The real estate market is showing unusual signs. As global tightening begins, experts worry that the bubble in the domestic real estate market, which benefited from the post-COVID-19 liquidity, may burst. They warn we should prepare for a potential impact on the real economy.

Since late last year, major central banks, including the U.S. Federal Reserve, have been raising interest rates to combat inflation. This has caused housing prices to decline, reducing household net worth and increasing losses for real estate developers, which could potentially trigger a recession.

Global Liquidity and the Surge in Housing Prices

Meanwhile, some investors are attempting to exploit the 'bubble' in the real estate market for profit. They expect prices to fall soon and aim for capital gains by buying low. Others seek arbitrage opportunities, assuming prices haven’t yet aligned with fair value. For these investors, it is crucial to assess whether current property prices are discounted or overpriced compared to intrinsic value.

Similarly, for financial institutions heavily involved in mortgage lending, analyzing the real estate market is key to the success of their loan business. This study examines why identifying the 'bubble' in the real estate, especially in auctions, is important and how it can be explored mathematically and statistically.

Importance of Real Estate Auction Market

Various stakeholders participate in Korea's real estate auction market, each with distinct objectives. Homebuyers, investors seeking profit opportunities, and financial institutions managing mortgages are all active players. The apartment auction market, in particular, is highly competitive, with prices often closely aligned with those in the regular sales market.

Financial institutions are closely connected to the auction market. In Korea, when a borrower defaults on a property loan, the property is handled through court auctions or public sales overseen by the state. Financial institutions recover the loan amount by selling the collateral through these auctions in the event of a default.

Therefore, one of the key factors for financial institutions in determining their lending limits is how much principal they can recover in the auction market in the event of a default, especially for fintechs (P2P lending) and secondary lenders such as savings banks and capital, which are not subject to loan-to-value (LTV) restrictions.

Since most financial institutions hold a significant portion of their assets in mortgage loans, lending the maximum amount within a safe limit is ideal for maximizing revenue. Thus, when financial institutions review mortgage loan limits, trends in the auction market serve as a critical decision-making indicator.

To See Beyond Prices in the Market

It's easy to assume that the winning bid for an apartment auction in a certain area of Seoul, at a particular time, would either come at a discount or a premium compared to the general market price. And, with a bit of rights analysis, setting a cautious upper limit wouldn't be all that hard. But, in reality, it's a bit more complex than just making those assumptions.

Furthermore, if we want to examine the market movement from a broader perspective rather than focusing on individual auction cases, we need to change our approach. For example, it's easy to track Samsung’s stock price trends in the stock market, even down to minute-by-minute data over the past year. However, in real estate, auctions for a specific apartment, like Unit 301 of Building 103 in a particular complex, don’t happen every month. Even expanding the scope to the whole complex yields similar results. Therefore, it's no longer feasible to analyze the market purely based on prices. Real estate analysis must shift from a [time-price] perspective, as in stocks, to a [time-location] perspective.

Errors in the Auction Winning Bid Rate Indicator

Just as the general sales market has a time-series index like the apartment sales index, the auction market has the winning bid rate indicator. This is a monthly indicator published by local courts, showing the ratio of auction-winning bids to court-appraised values in a given area. For example, if the court appraises a property at 1 billion won and the winning bid is 900 million won, the winning bid rate would be 90%.

Since court appraisals are generally considered market prices, the winning bid rate represents the ratio of the auction price to the market price. When calculated for all auctions in an area, it gives the average auction price compared to the market value for that month.

However, this indicator has significant flaws. The court appraisal is set when the auction begins, but the winning bid reflects the price at the time of the auction. Given that auctions typically take 7 to 11 months, this time gap can lead to errors if market prices drop or rise sharply. For instance, news reports during recent price surges claimed that the winning bid rate in Seoul exceeded 120%, which seems hard to believe—how could auction prices be 1.2 times higher than market prices? This is actually incorrect information.

If market prices rise sharply during the 7 to 11 months it takes to complete an auction, bidders place their bids based on current market prices, while the court appraisal remains fixed at the start. As a result, the appraised value becomes relatively lower compared to the current market price, creating the illusion of a 120% winning bid rate. Interpreting this rate at face value can lead to poor real estate decisions or significant errors.

Limitations of Previous Studies

This has prompted previous auction market studies to try addressing the shortcomings of the winning bid rate indicator. For instance, some researchers adjusted the court-appraised value—the denominator—by factoring in the sales index at the time of the auction, aiming to estimate a more accurate "true winning bid rate."

However, experts agree this is not a perfect solution. To estimate the true winning bid rate for Seoul, all auctions during that period would need to have their court appraised values corrected. Each auction has different appraised values and closing dates, and researchers would have to manually correct hundreds or thousands of auctions. Expanding the region would make this task even more challenging, and even if corrected, the values would only be approximations, not guarantees of accuracy.

If researchers selectively sample auctions for convenience, it could introduce sampling bias. This is similar to trying to find the average height of Korean men by only sampling from a tall group.

It highlights the need for time-series indicators from a market perspective when making business decisions, rather than focusing solely on price data. The winning bid rate, commonly used in auctions, is prone to errors. Although methods like adjusting the court-appraised value have been suggested, they are difficult to apply in real-world scenarios.

These are the same problems I encountered as a practitioner. When time-series analysis was needed for decision-making, the persistent issues with the winning bid rate made it hard to use effectively.

Winning Price vs. Winning Bid Rate

There is an important distinction to make here. Analyzing the auction "winning price" and analyzing the "winning bid rate" have different meanings and purposes. As mentioned earlier, analyzing the winning price of a single auction case poses no issue.

For example, focusing on apartments, bidders base their bids on the market price at the time of bidding. If the gap between the bidding and final winning is about 1 to 2 months, considering that real estate prices don’t fluctuate dramatically like stocks within a month, the winning price should not significantly deviate from the market price a couple of months earlier. Factors like distance to schools, floor level, and brand, which are known to affect market prices, are likely already reflected in the market price, meaning they won't heavily impact the auction winning price.

Most prior studies on real estate auctions, particularly for apartments, have concentrated on how accurately they can predict the "winning price" and identifying the key factors that influence it. However, in practice, as previously discussed, price prediction is not the primary concern.

Even a simple linear regression analysis reveals that the R-squared between winning prices and KB market prices from 1-2 months earlier exceeds 95%, indicating a strong linear relationship. There is no evidence of a non-linear connection. If future trend forecasting is needed, the focus should shift toward a time-series analysis.

Discounts/Premiums Changing Over Time

After a lengthy introduction, let's get to the main point—I want to analyze the auction market. The problem is, the data contains significant errors, and trying to correct them individually has its limitations, especially within the industry. We need a different approach. So, what alternative methods can we use? And what insights can this new analysis reveal?

This is the core topic and background of the study. I used statistics as a tool to solve a seemingly insurmountable business problem encountered in practice.

What I aimed to find in the market was the difference between the sales market and the auction market. This 'difference' can be expressed as the discount or premium of the auction market compared to the sales market. Additionally, a time-series analysis is essential because the discount/premium factors will change over time depending on the economic or market conditions.

Factors of Discount/Premium in the Auction Market

Existing studies on the factors of discount/premium in housing auctions are quite varied. Nonetheless, as mentioned earlier, both international and domestic research mainly focus on price analysis rather than market analysis, making it difficult to grasp the broader trends of real estate. Typically, they gather auction cases over several years, remove legal issues, compare with market prices, and conclude there was a discount/premium, attributing it to specific factors.

Moreover, overseas studies often allow private auctions and use bidding systems, making direct application to Korea difficult. In domestic studies, the few that exist lack market-based analysis.

The Challenge: 'Data Availability'

If we assume that there is a discount/premium factor in the auction market compared to the sales market, the auction sale rate can be restructured as follows:

[ Auction Sale Rate_{t} = \frac{\sum Market Price_{t} \pm Premium_{t}}{\sum Appraisal Price_{t-n}} ]

Now, interpreting the three elements of the auction sale rate as influential factors and transforming it into a linear regression model, it would look like this:

[ Auction Sale Rate_{t} = \beta_{0} + \beta_{1} EoM_{t} + \beta_{2} EoA_{t} + \beta_{3} EoP_{t} ]

• EoM: Effect of Market Price (influence of the general sales market)
• EoA: Effect of Appraisal Price (influence of court appraised values)
• EoP: Effect of Price Premium (influence of discount/premium)

To complete this regression model, data for all three variables is needed. The effect of market prices can be substituted with the sales index provided by the Korea Real Estate Board. The sales index should be transformed using a log difference to match the format of the auction sale rate.

A major challenge lies in obtaining data for the other two variables. First, acquiring court-appraised price data for research purposes is nearly impossible. The focus here isn't on the historical 'appraised price' itself, but rather on how much it influenced each analysis period (typically monthly). This means we need data adjusted to the auction's closing time. However, without digitizing all auction cases nationwide over the past 10 years, this task is virtually unachievable. The unobservable variables are intertwined, resembling a noisy background.

Factor Separation and Extraction

How can we isolate a specific male voice from a noisy mix of sounds? This is where the Fourier Transform comes in. It converts an input signal from the time domain to the frequency domain, separating each individual frequency. By applying an inverse Fourier Transform, we can find the unique voice, leaving it intact while setting other elements to zero, effectively filtering out noise.

In the same way, if we view the auction sale rate as a noisy input signal, we can separate its contributing factors independently. First, by removing the effect of market prices from the auction sale rate using a regression model, we can assume that the residual term contains hidden influences from court appraisals and discount/premium factors. Among the remaining elements in the residuals, we can assume the two strongest factors are the court appraised price and discount/premium. Fourier Transform can then be used to extract these two independent signals.

This assumption can be statistically verified. As shown in the table, when regressing the auction sale rate using the three initially assumed variables—two components extracted by Fourier Transform and the market price data—the adjusted R-squared is about 94%. In other words, the auction market can be explained by these three factors (market price, court appraisal, and discount/premium). Additionally, the ACF/PACF plot of the residuals after Fourier extraction (see figure below) shows no significant remaining patterns.

Through the Fourier Transform, I was able to resolve both the limitations of the auction sale rate as a time-series data and the issue of relying on external data. I successfully extracted the two remaining factors (court appraisal and discount/premium) from the residuals after removing the effect of market prices.

However, I must caution that using Fourier Transform on general asset market data, like stocks or bonds, is risky. This method is only applicable to data with consistent cycles. Unlike price or sales indices, auction sale rate data exhibits cyclic movements between 80-120%, driven by economic and market conditions, allowing the process to be performed without errors.

Court Appraisal Extraction

The two factors extracted through the Fourier Transform are currently only assumptions, believed to represent court appraised value and the discount/premium factor. Therefore, it is necessary to accurately verify if these factors are indeed related to court appraised values and discount/premium factors. First, I analyzed two aspects using around 2,600 auction cases:

• The average time gap between the court appraisal date and the auction closing date.
• The relationship between court appraisal prices and KB market prices at the appraisal date.

The time gap between appraisal and auction ranged from 7 to 11 months (within the 25% to 75% range), and the relationship between the court appraisal price and KB market price showed a Beta coefficient of 1.03, indicating almost no difference. Based on these two results, I reached the following conclusions:

• There is a lag relationship between court appraisal prices and market prices (lag = time gap).
• The lag variable of market prices can substitute for court appraisal prices.

Regression analysis showed that the lag variable of the sales index and court appraisal had about 54% explanatory power. This confirmed that the court appraisal component extracted via Fourier Transform could function as an actual court appraisal. Additionally, when comparing how well the lag variable and court appraisal component explained the auction sale rate, the appraisal component (50%) outperformed the lag variable (20%).

Discount/Premium Extraction

Next, I tested the discount/premium component, the core of this study, from two angles. First, whether the component extracted by the Fourier Transform can function as a discount/premium factor, and second, what the true identity of this component is.

For verification, I applied a sigmoid function to the discount/premium component to produce an on/off effect (0/1).

I attempted to compare this component with various data available from sources like the National Statistical Office, but I couldn't find any data showing similar patterns. The reason for this was simpler than expected.

The auction market is dependent on the sales market. Most macroeconomic variables we know likely influence housing prices, which have already been removed from the regression model. Therefore, the remaining factors are likely unique to the auction market, independent of sales prices. The variable that shows a similar pattern to the discount/premium component is the month-over-month and two-month differences in the winning bid rate, as shown in the figure.

The Nature of the Discount/Premium

To summarize the analysis so far: after excluding the effects of market prices and court appraisals, the remaining factor in the auction sale rate is the discount/premium factor. This factor exhibits a similar pattern to the month-over-month fluctuations (volatility).

In other words, if past volatility explains what the 'sales price' and 'court appraisal' couldn't, it suggests that the auction market has a discount/premium factor driven by volatility (the difference in past winning bid rates). As I will explain later, I have named this component the 'momentum factor,' believing it to explain trends.

Cluster Characteristics of the Momentum Factor

As we delve deeper into this analysis, it's essential to recognize that auction market dynamics are not static but evolve over time, necessitating a more adaptive model to track these changes effectively.

Unlike Ordinary Least Squares (OLS) regression, which assumes a fixed beta coefficient, the Kalman Filter's state-space model allows the beta coefficient to change over time. By tracking this time-varying coefficient, we can observe how the influence of different variables fluctuates over various periods. To analyze the 'momentum factor' in greater detail, I applied the Kalman Filter to assess whether the beta coefficient indeed varies over time.

Consequently, as shown in the figure below, we can observe that the regression coefficient of the momentum factor exceeds that of the sales price regression coefficient in certain intervals. Upon examining these intervals, it becomes clear that the momentum factor exhibits a type of clustering effect.

The True Meaning of the Momentum Factor

We need to think more deeply about the "intervals where the momentum factor's sensitivity exceeds market prices." The momentum factor explains the discount/premium. Therefore, when the discount/premium factor significantly impacts the auction sale rate, it suggests that the usual "average relationship" between the sales market and the auction market has been disrupted.

What does it mean when the "average relationship is disrupted"? For example, if the sales and auction markets typically maintain a gap of 10, this disrupted relationship means the gap has shrunk to 5 or expanded to 15. Such situations typically occur during overheated or excessively cooled markets, or just before such conditions arise. When everyone is rushing to buy homes, this can naturally lead to an "overheating" that breaks the usual relationship, which can be interpreted as increased "popularity" in the auction market.

However, one important thing to note is that when the sales market falls, the auction market typically falls too. This is because the market price has the largest influence on changes in the auction sale rate. In other words, even if the momentum factor is inactive, the auction market can rise or fall in response to the sales market. Therefore, the "activation of the momentum factor" doesn't necessarily indicate price increases or decreases.

The discount/premium factor is ultimately defined as the effect of market prices + '@'. The sensitivity analysis of the discount/premium factor indicates that '@' represents "excessive movement beyond the average." I named this the "momentum factor" because I believe it can detect changes in market sentiment or trends. As seen in the Figure 5, the momentum factor tends to signal market trend changes before and after its cluster periods.

I cautiously suggest that when the momentum factor shows excessive movement, it could signal a "bubble" or "cooling" sign. Further exploration of this idea is beyond the scope of this [paper discussion], but it certainly warrants future research.

Focus on Logic Over Technique

The reason I wrote this paper wasn't because I majored in real estate or specialized in the field. Most of my work in recent years involved changing systems to enable data-driven decision-making, one of which was loan screening, and another was related to real estate.

I understand that definitions of data science vary from person to person. However, for me, data science was the perfect tool for solving business problems. That’s why I chose a topic that was considered insurmountable in practice and applied the knowledge I learned in school.

The aspect I want to highlight in this paper is not the technical side but the logical one. The techniques used—regression analysis, Fourier transformation, and the Kalman filter—are not particularly advanced for graduate-level science and engineering. There was also an incentive to avoid using non-linear pattern matching techniques like ML/DL, which are unsuitable for financial data requiring clear interpretations. For me, it was more important to choose the method suited to the problem, and nothing more. The key was how to logically solve and approach this issue.

I believe that in solving business problems, logic should come first, and technology is just a tool. This is my ideal approach, and I wanted to keep the paper's concept simple, yet logically solid.

The Gap Between Business and Research

When I started researching for this paper, I remember thinking, "What problem should I try to solve?" My obsession with problem-solving came from the belief that there is a gap between the worlds of business and research, a bias I developed through experience.

As a practitioner, I think that in most fields, decisions are still largely based on subjective judgment rather than data. Furthermore, I know that many industries face challenges in successfully adopting data analysis systems, and I personally experienced this. While each field has its own circumstances, I believe one key reason for this gap is the disconnect between research and business.

From the industry perspective, I often felt that many research results focused on the study itself, neglecting "real-world applicability." On the other hand, from an academic perspective, I found that business often relied too heavily on subjective decisions, ignoring the complexities of the real world.

Bridging the Gap

Thus, the real intent of this paper was to bridge the gap between business and research, however small that contribution may be. I wanted to be a "conceptualizer" who actively uses data analysis to solve business problems. In this sense, I believe this paper sits somewhere between research and business. Throughout the writing process, I fought hard against the temptation to get lost in academic curiosity, focusing instead on practical applicability.

The quality and results of the paper will be judged by reviewers or proven in real-world industries, not by me. However, I anticipate that my future work will also be positioned between these two worlds. Connecting these two domains is an incredibly fascinating challenge. To view the article in Korean, please clickhere.

Ⅰ. What if we could detect a real estate bubble?

In 2021, the real estate market entered a recession as the bubble burst. The government is hastily preparing policy to activate the market, but it doesn't seem to be working as planned. However, if we could detect the real estate bubble in advance, wouldn’t it be possible to prevent the market from entering a recession?

The impact of the real estate bubble

The impact of the real estate bubble was immense. Apartment prices recorded the largest decline and Seoul's overall housing sale prices experienced the biggest drop since the subprime mortgage crisis.

Along with the decline in real estate prices, real estate transactions have also decreased as financial authorities sharply raised the base rate to reduce liquidity. In Seoul, homeowners are even offering to cover maintenance fees, moving costs, and luxury bags to promote leases and sales, but unlike before, transactions are not happening as actively. As seen in Figure 1, the number of apartment transactions in Seoul from January to September 2022 fell below 10,000, a decrease of 73.7% compared to the last year.

Many experts are concerned that the real estate recession could directly lead to a shock in the economy. To prevent this, they argue that the current real estate policies must be swiftly overhauled with new taxation policies suited to the era of high interest rates. They think the comprehensive real estate tax and capital gains tax are too burdensome, and coupled with high loan interest rates, real estate transactions are not occurring actively. Therefore, they suggest that easing real estate taxes and removing regulations are necessary to boost transaction volumes.

On the other hand, hastily changing real estate policies could be risky. The massive liquidity and low-interest rate environment caused by the global pandemic led to speculative behavior among people in their 20s and 30s, trapping many young adults in debt. Considering this reality, indiscriminately easing or removing real estate regulations could be dangerous. Therefore, the government should maintain appropriate levels of regulation to curb speculation and ensure that housing opportunities are available for real homebuyers.

Governments have alternated between these two approaches to find policies that would minimize the impact of housing bubbles. However, they have been unable to find an ideal solution that satisfies everyone. Excessive regulation risks violating the basic market principle that prices are determined by supply and demand, while too much leniency can lead to market disruptions like speculation, over-leveraging. It's extremely difficult to strike a balance in policy-making that satisfies everyone's needs.

What if we could identify a real estate bubble in advance?

Many people fail to recognize a bubble in the real estate market until prices crash. This is because accurately measuring the intrinsic value of real estate is difficult. As a result, most market participants mistake rising property prices for an increase in intrinsic value and get swept up by the decisions of others, leading to a bubble. When the bubble bursts, the inflated asset prices drop, which can lead to increased household debt, large-scale bad debts for financial institutions, and, in severe cases, an economic recession.

However, what if we could detect a real estate bubble in advance? It could help resolve many of the concerns mentioned above. The government would be able to identify overheating in the real estate market early and take an action before the bubble negatively impacts the real economy. Moreover, given the high level of global interconnectedness today, a bubble in one country can have significant effects on the global economy, making the prediction of bubbles increasingly important.

In this article, I will explain the steps taken to identify factors related to a real estate bubble and statistically verify them. Specifically, we will examine whether the "winner’s curse", often cited as a cause of real estate overheating, truly corresponds to a bubble by using regression analysis and statistical testing.

II. The History of Bubbles and the Reasons They Recur

A bubble refers to a phenomenon where the price of a specific asset significantly exceeds its intrinsic value due to excessive demand. This typically occurs when the economy becomes overheated.

When a bubble bursts, it leads to massive losses for investors and delivers a significant blow to financial institutions whose main business is providing mortgage loans. This could result in systemic risk across the financial market.

The History of Bubbles

Bubbles have historically repeated themselves multiple times in the global financial market. Examples include the Dutch Tulip Bubble of the 1630s, the South Sea Bubble in Britain during the 1720s, the Japanese real estate and stock market bubble of the 1980s, the Dot-com Bubble of the 1990s, and the U.S. housing bubble of the 2000s.

Let’s first take a closer look at the Japanese real estate bubble. In the early 1980s, as the yen surged and Japan's trade situation worsened, the Japanese government implemented monetary policies to stimulate the economy. With increased liquidity in the market, speculation was fueled, leading to a bubble between 1985 and 1989, during which the value of Japanese stocks and urban land tripled. At the peak of the real estate bubble in 1989, the value of the Imperial Palace grounds in Tokyo exceeded the total real estate value of the state of California. Ultimately, the bubble burst in 1991, leading to Japan's prolonged economic stagnation, known as the "Lost Decade."

Next, let’s look at the U.S. subprime mortgage crisis. After the dot-com bubble burst, many investors, learning from the experience, shifted money into real estate, which was considered a relatively safe asset. As a result, U.S. housing prices nearly doubled between 1996 and 2006. Additionally, as interest rates dropped, people rushed to buy homes using mortgage loans. However, to control the skyrocketing housing prices, the U.S. government sharply raised interest rates, leading to a wave of defaults by subprime borrowers who were unable to repay their loans. This turned mortgage-backed securities into worthless assets and pushed banks and other financial institutions on the brink of collapse.

The U.S. subprime mortgage crisis had a significant impact on South Korea as well. Major hedge funds and investment banks, including prominent U.S. financial institutions like Bear Stearns and Lehman Brothers, faced bankruptcy. Learning from this, foreign investors began favoring safer assets, causing significant volatility in the Korean foreign exchange market. As the interest rate spread between the U.S. and South Korea widened, the carry trade became widespread. It leads to negative effects on both the domestic financial market and the real economy in Korea.

The Reasons Bubbles Recur

The main method of detecting a real estate bubble is by examining the ratio between the money supply and the market capitalization of apartments. Generally, when the money supply increases, the value of money declines, leading to a rise in apartment prices. However, if the gap between the growth rates of the money supply and apartment market capitalization becomes unusually large, it may indicate a bubble. This suggests that apartment prices are rising independently of the available money supply. In November 2021, when the bubble was at its peak, the ratio of apartment market capitalization to the money supply soared to 147%. This shows that apartments were highly overvalued compared to their intrinsic value.

So why do bubbles continue to recur in cycles? Despite experts consistently presenting objective indicators, such as the one mentioned above, and warning about signs of a real estate bubble, why do people persist in risky behaviors like over-leveraging ("all-in" borrowing) and speculative investments ("betting with borrowed money")?

Robert Shiller, the 2013 Nobel Prize winner in economics, argued in his renowned book "Irrational Exuberance" that most market participants do not fully understand the true nature of the market. He further stated that people often don’t even care about why the market might be undervalued or overvalued. In such an environment, people’s investment decisions are heavily influenced by easily accessible information. In other words, instead of conducting deep quantitative and qualitative analysis, most investors are drawn to shallow, hearsay-like information, leading them to make decisions that are closer to gambling.

The core principle of a bubble can be summarized in one word: "Herd effect". In these situations, the independence of individuals breaks down, leading to irrational decisions made collectively.

III. Characteristics of the Real Estate Auction Market in Korea

In this section, let's explore why we should examine the auction market as a tool for predicting bubbles in the real estate sales market.

Recently, the real estate sales market has suffered a transaction freeze. The auction market typically comes into the spotlight when the real estate market enters a downturn. Due to the bleak outlook, competition for successful bids significantly decreases, and the bid price ratio — the ratio of the final bid to the appraised value — also drops noticeably. This, in turn, increases the incentive for investors.

In fact, the auction market appears to be reviving. According to a report published by the court auction specialist firm, Gigi Auction, the number of apartment auctions nationwide in October 2022 was 1,472, marking an upward trend after recording 1,330 cases in June of the same year. Additionally, the nationwide apartment bid price ratio in October 2022 was 83.6%, only a 0.5 percentage point increase from 83.1% in September of the same year, which was the lowest level since 2019. Considering that appraised values are like market prices, this suggests that recently, prices in the auction market are lower than in the sales market.

Savvy investors are turning their attention to the real estate auction market to exploit niche opportunities. If they can properly analyze the real estate market and identify undervalued areas, they will be able to fully enjoy excess returns in the auction market.

Bubbles can also occur in the auction market

A question naturally arises: Can bubbles occur in the auction market just like in the sales market? To answer this, we need to understand the real estate auction market in Korea.

In Korea, auctions are widely perceived as a way to buy real estate at a lower price. However, one peculiar aspect of the Korean auction market is the frequent occurrence of the so-called "Winner's Curse," where the winning bidder ends up overpaying due to overestimation of the property's value or intense competition. This phenomenon is often attributed to Korea's unique real estate auction system, which sets it apart from those in other developed countries.

Korea’s real estate auction system employs a sealed-bid auction and a first-price auction format. In a sealed-bid auction, bidders cannot see the prices offered by others, ensuring independence between participants. In the first-price auction, the highest bidder wins and pays the amount they submitted. Bidders aim to submit a price lower than the market value but higher than their competitors, so the bid prices generally don’t vary greatly. Unless the bidder is a stakeholder, such as a tenant or creditor, it is rare for someone to submit an overwhelmingly higher bid than others.

However, if current real estate prices do not accurately reflect intrinsic value, or if expectations of future price increases take hold in the market, the situation changes. Market participants, anticipating excess returns, will flood into the auction market, leading to intense competition. As a result, the gap between the winning bid and the second-highest bid will widen significantly. Moreover, as bidders are driven by herd effect and inflate bid prices, this phenomenon closely mirrors the bubble seen in the real estate sales market.

In addition, the real estate auction market is known to precede the sales market. As we observed earlier, when real estate prices begin to rise, properties listed in the sales market often move to the auction market at lower prices, activating the auction market. Therefore, if we can detect a bubble in the auction market through data analysis, it could also serve as an indicator to identify a bubble in the real estate sales market.

Ⅳ. Bubble index: Price Differences between the 1st and 2nd Place Bids in Auction Market

Literature review

Previous studies on predicting real estate market prices or auction winning bids have employed the Hedonic Price Model and Time Series Model. The Hedonic Model is a regression model based on the assumption that the price of a good is the sum of the quantities of its inherent characteristics. In prior research, the focus was on increasing accuracy by adding as many variables as possible that could represent the characteristics of real estate. However, adding a large number of variables poses a risk. Including unnecessary variables without sufficient validity can lead to multicollinearity, which increases the variance of the estimates and results in unreliable outcomes. This is also the reason why research using the Hedonic Price Model has not been conducted since the mid-2010s.

Data and variable selection

In this study, we aim to address the issues found in previous research by introducing the price differences between the 1st and 2nd place bids in the auction market as an indicator of a bubble and statistically verifying this index.

In this study, the data consists of the quarterly transaction volumes from 2014 to 2022 for the Gangnam and Nowon districts. These areas were deemed most suitable for the study, as Gangnam and Nowon are the regions in Korea with the most active transaction and bidding activities.

Additionally, while this study is based on the Hedonic Price Model as a foundation, it introduces some modifications. The dependent variable is the corrected winning bid rate(y), while the independent variables include the number of unsuccessful bids(FB_NUM), the number of bidders(BD_Num), the bubble index (Index_5), and the M2 currency volume(M2), distinguishing the variable selection from previous studies.

To elaborate on the dependent variable, the corrected winning bid rate, the original winning bid rate is calculated by using the court appraised value (typically the KB market price) as the denominator and the winning bid as the numerator. However, there is a time gap of about 7 to 11 months between the appraisal and the winning bid. Therefore, this study adjusts the court appraised value, which is the denominator in the traditional winning bid rate, to reflect the market price at the time of the winning bid using the following formula.

To further explain the independent variables in this regression model, first, the number of unsuccessful bids(FB_NUM) serves as a control variable that explains auction risk factors. A high number of unsuccessful bids indicates that the auction price is set higher than the market price, leading participants to forgo the auction. Consequently, this suggests a higher likelihood that the next auction will also fail. Second, the number of bidders(BD_Num) refers to the number of people who participated in the auction. As market overheating(a bubble) occurs in the auction market, the number of bidders tends to increase, making it a potential indicator of a bubble. Third, the price differences between the 1st and 2nd place have been explained several times before, so it will be omitted here. Lastly, the inclusion of the M2 currency volume(M2) in the model considers the general trend that an increase in money supply often leads to a sharp rise in real estate prices.

Exclusion of the Intrinsic Value of Real Estate and the Necessity of the Chow Test

There are numerous factors that determine real estate prices, such as school districts, job opportunities, apartment floor levels, proximity to roads, building age, apartment structure, and transportation convenience. These elements that influence intrinsic value are extensive and complex. Therefore, in the regression model for the corrected winning bid rate, I applied a logarithmic transformation to eliminate the intrinsic value, allowing the focus to remain solely on the auction characteristics, which is the main purpose of this study. The detailed process is as follows.

This paper also assumes that a structural break occurs when a bubble forms, so conducts a Chow test to verify this. The assumption is that during periods of market overheating, such as a bubble, the market will behave differently due to irrational investment sentiment compared to other periods. The Chow test is a statistical test used to determine whether there has been a "structural shock or change" by comparing the regression coefficients of two linear regression models before and after a specific period in time series data.

Regression Analysis and Chow Test

First of all, I conducted regression analysis over the entire period. The dependent variable is the corrected winning bid rate(y), and the independent variables are the number of unsuccessful bids (FB_NUM), the number of bidders (BD_Num), the price difference between the first and second bidders (Index_5), and the M2 currency volume(M2).

As shown in Figure 5, the adjusted R-squared is 0.774, and all independent variables are statistically significant. One notable point is that in the regression model for the entire period, the coefficient for Index_5 (the price difference between the first and second bidders) is 0.039, indicating that it has a relatively small impact on the winning bid rate compared to the other independent variables.

Next, let's identify the structural break point through the Chow test. As shown in Figure 6, it can be statistically confirmed that a structural break occurred at point 226 (Q2 2016).

Let’s divide the data into two regression models based on the structural break at point 226 and examine whether there is a significant change in Index_5(bubble index) before and after the structural break.

As seen in Figure 8, Index_5 was not rejected at the 0.05 significance level before the structural break, but after the structural break, Index_5 (t-stat = 2.613, p-value = 0.01) became statistically significant. This indicates that the bubble index (Index_5) significantly increased at the point where the actual bubble occurred (the structural break point).

Figure 9 also shows that Index_5 began to fluctuate significantly around the structural break point at 226. After point 226, there was a notable increase in Index_5, reflecting the overheated real estate market at that time, as liquidity in the low-interest-rate environment flooded into the Gangnam reconstruction market and new apartment developments in Q2 2016.

For this reason, I argue that the price difference between the first and second bidders can be used as a bubble indicator. By utilizing this metric, we can prevent the bubble from inflating further and take action before the bubble bursts unexpectedly, leading the real estate market into a downturn.

Recently, the U.S.-China conflict has intensified. Beginning with trade restrictions on China under the Trump administration in 2018, the Biden administration, which took office in 2021, has continued to take bold steps aimed at domestic manufacturing recovery and tightening control over China. These efforts include bipartisan legislation such as the Infrastructure Investment and Jobs Act and the CHIPS and Science Act. As a result, the global industrial landscape is undergoing significant restructuring.

At one time, U.S. manufacturing was unrivaled, but it gradually declined under the pressure of "Triffin's Dilemma," a result of the dollar's status as the global reserve currency and increasing global competition. However, a recent wave of U.S.-China "decoupling" has shifted the comparative advantage of capital and labor between the two countries. As a result, U.S. manufacturing is showing signs of revival, with a sharp increase in employment. In contrast, China's once-explosive manufacturing growth is shrinking due to the aggressive trade sanctions imposed by the U.S. Meanwhile, America's IT and financial sectors are facing a wave of layoffs, bringing cold news to industry workers.

This column reinterprets the newly evolving industrial structures of the United States and China by applying the Heckscher-Ohlin theorem, an international trade theory based on comparative advantage in economics, the Cobb-Douglas production function, a mathematical tool of microeconomics, and regression analysis from statistics.

Heckscher-Ohlin Theorem and Cobb-Douglas Production Function

Before examining the U.S.-China conflict, it is essential to introduce the theoretical tools that will aid in its interpretation. First, the Heckscher-Ohlin theorem is a theory in international economics which states that trade occurs between countries due to differences in their factor endowments and the varying factor intensities required to produce different goods. The theorem builds upon David Ricardo's theory of comparative advantage but introduces a slightly different perspective. While Ricardo’s theory suggests that comparative advantage arises from differences in technological capabilities, the Heckscher-Ohlin theorem asserts that comparative advantage and trade stem from differences in factor intensities, such as labor (L) and capital (K), which are used in the production process. For example, China has an abundance of labor, while the U.S. is rich in capital and technology. As a result, the two countries trade labor-intensive goods for capital-intensive goods.

The Heckscher-Ohlin theorem can be expressed in the form of a Cobb-Douglas production function, which is commonly used in economics. This function illustrates the relationship between factor intensities (L for labor, K for capital) and output (Y) and is widely used to analyze how the inputs of labor and capital contribute to factor productivity in a country or a specific industry. It helps to determine how much productivity is generated when labor and capital are employed in production.

Estimating the "Elasticities" of the Cobb-Douglas Production Function through Regression Analysis

Let's formulate the Cobb-Douglas production function for a country, where the total output is denoted by $Yi$, labor input by $Li$, capital input by $Ki$, and the remainder of the output that cannot be explained by labor and capital (the residual) is represented by $exp(u)$, as follows:

$Y = exp({\beta_ 0}) \cdot L^{\beta_ L} \cdot K^{\beta_K} \cdot exp(u) \quad \cdots ~(1)$

In the Cobb-Douglas production function, "elasticities" or "factor productivities" are represented by $\beta_ L$ and $\beta_K$, and these can be easily estimated through a slight transformation of the equation. First, by taking the logarithm of both sides of equation 1, it transforms into equation 2.

$\log{Y} = \beta_ 0 + {\beta_ L} \log{L_ i} + {\beta_ K} \log{K_i} + u \quad \cdots ~(2)$

Equation 2 takes a familiar form. It is a linear regression equation where $\log{Y}$ is the dependent variable and $\log{L}$ and $\log{K}$ are the independent variables. By taking the partial derivatives of $\log{L}$ and $\log{K}$, we can obtain $\beta_ L$ and $\beta_ K$. In microeconomics, these regression coefficients are referred to as the "elasticities of substitution" for the factors of production, denoted as $e_ {LY}$ and $e_ {KY}$.

$$
\beta_ L = \cfrac{\partial \log{Y}}{\partial \log{L}} = \cfrac{dY}{Y}*\cfrac{L}{dL} = \cfrac{dY/Y}{dL/L} = e_{LY}
$$

$$
\beta_ K = \cfrac{\partial \log{Y}}{\partial \log{K}} = \cfrac{dY}{Y}*\cfrac{K}{dK} = \cfrac{dY/Y}{dK/K} = e_{KY}
$$

By using OLS (Ordinary Least Squares) estimation in equation 2, we can get $\beta_ L$ and $\beta_K$, allowing us to quantitatively analyze how elastically the output of a given country increases when additional labor and capital are input.

Trade Arising from Differences in the Regression Coefficients

In summary, by estimating the $\beta_ L$ and $\beta_K$ of the Cobb-Douglas production function, we can quantitatively assess the labor and capital productivity of each country. Based on this Cobb-Douglas production function, the reasons for trade between countries can be explained using the Heckscher-Ohlin theorem.

For example, let’s assume that South Korea is relatively capital-abundant, while Chile is labor-abundant. In this case, smartphones are considered capital-intensive goods, requiring more capital than labor, while wine is labor-intensive, requiring more labor than capital. Thus, in the production of smartphones, $\beta_ K$ would be greater than $\beta_ L$, whereas in the production of wine, $\beta_ L$ would be greater than $\beta_ K$. According to the Heckscher-Ohlin theorem, due to differences in factor intensities ($\beta_ L$, $\beta_K$), South Korea would have a comparative advantage in producing capital-intensive smartphones at a lower cost, while Chile would have a comparative advantage in producing labor-intensive wine at a lower cost, leading to trade.

At this point, we have all the necessary tools for analysis. However, before diving in, let's first take a look at the recent shifts in the industrial landscape of the U.S. and China due to their ongoing conflict

Increase in U.S. Manufacturing Employment Rate

The U.S. export-based industries, particularly manufacturing, have consistently recorded trade deficits and gradually declined due to the petrodollar system that began in the mid-to-late 1970s. This decline accelerated further during the 1980s under Ronald Reagan's administration, as the consecutive oil shocks triggered domestic economic recessions, leading to a steep downfall of labor-intensive manufacturing industries. To counter this, the government implemented tax cuts, public spending, and massive defense expenditures to curb the recession dramatically. Subsequently, the U.S. experienced nearly 40 years of low interest rates and low inflation, and until the onset of the COVID-19 pandemic, the manufacturing employment rate had been steadily recovering.

However, in the wave of the COVID-19, the economy contracted once again, leading to the overnight shutdown of major automobile assembly plants and dealerships. Automobile manufacturing came to a sudden halt, and in the food industry, numerous reports emerged of workers contracting and dying from the virus, prompting factories to shut down and causing a massive loss of jobs. In response, the U.S. government implemented a large-scale quantitative easing of $4.5 trillion, which helped boost employment rates, and the manufacturing sector began to recover.

Amidst these developments, the U.S., under the leadership of President Biden, aggressively pushed forward the Bipartisan Infrastructure Law (BIL) and the CHIPS and Science Act. These initiatives aimed to strengthen the domestic economy by overhauling bridges, roads, and rural areas across the country. As a result, not only did the U.S. curb the growth of China's semiconductor industry, a key sector of the Fourth Industrial Revolution, but it is also now moving to reshape the global semiconductor landscape with the U.S. at its center.

Buoyed by this political momentum, the manufacturing sector has continued its upward trend. According to the Financial Times, as of June, U.S. manufacturing employment has increased by nearly 800,000 since President Biden took office, with around 13 million people now employed in the sector, the highest level since the 2008 global financial crisis. As shown in Figure 1, employment rates rose sharply during key periods such as the pandemic-era quantitative easing and the passage of the Bipartisan Infrastructure Law and CHIPS and Science Act.

The Grim U.S. Financial and Tech Sectors

In stark contrast to the manufacturing sector, which is gearing up for significant growth, major U.S. tech companies have been undergoing large-scale layoffs since the latter half of last year. Leading the wave of employment cuts are global IT giants such as Amazon and Meta, with many other U.S. tech firms following suit. According to Layoffs.fyi, a site that tracks layoffs in the U.S. tech industry, 1,058 IT companies laid off a total of 164,709 employees last year alone. Notably, in November, Amazon laid off 10,000 employees, while Meta cut 11,000 jobs.

The layoffs at major tech companies have continued into this year. Following last year's cuts, Amazon has laid off an additional 17,000 employees so far in 2023. While Apple announced plans to reduce its operating budget to avoid restructuring, some reports suggest that the company began cutting staff in April, starting with its retail team at the U.S. headquarters in California.

Meanwhile, the U.S. financial sector has not been spared from the wave of layoffs. In May, CNBC reported that major Wall Street investment banks such as Morgan Stanley, Bank of America, and Citigroup carried out significant job cuts. Morgan Stanley laid off around 3,000 employees by the end of June, amounting to 5% of its total workforce based in New York. Additionally, Citigroup and Bank of America also announced the dismissal of hundreds of employees in May, bringing grim news to the financial industry.

Declining China's Manufacturing Sector Amid U.S.-China

China, which is in direct hegemonic competition with the U.S., is experiencing significant losses in its manufacturing sector. This is due to the intensifying "decoupling" between the two countries, driven by the Biden administration’s domestic-focused initiatives such as the Bipartisan Infrastructure Law and the CHIPS and Science Act, which aim to strengthen U.S. technological sanctions on China. Additionally, ongoing tariff disputes between the two nations have further escalated the situation.

China's manufacturing growth has significantly slowed due to geopolitical tensions with the U.S. As shown in Figure 2, China's secondary industry (related to manufacturing) began to decline starting in 2021, when the U.S. started implementing parts of the Bipartisan Infrastructure Law as part of its efforts to strengthen domestic competitiveness and achieve self-sufficiency.

In the same context, data released by China’s General Administration of Customs on July 13 showed that China's export value in June was $285.3 billion, a 12.4% decrease compared to the same month last year. This figure falls short of both the previous month's -7.5% and the market expectation of -9.5%.

An Explanation of the U.S.-China Conflict Based on the Heckscher-Ohlin Theorem

Through the discussion so far, we can see that between 2021 and 2022, when national efforts toward "self-sufficiency" and "decoupling" between the U.S. and China began to intensify, the two countries started to take divergent paths in manufacturing. Additionally, we've observed that U.S. sectors such as Wall Street and Big Tech are facing a wave of layoffs, leading to a general downturn in the financial and IT industries.

Here, we can raise a few questions. Why have the industrial structures of the U.S. and China evolved in opposite directions? For instance, why have the Bipartisan Infrastructure Law and the CHIPS and Science Act, aimed at reducing U.S. dependence on China, revitalized U.S. manufacturing while leading to the decline of China's manufacturing sector? Could U.S. and Chinese manufacturing not have risen together when the U.S. decided to focus on strengthening its domestic market? And why, seemingly unrelated, has the U.S. financial and IT sectors faced a downturn?

Let’s view the U.S.-China conflict and industrial structural changes through the lens of the Heckscher-Ohlin theorem. From 1970 to 2022, the U.S. experienced a decline in its export competitiveness in manufacturing due to the dollar's status as the global reserve currency and instead focused on capital-intensive industries like finance and IT. It’s also important to acknowledge that the flow of global capital into the U.S. as a reserve currency issuer played a significant role in this shift. During this period, the U.S. had a comparative advantage in "technology" or "capital," whereas China, by leveraging its cheap labor, focused on labor-intensive industries based on primary sectors, giving it a comparative advantage in "labor." By concentrating on what they did best, the U.S. and China grew their economies by trading technology-intensive goods and labor-intensive goods with one another.

However, with the recent passage of laws such as the Bipartisan Infrastructure Law and the CHIPS and Science Act, the U.S.-China technological competition has intensified, leading to a resurgence in U.S. manufacturing. Additionally, as labor costs have declined in the U.S., the comparative advantage between the U.S. and China has started to shift. With the relative price of labor compared to capital decreasing in the U.S., resources have gradually been reallocated toward manufacturing. As a result, capital-intensive industries like finance and IT have seen a reduction in labor input. In contrast, China’s labor-intensive manufacturing sector has been shrinking due to aggressive U.S. sanctions, causing the relative price of labor to rise compared to capital.

Changes in the U.S.-China Industrial Structure Due to Differences in the Regression Coefficients

Let's revisit the U.S.-China conflict through the lens of factor productivity (elasticity) in the Cobb-Douglas production function. The total output of a country is denoted by $Y_ i$, labor input by $L_ i$, capital input by $K_ i$, and the remainder of the output that cannot be explained by labor and capital (the residual) by $exp(u)$. The Cobb-Douglas function can be expressed as follows. In this case, let’s assume there are only two countries, where $i$ represents either the U.S. ($U$) or China ($C$).

$$
Y = exp(\beta_ 0) \cdot L^{\beta_ L} \cdot K^{\beta_K} \cdot exp(u)
$$

Before the implementation of the petrodollar system and the resulting large trade deficits, the U.S. in the pre-1980s likely had higher values for both $\beta_ L^U$ (labor elasticity) and $\beta_K^U$ (capital elasticity) compared to China. The reason is that the U.S., with its robust manufacturing sector focused on secondary industries, would have had higher labor productivity than China, which was still primarily focused on the primary sector (light industries) during its early stages of economic growth. Additionally, the U.S. had significantly increased productivity through capital investments in mechanization and automation, suggesting that its capital productivity was also higher than that of China, which at the time relied more on manual labor.

On the other hand, from the 1980s until just before the COVID-19 outbreak, the U.S. saw its manufacturing sector decline under the petrodollar system, while China experienced explosive growth due to continuous market reforms and its entry into the WTO. As a result, in the manufacturing sector, both $\beta_ L^U$ (labor elasticity) and $\beta_ K^U$ (capital elasticity) in the U.S. likely became lower than those of China during this period.

However, between 2021 and 2022, as the U.S. took aggressive measures to counter China, its manufacturing sector began to revive. As a result, in manufacturing, the U.S.'s $\beta_ L^U$ (labor elasticity) and $\beta_ K^U$ (capital elasticity) are now catching up with those of China.

Applying the Heckscher-Ohlin theorem, it appears that due to the recent U.S.-China conflict, the U.S. has not only gained an absolute advantage in both $\beta_ L^U$ (labor elasticity) and $\beta_ K^U$ (capital elasticity) over China, but there are also signs that $\beta_ L^U$ may be surpassing $\beta_ K^U$. This indicates that labor is shifting toward the manufacturing sector in the U.S., while in China, trade sanctions are causing $\beta_ L^C$ (labor elasticity) to decline, making $\beta_ K^C$ (capital elasticity) relatively larger.

The Gordon School of Business (GSB) is the educational branch of the Swiss Institute of Artificial Intelligence (SIAI). While SIAI now serves as a hub for global conferences and high-level research, GSB continues the Institute’s original educational mission: training the next generation of thinkers capable of applying deep statistical reasoning to real-world complexity. This Yearbook offers a rare glimpse into that mission in action—not through official reports or faculty summaries, but through the voices of the students themselves.

A Window Into Transformation

From the outset, GSB’s intent in producing this Yearbook was simple yet ambitious: to let students tell the story of how their thinking has evolved over the course of their studies. We hoped to capture a vibrant mosaic of personal growth—moments of struggle, breakthroughs, and even the emotional highs and lows that inevitably accompany rigorous academic training.

In reality, many students responded with something more restrained. Rather than dramatic narratives, they chose quiet reflection, focusing on the substance of their dissertations and the precision of their methods. To some, this may seem understated. Yet to those familiar with the transformation our students undergo, these reserved voices speak volumes. Their very restraint signals a maturity born of rigorous training: an ability to think deeply, write clearly, and ground ideas in evidence rather than flourish.

Beyond Code and Models

One common misconception about our program is that it produces graduates armed solely with lines of code and algorithms. This Yearbook aims to challenge that view. The public often equates “artificial intelligence” with mystical black-box technologies. What our students discover, and what these pages reveal, is that AI at its core is computational statistics—tools and frameworks that unlock patterns in data and transform how we understand the world.

By the end of the program, students move beyond chasing models for their own sake. They learn to ask deeper questions: What is the cause behind this pattern? How does this insight translate into real-world impact? How can rigorous theory coexist with the messy realities of human systems? The diversity of topics in this Yearbook—spanning finance, health, policy, and beyond—reflects not only the breadth of AI’s applications but also the versatility of the thinkers GSB aims to cultivate.

The School’s Philosophy

GSB is founded on the motto Rerum cognoscere causas—“to seek the causes of things.” This principle runs through every lecture, seminar, and case study. Our teaching begins with theory: statistical frameworks, mathematical reasoning, and computational rigor. But theory alone is never enough. Students are challenged to apply their knowledge to real-world data, confront ambiguous problems, and produce outcomes that are both scientifically robust and practically relevant.

This philosophy shapes a particular kind of graduate. Our alumni are not limited to any single industry or role; they are versatile minds who can move seamlessly between disciplines, bridging technical insight and strategic thinking. In an era where data touches every domain, this breadth is not optional—it is essential.

Reading the Yearbook

As you read this Yearbook, you will notice that the tone is often modest. These are not glossy marketing pieces. They are personal reflections on work that was, for many students, the most demanding intellectual challenge of their lives. Some entries may feel technical; others may feel restrained. But beneath the surface lies something remarkable: evidence of how much their thought processes have changed.

A student who once saw AI as code now sees it as a lens for understanding reality. A student who once sought quick answers now wrestles with root causes. A student who once worked in isolation now frames their work in conversation with society. These subtle shifts mark the true success of GSB’s educational model.

A Quiet Invitation

In sharing this Yearbook, we do not simply showcase dissertations or celebrate milestones. We invite you—whether you are an academic, an industry professional, or simply curious—to witness the growth of minds in motion. This is what a year of intense study at GSB produces: not just technical skill, but a way of thinking that endures long after the program ends.

We hope you read these pages not only for their content, but for what they represent: a generation of students who have learned, in their own quiet ways, to seek the causes of things.