaction and when banks assess risk, the accuracy of price trends is very important. Data is key in real estate. The role of the KB Real Estate Big Data Center is to connect data accumulated since the Housing Bank era and show as much information as possible. Measuring transactionable prices of apartments nationwide and predicting future prices will also help financial stability. When you enter KB Real Estate, you will be able to see actual transaction prices, listings, market prices, and AI-estimated prices all at once to facilitate transactions. This will be an important indicator for sellers, buyers, and brokers.

▲What is the outlook for the real estate market?

-Prices have risen sharply over the past 2 to 3 years. Currently, the cycle is in a downward trend. However, due to the special situation of COVID-19 since 2020, liquidity has been very abundant. Although prices have recently risen slightly, it is hard to say purchasing power has significantly increased. It will take more time before prices rise again. Among the macro variables affecting real estate prices, income is the most important. Employment in Korea is relatively stable. Even if people borrow to buy homes, they repay diligently. That is why even during economic crises, the possibility of a sharp drop in real estate prices has been low. How well this employment is maintained will have a major impact on the real estate market going forward.