Lee Jaewon on Leaving the Bank of Korea: "In the AI Era, Korea's Key Asset Is 'Data'... Semiconductor Boom Is the Best Time for Structural Reform" [Interview]

Semiconductors Require "Technological Edge and Supply Chain Dominance" for Sustainable Income
Boom Benefits Could Be Concentrated Among Large Corporations and Highly Skilled Workers
"Income and Time Earned in the Semiconductor Boom Should Be Use

From the competitive college admissions environment and life-sustaining treatments to corporate innovation and, most recently, artificial intelligence (AI), the Bank of Korea Economic Research Institute led by Lee Jaewon has been at the center of South Korea’s structural reform debate for the past three years. Lee’s sense of urgency about the “golden time” needed to solve issues such as low birth rates, aging society, and productivity slowdown has been the driving force behind this. As a result, in the past three years, the scope of research at the Bank of Korea Economic Research Institute has expanded from analyzing short-term economic cycles and monetary policy to diagnosing structural problems the South Korean economy faces and proposing concrete solutions. Through this process, the research capabilities of the institution’s staff have been strengthened, and its external research network has also expanded.

In an interview at the Bank of Korea on July 30, Lee stated, “If our work as a think tank has contributed to enriching public debate and providing solutions to our economic issues, that is the most rewarding achievement.” He credited the research team who conducted individual projects for these advancements. Expressing his gratitude to those who supported and encouraged these efforts both internally and externally, he humbly remarked, “I was fortunate.”

Lee said he feels a deep sense of regret as he prepares to leave his post, but added, “The greatest asset I gained from my time at the Bank of Korea is not answers, but questions.” He hopes to further develop the questions he encountered in policy circles into rigorous academic research. Here, Lee shared his thoughts and ambitions regarding the present and future of Korea’s economy.

Lee Jaewon, President of the Bank of Korea Economic Research Institute, emphasized in an interview with The Asia Business Daily, "It is difficult to promote structural reforms during a crisis, and the capacity to prepare buffer measures is also limited." He added, "Rather, the best time for structural reform is during normal times, especially when there is room for maneuver due to external booms such as in semiconductors."

"Semiconductor Boom: The Perfect Time for Structural Reform"

It is impossible to discuss the current South Korean economy without mentioning semiconductors. Lee emphasized that, in this phase of semiconductor-driven growth, it is crucial to consider: ▲ To what extent can this boom be converted into sustainable income? ▲ How broadly will semiconductor profits be distributed throughout the domestic economy via investment, employment, and consumption? ▲ And how widely will these benefits be shared across companies, industries, and different social groups?

He explained, “The significant increase in real national income in the first half of this year was due to improved terms of trade as semiconductor prices rose, meaning the same production volume generated more income.” While this is a positive development, he cautioned that the income gains from price increases can be temporary. In order to convert these into structural income, he explained, it is necessary to maintain technological advantages and supply chain dominance.

Korean companies currently possess pricing power in high value-added semiconductors such as high bandwidth memory (HBM), thanks to their quality, yield, and large-scale supply capacity. Lee stated, “These advantages are not permanent but provide a temporary window of opportunity.” He stressed, “Before competitors and latecomers catch up, Korea must move beyond simple capacity expansion and volume competition, and solidify its leadership in fields such as design, next-generation memory, low-power AI semiconductors, and materials, parts, and equipment.”

He also noted the need to be mindful of the fact that the benefits of the boom may be relatively concentrated among large corporations, highly skilled workers, and employees in related firms. Because semiconductors are a capital and technology-intensive industry, even if production and export volumes surge, employment does not necessarily rise at the same rate. Lee observed, “Macroeconomic indicators may improve rapidly, but some households, the self-employed, and those outside the Seoul metropolitan area may not fully experience the economic recovery.” He said, “Income and time gained from the semiconductor boom must be channeled into innovation among small and medium-sized enterprises and the service industry, into workforce development, and into the revitalization of the regional economy and domestic demand.” Only then, he argued, can this boom become more than just a temporary surge in exports and instead lay the groundwork for enhancing productivity across the entire economy.

Lee emphasized, “It is difficult to promote structural reforms during a crisis, and the capacity to prepare buffer measures is also limited.” He added, “Rather, the best time for structural reform is during normal times, especially when there is room for maneuver due to external booms.”

Director Lee stated, "The nationwide health insurance data and vast hospital clinical data can serve as the foundation for creating new growth engines that combine the bioindustry and AI," adding, "We need to establish a system that standardizes and securely links different data so that companies and researchers can practically utilize it."

"Combining Korea’s Manufacturing Data and AI Will Secure Competitiveness"

As a key asset that will determine competitiveness in the AI era, Lee cited “high-quality data.” He explained, “Korea is a manufacturing powerhouse, with abundant data accumulated on production floors.” He stressed, “In addition to capabilities in stable mass production, meticulous process management, skilled technical personnel, and a robust ecosystem for materials, parts, and equipment, Korea also has the agility to swiftly integrate new technologies into products and processes.” By combining manufacturing data with AI, he argued, Korea can achieve differentiated competitiveness in “physical AI” sectors such as robotics, autonomous driving, and smart factories.

He suggested that a similar approach could be taken in service industries where high-quality data has also been amassed, such as healthcare, finance, logistics, and content. A prime example is the medical and bio sectors. Lee stated, “Korea has accumulated nationwide health insurance data and vast hospital clinical data through its single-payer health insurance system.” He added, “This is a nationally valuable asset with global rarity and can become the foundation for creating new growth engines by combining the bioindustry with AI.”

Looking ahead, Lee argued that simply gathering more data is not enough. Instead, there is a need to establish a system that standardizes and securely links different types of data so companies and researchers can practically utilize it. He also called for achieving a balance between data protection and data utilization, as well as simultaneously expanding AI talent and computing infrastructure. Ultimately, Lee diagnosed, Korea’s real strength lies in its capacity to leverage the extensive data and on-the-ground experience of traditional industries by integrating them with AI to create new industrial competitiveness.

Director Lee emphasized, "AI is a complementary means to adapt to population decline," stating, "Expanding labor market participation among women and the elderly, extending healthy life expectancy, labor mobility and re-education, service industry innovation, revitalization of the regional economy, and reforms in pensions, healthcare, and finance must be pursued together to effectively increase potential growth rates."

Declining Potential Growth Rate: The Key to a Turnaround is the "Spread of AI Productivity Effects"

Lee assessed that the degree to which AI can offset Korea’s declining potential growth rate depends on the actual productivity impact of AI. He emphasized, “AI’s productivity effect will depend on how much the scope of work it can actually perform expands, how quickly companies—including small and medium-sized firms—adopt it, how work procedures and organizational structures are innovated, and whether AI is used to complement human capabilities rather than simply replace humans.” While time savings and productivity improvements are already being seen at the task level, he observed that, as with previous general-purpose technologies, it will take considerable time before these gains translate into broad-based productivity improvements for companies and the economy as a whole.

Lee stated that, while AI has the potential to significantly ease the long-term economic shock of population decline, it will not be able to resolve all challenges that arise during the transition period. This is because, before AI is fully diffused, the labor force may diminish first, and the retirement of skilled workers along with a decrease in new entrants could lead to skill shortages across industries and occupations. In particular, in fields where physical presence or manual labor is essential—such as caregiving, healthcare, construction, transportation, and food services—current forms of generative AI are limited in addressing labor shortages.

He said, “Spending on pensions, healthcare, long-term care, and demand for caregiving will keep rising as the population ages. Ultimately, AI is not a substitute to solve population decline, but an extremely important complementary tool for adapting to it.” He further explained that, along with the spread of AI, measures such as expanding labor market participation among women and older adults, extending healthy life expectancy, promoting labor mobility and retraining, innovating the service industry, revitalizing the regional economy, and reforming pensions, healthcare, and finances must all be pursued in tandem to truly enhance potential growth rates.

"AI Can Sequentially Deliver Demand, Supply, and Financial Shocks to the Macroeconomy"

Lee analyzed that, in the shorter term, AI could generate three distinct shocks to the macroeconomy. Initially, an expansion in investment could create a demand shock. If adoption is successful, this would shift into a productivity shock, but if expectations become excessively high, it could eventually result in a financial shock.

He explained, “In the early stages, the impact is characterized mainly as a demand shock, as investments in data centers, semiconductors, power grids, software, and specialist personnel all rise before widespread productivity gains take hold.” Consequently, increases in facility investment, power demand, and wages for AI-related personnel could temporarily push up aggregate demand, raise production costs, and heighten inflationary pressures.

He went on to predict that a positive supply shock could follow. “If companies reorganize work procedures and structures around AI, and the technology spreads widely to service sectors and small businesses, the productivity of both labor and capital will rise.” At this stage, even if the economy grows rapidly, it does not necessarily lead to higher inflation.

The third possibility is a financial shock. Lee explained that current valuations and investments in AI-related firms already reflect substantial anticipated future productivity, so if productivity improvements are delayed or profitability falls short of expectations, stock and credit markets could undergo sharp corrections. He warned, “Particular attention should be paid to the concentration of high valuations and leverage in AI-related companies, as well as private credit, as these financial structures could intensify a market downturn.”

Director Lee said, "Our country's government debt is still at a manageable level compared to GDP," but pointed out that "aging and other factors could quickly pressure public finances." He emphasized that current financial capacity should not be seen solely as a reason for reassurance but rather as an opportunity to proactively reform pensions, healthcare, taxation, and expenditure structures.

"If Another Global Financial Crisis Occurs... Be Wary of the Government Bond Market"

Lee cautioned that crises can often originate from places considered safe. In this context, he warned that imbalances between supply and demand in the government bond market and liquidity squeezes could amplify financial instability. The foremost issue, he pointed out, is that government bond issuance is structurally increasing. Last year, global public debt reached nearly 94% of the world’s GDP. According to the International Monetary Fund (IMF), if current trends continue, global public debt is projected to hit 100% of GDP by 2029. Lee noted, “Taking into account factors such as aging populations, defense spending, climate and energy transitions, industrial policies, and rising interest charges, it will be very difficult for fiscal deficits and public debt in major economies to decline meaningfully in the short term.”

He said, “In the past, a global savings glut and large-scale government bond purchases by central banks absorbed increased bond supply. Now, with persistent fiscal deficits raising the risk of a ‘bond glut’ and diminished central bank purchases, the environment for absorbing excess supply is much weaker than before.” He warned, “This could create structural upward pressures on long-term government bond yields and term premiums.” Most importantly, since U.S. Treasury yields serve as the benchmark for global financial asset prices, their influence is not limited to the United States alone.

Lee also underscored the significance of changes in holders of government bonds. “Nonbank financial institutions such as hedge funds and investment funds rely relatively heavily on leverage and short-term funding,” he said. “If interest rates spike, these players may be forced to quickly liquidate government bonds to cover margin calls, resulting in a vicious cycle where price declines trigger further selling. Institutions that provide liquidity in normal times can actually increase volatility during a crisis.”

He also pointed out that the growth of dollar stablecoins can serve as a new linkage. “Issuers manage a substantial portion of their reserves through U.S. short-term Treasurys, Treasury-backed repurchase agreements (repos), and bank deposits. While these support government bond demand under normal conditions, if confidence is shaken and large-scale redemptions occur, issuers may be compelled to rapidly liquidate reserves, setting off a vicious cycle of ‘redemption – bond sales – price declines and liquidity deterioration – further redemption.’”

Lee noted that Korea, too, is not immune to such shocks. While fiscal risks at home are not extreme for now, a rise in U.S. Treasury yields could put upward pressure on domestic long-term and corporate bond yields. If the dollar strengthens while risk aversion grows, this could threaten exchange rates, capital flows, and the real economy. He said, “Policy space for monetary management tailored to domestic economic conditions and inflation may also become constrained.” Ultimately, he warned, “Future financial crises may not only take the traditional form of distressed risky assets contaminating safe assets, but could also manifest as instability in the government bond market—which was previously considered a safe haven—spreading throughout the financial system as a whole.”

Excessive Government Debt Also Constrains Monetary Policy

Excessive government debt can likewise restrict the scope of monetary policy. Lee cautioned, “If government debt is too high relative to GDP, any increase in interest rates will sharply raise the government’s interest burden, causing bond prices to fall and financial instability to ensue.” In such cases, the central bank might want to increase interest rates to control inflation but is hamstrung due to the financial and fiscal repercussions. “This is what’s traditionally called ‘fiscal dominance,’” he said.

There is an even more fundamental risk: when the government fails to inspire confidence that it will manage increased debt by raising taxes or reducing expenditures in the future. “If rates rise under these conditions, government interest expenses will increase further, creating a vicious cycle in which debt continues to balloon—thereby actually stoking inflation.” In such a situation, the central bank does not merely hesitate to raise rates; it may actually resist doing so because the hikes would be ineffective or even counterproductive for price stability.

Lee noted that Korea’s government debt remains manageable relative to GDP, so there is no need to be immediately concerned. However, he warned, “As aging pushes up pension, healthcare, and long-term care costs rapidly, while slower growth likely erodes the tax base, risks could build.” The National Assembly Budget Office projects that, if existing laws and systems are maintained, the national debt-to-GDP ratio will rise from 49.4% last year to 173.0% by 2072. The IMF’s separate long-term scenario predicts the ratio for the government could reach 90–130% by 2050 if reforms are not undertaken. “Although these forecasts differ on the definitions and preconditions for debt and are not set in stone, they demonstrate how quickly aging can strain fiscal resources,” Lee said. “Current fiscal capacity should not be viewed merely as reassurance, but used proactively as an opportunity to reform pensions, healthcare, taxation, and the expenditure structure in advance.”

Director Lee said, "There is a need to analyze vulnerable groups in detail based on debt, liquid assets, consumption patterns, housing types, age, and employment status," adding, "We need to look not only at 'who is hurt by austerity' but also at 'who suffers more if inflation is not controlled.'"

"Segmenting Vulnerable Groups to Analyze Inflation and Interest Rate Impacts"

In mid-August, Lee will leave the Bank of Korea and return as a professor at Seoul National University. Upon his return, he plans to analyze in detail who bears the greatest burden from inflation and interest rate changes. Raising interest rates in response to higher inflation hits households with high debt and unstable income particularly hard. However, if high inflation is simply left unaddressed, the most vulnerable—those with a high proportion of spending on necessities and few real assets to hedge against price increases—stand to suffer even more. Lee said, “In Korea as well, rather than treating the vulnerable as a homogeneous group, I think it is necessary to segment them by debt, liquid asset holdings, consumption structure, housing type, age, and employment status—and I have become interested in such analysis.” He emphasized, “We need to look not only at ‘who is hurt by austerity,’ but also at ‘who is even more hurt if inflation is not controlled.’”

He is also interested in studying how AI will change the way companies set prices and how monetary policy functions. “As AI spreads, more companies will be able to analyze market conditions in real time and adjust prices more frequently, making prices across the entire economy more flexible than in the past,” he said. “This could cause changes in interest rates to affect prices more rapidly than production or employment.” The sacrifice of growth and employment necessary to stabilize prices could be reduced, but at the same time, even small policy errors might show up more quickly in inflation data. Lee remarked, “I want to examine what such changes mean for the trade-offs among policy goals and the framework for implementing monetary policy.”

He also identified the role monetary policy should play during periods of structural transformation—when the industrial structure is permanently changing—as an important topic for study. AI, population aging, carbon neutrality, and supply chain restructuring are not simply temporary fluctuations but causes of long-term, structural changes that move people and capital across industries. “These transformations do not happen overnight, and in the process, production, employment, and prices may move in complex ways. Monetary policy cannot determine the final direction of economic structure, but how it affects the speed, costs, and dynamics of this transformation is an area that requires important future research,” Lee concluded.

Lee added, “(When I return to university) I want to be a professor who helps students view real-world issues through an economic lens and develop their own new questions, both in education and research.” He also plans to complete a revised edition of the macroeconomics textbook Macroeconomic Theory, co-authored with Professors Unchan Jung and Youngsik Kim. “We are already overdue on the revision, and I feel sorry to my co-authors for the delay while I was at the Bank of Korea. Nonetheless, by incorporating my central bank experiences with monetary and economic policy into the book, I think I can make the content richer and more vivid than before,” he said with a smile.

About Lee Jaewon

Born in 1975, Lee graduated from Eonnam High School and received his bachelor’s degrees in economics and mathematics from the University of Wisconsin, followed by his master’s and Ph.D. in economics from Princeton University. After serving as a professor at Rutgers University and the University of Virginia, he became a professor in the economics department at Seoul National University. In September 2023, he was appointed President and Chief Economist at the Bank of Korea Economic Research Institute. Lee has been a visiting scholar at the Federal Reserve Bank of St. Louis, and he currently serves as an external research fellow at the Dallas Fed, a committee member of the International Journal of Central Banking (IJCB), and a board member of the Central Bank Research Association (CEBRA).

This content was produced with the assistance of AI translation services.