In equilibrium markets where supply and demand determine transaction volume and price, a strong increase in demand tends to raise both, but when the increase originates from supply, transaction volume expands while price declines. This pattern holds in funding markets: rising borrowing demand (fund demand) increases both credit volume (debt) and interest rates (credit price) at equilibrium, whereas expanding fund supply through higher household savings can increase credit volume even as borrowing rates fall (Figure 3). Over the past two decades, interest rates have continued to decrease while debt among households, firms, and government has steadily risen, indicating that fund supply expansion outweighed demand growth.

Debt for one party is an asset for another, and borrowing requires a counterparty to defer consumption and supply funds. In parallel, the sustained rise in household debt coincides with the expansion of household asset accumulation. When economic conditions strengthen incentives for asset accumulation, the supply curve of funds shifts to the right, lowering interest rates even as economy-wide debt rises. This study identifies rising life expectancy as a primary driver of stronger incentives for household asset accumulation and the consequent expansion of the supply of funds. It further examines the structural underpinnings of the long-term coexistence of falling interest rates and growing household debt.

Ⅱ. Theoretical Analysis of the Drivers of Household Debt

The existing literature on sustained growth in household debt has largely focused on disparities in income and wealth. Mian et al. (2021) find that households with greater net wealth have lower marginal propensities to consume (MPC). As wealth inequality widens, highwealth households save a larger share of their income, expanding the supply of loanable funds, which low-wealth households borrow, driving up household debt. This framework is particularly effective in explaining the U.S. situation before the global financial crisis. However, in the past decade, household debt has continued to grow in Korea and other major economies despite little change in income and wealth inequality, suggesting that structural drivers beyond inequality warrant closer examination.

Figure 4 illustrates how household debt-to-GDP ratios vary with remaining life expectancy (life expectancy – median age) and oldage dependency ratios. Panel (a) shows a clear positive correlation between longer remaining life expectancy and higher household debt-to-GDP ratios. Remaining life expectancy increases when life expectancy rises or median ages fall. By contrast, Panel (b) shows a negative correlation between the old-age dependency ratio, defined as the ratio of the elderly population (aged 65 and over) to the workingage population (aged 15–64), and the household debt-to-GDP ratio. These correlations remain clear when household debt is measured relative to disposable income (see Appendix Figure 1).

These findings suggest that demographic factors, such as rising life expectancy and shifts in population age structure, can meaningfully shape household debt trends over the medium and long term. Building on prior research that emphasized household inequality, this study examines how these demographic changes affect the household debt trajectory and provides projections based on these factors.

Rising life expectancy reinforces households’ incentives to accumulate assets, also driving up household debt through age-specific patterns of asset accumulation. Households typically seek to smooth consumption over their lifetime, adjusting spending in response to income variations at different stages in life. With life expectancy increasing while the retirement age from primary jobs with stable income flows remains largely unchanged, working-age cohorts tend to reduce present consumption and build financial assets in preparation for extended retirement years. Retired older cohorts also tend to preserve accumulated assets by reducing spending over these extended periods.

Since the 1980s, Korea’s life expectancy has steadily increased by 0.4 years annually through the 2000s—twice the OECD average (0.2 years)—alongside its rigid retirement age from primary employment (Figure 5). The resulting longer retirement phase has strengthened incentives for households to build assets. Based on the Survey of Household Finances and Living Conditions, net assets were calculated by age group, normalized to 2024 values using the ratio of nominal GDP in 2024 to that in 2014. The results show broad increases in net assets across most age groups over the past decade, with the trend led by older cohorts (Figure 6).

Explaining the rising household debt requires considering not only stronger incentives for asset accumulation but also heterogeneity across households. Debt arises only when some households supply funds while others borrow, even within the same macroeconomic environment. This study examines one key dimension of heterogeneity—variations across age groups. While rising life expectancy strengthens the incentive to accumulate assets, asset portfolios vary by age. Household assets are broadly categorized into financial assets, which are generally liquid with low transaction costs, and real assets, primarily housing, which require substantial time and costs to trade.

Older cohorts tend to have higher homeownership rates than younger and middle-aged cohorts. Middle-aged and older cohorts, already homeowners with shorter remaining years, tend to favor financial assets over housing, which is costlier to trade. Their tendency increases demand for financial assets and expands the supply of loanable funds, which may place downward pressure on interest rates.

By contrast, younger cohorts with longer life horizons show stronger demand for housing assets. Their lower sensitivity to transaction costs, combined with the ability to enjoy housing services over an extended period, makes long-term homeownership a more attractive prospect. To acquire housing, they typically borrow large sums upfront, repaying principal and interest over time while gradually building real assets. In effect, household debt arises as younger cohorts borrow from older ones to finance housing purchases. While homeownership is a form of durable consumption, it also serves as a form of savings, as homes can later be liquidated to fund consumption in retirement.

Figure 7 shows net financial assets by age of the household head. Net financial assets are defined as financial assets minus debt,4) representing net worth excluding real assets. Contrary to the broad increase in net assets across most age groups amid rising life expectancy, net financial assets tend to decline for groups under age 45 but increase for those aged 50 and above. This suggests that, as life expectancy has increased, older cohorts tend to accumulate primarily financial assets, while younger cohorts expand housing assets through debt. Put differently, rising life expectancy has produced different patterns of asset accumulation across age groups. Younger cohorts have driven debt growth, borrowing funds supplied by older cohorts to finance housing purchases. National data on household debt comparing 2013 and 2023 show that real per capita debt rose among younger cohorts but declined among older ones (Figure 8).

Beyond rising life expectancy, changes in the population age structure can influence household debt trends. Households make borrowing and saving decisions to smooth consumption―including housing services―over the life cycle as income varies with age. Household income is typically low in early adulthood and rises gradually, peaking around the mid-50s before declining. Despite this income pattern, households in their 30s and 40s borrow against future income to smooth consumption, gradually repaying debt as earnings rise while preparing for retirement.

These lifecycle behaviors imply that younger and older households play distinct roles in the supply and demand for funds. Cohorts aged 50 and above typically serve as lenders, while younger cohorts are primarily borrowers. Aggregate demand for savings and borrowing moves in tandem with shifting age distribution. When the age profile is skewed toward the 30s and 40s, borrowing demand rises and household debt expands. By contrast, in aging societies with a growing share of people in their 60s and 70s amid persistently low fertility, borrowing demand weakens and household debt tends to decline.

Ⅲ. Empirical Analysis of the Drivers of Household Debt

This section presents empirical findings on the theoretical frameworks outlined earlier, drawing on panel data from Statistics Korea’s Korean Statistical Information Service (KOSIS), the Bank of Korea’s Economic Statistics System (ECOS), Eurostat, the OECD Data Explorer, the United Nations, the World Inequality Database, and the IMF’s Integrated Macroprudential Policy (iMaPP) Database. It also provides projections on Korea’s household debt-to-GDP ratio.

The analysis constructs a panel dataset of 35 countries, primarily OECD and EU members, covering life expectancy and population shares in five-year age groups. To quantify the impact of shifting age distribution on the household debt-to-GDP ratio, a quadratic regression model incorporates these population shares to reflect lifecycle patterns of saving and borrowing, where younger cohorts tend to act as demanders of funds and older cohorts as suppliers.

The analysis also incorporates their Gini coefficients for net wealth, following Mian et al. (2021), to capture the impact of rising wealth inequality on household debt. While key controls include changes in macroprudential regulations, the model uses a fixed-effects specification with country fixed effects and dummy variables for the global financial crisis and the COVID-19 pandemic. Robustness tests with additional covariates, including population density, real per capita GDP growth, consumer price inflation, and time trend, confirm that the results remain stable regardless of their inclusion.

Table 1 shows the estimated effects of these variables on the ratio of household debt to nominal GDP. The results indicate that rising life expectancy significantly increases the ratio, while a declining share of younger cohorts and an aging population structure meaningfully reduce it. Specifically, a one-year increase in life expectancy is estimated to raise the household debt-to-GDP ratio by approximately 4.6%p. A 1%p decrease in the share of the population aged 25–44 and a corresponding 1%p increase in the share aged 65 and over is associated with a 1.8%p decline (Figure 9). Moreover, consistent with previous studies, the impact of wealth inequality, as measured by the Gini coefficient for net wealth, is also estimated to be statistically significant: a one standard deviation increase in the Gini coefficient raises the ratio by roughly 4%p.

These estimates reveal that key explanatory variables account for the 33.8%p rise in Korea’s household debt-to-GDP ratio between 2003 and 2023, with 28.6%p explained by rising life expectancy and 4.0%p by shifts in age distribution. The dominant life expectancy effect reflects a 6.2-year gain (77.3 → 83.5) over the two decades. Other factors had only limited effects. A modest increase in wealth inequality (a 0.26 standard deviation rise in the net wealth Gini coefficient) raised the ratio by 1.0%p, while significantly tighter macroprudential regulations reduced it by 2.3%p. Taken together, these results suggest that structural demographic changes have been the primary drivers of Korea’s household debt growth.

However, projections based on life expectancy and age profile suggest that Korea’s household debt-to-GDP ratio will peak in the coming years before trending downward (Figure 11). This outlook reflects two structural forces as life expectancy gains slow—a declining share of younger cohorts with demand for new debt and a rising share of older cohorts drawing down accumulated assets to sustain consumption.

These patterns are likely to become more pronounced over the medium to long term. According to Statistics Korea’s Population Projections, life expectancy will gain 6.4 years (84.5 → 90.9) by 2070, pushing the ratio up by 29.5%p, whereas age distribution shifts due to population aging will reduce it by 57.1%p. On balance, these structural factors are projected to lower the ratio by 27.6%p by 2070 relative to current levels.

Ⅳ. Conclusions and Implications

Korea’s rapid gains in life expectancy, outpacing other major economies, have substantially strengthened household incentives to accumulate assets. As life expectancy rises, older cohorts tend to build financial wealth, while younger cohorts borrow these funds to acquire housing. These incentive dynamics, together with variations in portfolio patterns across age groups, have been the primary drivers of Korea’s household debt growth.

Concurrently, persistent low fertility is accelerating population aging. If the older population share expands under stagnant life expectancy, the economy’s capacity to supply funds is likely to weaken, while a shrinking younger population reduces household borrowing demand. As a result, household debt is likely to enter a gradual downward trajectory, and simulation results indicate that the household debt-to-GDP ratio will peak within the next few years before trending down.

These findings offer important policy implications. Foremost, household debt dynamics are shaped not only by financial policies but also by nonfinancial factors, such as labor market rigidity and wealth inequality. Despite rising life expectancy, tenure in stable lifetime jobs has stagnated, pushing workers into less secure, low-wage post-retirement jobs. These employment insecurities strengthen household incentives for asset accumulation, which, alongside portfolio differences across age groups, has driven household debt growth. Han (2024) finds that Korea’s rigid, seniority-based wage system exacerbates job insecurity among middle-aged workers. A shift to a flexible, performance-based wage system could enhance labor market efficiency and mitigate household debt growth.

In addition, as core economic fundamentals such as shifts in population structure materially influence the household debt-to-GDP ratio,household debt policies built around arbitrary aggregate targets are likely to cause unintended distortions, including excessive market frictions and high adjustment costs. The analysis demonstrates that household debt trends largely reflect natural shifts in saving and borrowing patterns, driven by rising life expectancy and shifting age distribution. Therefore, in a normally functioning financial market, debt policy should prioritize assessing the repayment capacity of borrowers and safeguarding the financial soundness of financial institutions,rather than excessively restricting fund flows unless in exceptional circumstances.

To mitigate both borrower default risk and systemic risk, it is essential to gradually roll back exemptions to the Debt Service Ratio (DSR) regulations. Current rules allow broad carve-outs or looser ratios for policy mortgages (Bogeumjari Loans and qualified mortgages), group loans for interim payments or relocation, and jeonse loans. Some loans are extended with scant regard for repayment capacity on the grounds of their public purpose. While such exemptions improve credit access for vulnerable groups and support short-term cyclical flexibility, they undermine the effectiveness of DSR regulations, compromise the fairness and predictability of the debt regulatory regime, and distort market-based risk assessments. Accordingly, policymakers should prioritize narrowing DSR exemptions. Where exemptions are unavoidable, they should be paired with strict repayment capacity assessments and a risk-based framework that differentiates by loan purpose and repayment structure.

Finally, although not the primary focus of this study, excessive policy financing can also be identified as a structural driver of household debt expansion in Korea. While policy financing serves as a tool to address market failures arising from information asymmetry and achieve public policy goals, it risks extending credit at below-market interest rates with relaxed screening criteria to borrowers who would otherwise be excluded by private lenders. This likely fuels household debt growth, while overly high guarantee ratios combined with fees set too low relative to coverage levels may lead to inefficient debt accumulation under weak market-based risk assessments. To mitigate these risks, guarantee ratios should be adjusted and fees aligned more closely with underlying risks. While reduced fees for vulnerable groups may be justified as part of social safety net policies, policy financing programs should be thoroughly assessed to prevent the extensive application of high-guarantee, low-fee structures.

Appendix

This study employs the methodology of Higgins (1998) to model the effects of life expectancy and age-group population shares on the household debt-to-GDP ratio, as specified below:

This method reduces age-group population shares to their first and second moments and incorporates them into the regression model as explanatory variables, based on assumptions about the regression coefficients.