## Introduction — "Graying Asia: How Aging Is Reshaping Banking" (Working Paper No. WP/2026/150)

## Source details

**Canonical URL:** [Introduction — "Graying Asia: How Aging Is Reshaping Banking" (Working Paper No. WP/2026/150)](https://www.imf.org/-/media/files/publications/wp/2026/english/wpiea2026150-source-pdf.pdf)

## Other formats

- [Markdown version](/-/media/files/publications/wp/2026/english/wpiea2026150-source-pdf.pdf.md)
- [Structured JSON version](/-/media/files/publications/wp/2026/english/wpiea2026150-source-pdf.pdf.json)

---

### Context and motivation
- The Asia‑Pacific region is undergoing a pivotal demographic transition characterized by rapid population aging and, in some economies, population decline.
- Total regional population is projected to peak in 2054 (United Nations, 2024).
- Demographic shifts are uneven across the region and will shape economies and financial markets, with specific consequences for financial systems and banks.

### Why the demographics–banking link matters for Asia‑Pacific
- Many Asia‑Pacific emerging markets remain bank-dominated, with banks intermediating the majority of savings and credit.
- As populations age, household financial behavior typically shifts toward lower leverage and higher liquidity, including greater deposit holdings—affecting:
  - banks’ funding structures,
  - lending patterns,
  - profitability,
  - business models,
  - risk profiles.
- Demographic divergence may prompt global and regional banks to reallocate assets toward younger, faster‑growing markets, influencing capital flows and financial stability.

### Research questions and contributions
- Research questions:
  - How will demographic change (aging and population growth) affect banks’ balance sheets and risk‑taking in the Asia‑Pacific region?
  - What are the potential implications for financial stability?
- Main contributions:
  1. Integrated framework tracing impacts from households to banks and cross‑border channels.
  2. Joint analysis of global and Asia‑Pacific samples, including many emerging markets.
  3. First exploration (to authors’ knowledge) of cross‑border banking changes arising from demographic change.

---

### Demographic developments and channels to banking

### Regional demographic patterns
- Asia‑Pacific exhibits rapid yet uneven demographic transition:
  - Post‑dividend: Japan, Korea, China.
  - Late‑dividend: Vietnam, Malaysia.
  - Early‑dividend: Philippines, Bangladesh.
- Over the next two decades key aging metrics (share of old‑age population and total population growth rate) will approach advanced‑economy levels (UN projections).
- Two overarching patterns:
  - Timing: Europe and Japan aged earlier; before 2010 aging largely confined to these economies.
  - Speed: since 2010 many Asian economies moved from late‑ to post‑dividend within a single generation; by 2050 most are projected to reach severe aging.
- Many Asian economies age at income levels below high‑income thresholds and with less comprehensive social safety nets.

### Primary transmission channel to banks
- Household financial behavior (saving, borrowing, investing) shifts with age and transmits to corporates, the public sector, and banks.
- Anticipated bank impacts include changes in funding structure, business models, risk‑taking, overseas expansion, M&A, and interactions with non‑bank financial institutions.

---

### Data and empirical approach

### Data sources and coverage
- Household surveys: Australia, China, Indonesia, Japan, New Zealand, Philippines (primary); Korea, Malaysia, Nepal, Thailand (partial).
- Bank‑level panel: Fitch Connect Database; global panel of 4,256 banks covering 1989–2023; unconsolidated financial statements.
- Country‑level panel: aggregated averages across 56 economies (unbalanced); balanced panel constructed for 2000–2023 for robustness.
- Cross‑border allocation: Ultimate Ownership Dataset from Fitch Connect; final cross‑border sample: 166 banking groups headquartered in 46 economies, yielding 2,504 bank–country–year observations.
- Macroeconomic controls: GDP growth rate; ln(real GDP per capita); CPI inflation; trade openness; policy interest rate (WEO); credit‑to‑GDP (World Bank GFDD).
- Demographics: UN World Population Prospects — old‑age share (65+), total population growth, life expectancy, active‑age share (15–49).

### Panel and regional distribution (high‑level)
- Global bank‑level panel: 4,256 banks (1989–2023).
- Country‑level panel: 56 economies.
- Cross‑border final sample: 166 banking groups; 2,504 bank–country–year observations.
- Asia‑Pacific: Number of Economies 16; Number of Banks 1,732; Demographic Status: Early Dividend 4, Late Dividend 6, Post Dividend 6.
- Panel is unbalanced with maximum coverage 1989–2023.

---

### Household‑level stylized facts and scenario

### Stylized household findings
- Wealth accumulation: household financial assets and total assets peak around or before retirement; liabilities peak slightly earlier; net worth/net financial assets peak around retirement.
- Debt participation: share of indebted households rises in mid‑life and declines steadily with age; a non‑negligible share of elderly households still hold debt.
- Portfolio composition: shift toward safer assets with age; ratio of deposits to total financial assets is U‑shaped across ages—falls in mid‑life and rises for elderly households.
- Financial leverage: average loan‑to‑asset ratio and average loan‑to‑deposit ratio decline with age.

### Simple forward scenario (method and illustrative outcomes)
- Method: hold within‑age‑group behavior constant; let population size and composition evolve to 2050 per UN projections using:
  - x_t = sum_w_{i,t} x_{i,t} and ratios v_{t+n} = sum w_{i,t+n} l_i / sum w_{i,t+n} y_i (formulae as in source).
- Country illustrations: Japan (post‑dividend), New Zealand (late/younger), Philippines (early‑dividend).
  - Age‑distribution shifts alone: increase total deposits and decrease total loans (except in younger Philippines).
  - Including population size change: shrinking populations (Japan) imply fewer deposits and fewer loans; growing populations (New Zealand, Philippines) imply increases in both aggregates.
  - Across examples, average loan‑to‑asset ratio (LAR) and average loan‑to‑deposit ratio (LDR) fall between 2025 and 2050.
- Key insight: aging with positive population growth can raise aggregate levels while lowering leverage ratios; aging with population shrinkage reduces both levels and ratios.

---

### Country‑level regression results

### Specification
- Y_it = β A_it + α X_it + μ_i + μ_t + ε_it
  - Dependent variables: loan‑to‑deposit ratio, deposit‑to‑funding ratio, loan‑to‑asset ratio (country‑year aggregates).
  - Key independent A_it: old‑age share (65+/Total). Robustness uses population growth, life expectancy, active‑age share (15–49).
  - Controls X_it: real GDP growth rate; ln(real GDP per capita); CPI inflation; policy interest rate; trade openness; credit‑to‑GDP ratio.
  - Country and year fixed effects included.

### Baseline results (sample: Observations = 1,038; R2 reported per column)
- Loan‑to‑Deposit (column 1):
  - Old‑age Share (65+/Total): -5.162*** (0.764)
  - CPI Inflation: -0.283 (0.195)
  - Real GDP Growth Rate: -0.295** (0.142)
  - Ln Real GDP per Capita: 36.669*** (5.440)
  - Short‑term Interest Rate: -4.233 (6.402)
  - Credit‑to‑GDP Ratio: 54.515*** (7.205)
  - R2 = 0.579
- Loan‑to‑Asset (column 2):
  - Old‑age Share (65+/Total): -1.414*** (0.243)
  - CPI Inflation: -0.004 (0.067)
  - Real GDP Growth Rate: -0.163*** (0.047)
  - Ln Real GDP per Capita: 3.227 (2.041)
  - Short‑term Interest Rate: 1.427 (1.806)
  - Credit‑to‑GDP Ratio: 21.759*** (2.565)
  - R2 = 0.696
- Deposit‑to‑Funding (column 3):
  - Old‑age Share (65+/Total): 1.296*** (0.305)
  - CPI Inflation: 0.089 (0.056)
  - Real GDP Growth Rate: -0.025 (0.068)
  - Ln Real GDP per Capita: -12.596*** (2.014)
  - Short‑term Interest Rate: -1.083 (2.365)
  - Credit‑to‑GDP Ratio: -15.921*** (2.397)
  - R2 = 0.615

### Interpretation and robustness
- A 1‑percentage point increase in the old‑age share is associated with:
  - a 5.2‑percentage point decline in loan‑to‑deposit ratio.
  - about a 1.4‑percentage point decline in loan‑to‑asset ratio.
  - about a 1.3‑percentage point increase in deposit‑to‑funding ratio.
- Including population growth, life expectancy, and active‑age share preserves relationships; population growth often offsets aging effects.
- Robustness checks (excluding low‑income countries; balanced panel 2000–2023) support results.

### Forward‑looking application
- Estimated elasticities applied to UN demographic projections (assumes linear marginal associations hold forward).
- Illustration shows Asia projected to experience larger declines in loan‑to‑deposit and loan‑to‑asset ratios than other regions.
- Within Asia:
  - Australia and New Zealand retain high loan‑to‑deposit and loan‑to‑asset ratios.
  - China, Korea, Thailand could see diversification away from loans and decreased loan‑to‑asset ratios.
- Emphasis: direction and speed of change; linear interpretation noted.

---

### Cross‑border allocation results

### Specification
- Y_{b,i,t} / G_{b,t} = β1 A_{f,t} + β2 X_{i,t} + μ_{b,i} + γ_{h,t} + ε_{b,i,t}
  - Dependent variable: share of banking group b’s assets in host country f relative to group global total assets G_{b,t}.
  - Key independent: host‑country old‑age share (65+).
  - Controls: same macro variables as country‑level regression.
  - Fixed effects: bank‑country pair μ_{b,i}; home‑country*year γ_{h,t}.
  - Home‑country observations excluded to focus on cross‑border allocations.

### Main cross‑border findings (Observations = 2,504; R2 = 0.827)
- Old‑age Share (65+/Total): -0.851** (0.418)
- Real GDP Growth Rate: 0.170*** (0.059)
- Ln Real GDP per Capita: -9.526*** (3.034)
- Short‑term Interest Rate: -0.025 (0.039)
- Credit‑to‑GDP Ratio: 11.623*** (3.081)
- CPI Inflation: 0.018 (0.044)
- Trade‑to‑GDP Ratio: 2.159 (2.715)

### Interpretation
- A one‑percentage point increase in host‑country old‑age share is associated with an approximately 0.85‑percentage point decrease in the subsidiary asset share relative to parent group total assets.
- Linear projections of old‑age share changes between 2023 and 2050 correspond to decreases ranging from 8 to 32 percentage points in subsidiary asset shares relative to ultimate groups’ total assets.
- Banks allocate more to countries with higher real GDP growth, lower ln(real GDP per capita), and higher credit‑to‑GDP ratios.

---

### Conclusion, risks, and policy implications

### Key conclusions
- Aging reshapes bank behavior primarily via household channels: older households hold safer assets and exhibit reduced leverage, leading to more deposits and reduced loan participation.
- Panel evidence: rising old‑age share is associated with lower loan‑to‑deposit and loan‑to‑asset ratios—indicating shrinking credit demand/supply and reallocation away from loans.
- Cross‑border evidence: banks allocate fewer assets to aging host economies and more to younger host economies.

### Implications for bank balance sheets and risk
- Lower household leverage and stronger deposit preferences reduce share of loans on bank balance sheets.
- Banks likely to diversify into non‑loan assets and fee‑generating activities (trading books, wealth/retirement services), introducing exposures to market, duration, and fee‑income volatility.
- Increased liquidity supports regulatory buffers but new risks may arise in less regulated or non‑bank activities.

### Cross‑border and country‑group considerations
- Systematic reallocation toward younger host markets could channel foreign savings to younger and lower‑income host economies.
- In advanced aging economies with deeper capital markets, reallocation may reflect portfolio diversification away from subdued credit demand.
- In aging emerging markets with weaker safety nets, findings underscore importance of deepening alternative financing sources and strong supervision.

### Policy recommendations
- Strengthen regulatory agility to monitor financial innovation and banks’ shift into non‑loan activities.
- Deepen domestic capital markets to furnish non‑bank financing channels, especially in bank‑dependent emerging markets.
- Bolster supervision of non‑bank financial intermediation that may absorb bank reallocation and systemic risk.
- Tailor policies to country financial structures: bank‑based economies show a sharper decline in loan‑to‑deposit ratios with aging; deposit shares increase with aging in bank‑based systems but not significantly in market‑based systems.

*Source: IMF staff analysis, "Graying Asia: How Aging Is Reshaping Banking", Working Paper No. WP/2026/150*

### Introduction ...........................................................................................................

### Introduction

### Context and motivation
- The Asia-Pacific region is undergoing a pivotal demographic transition characterized by rapid population aging and, in some economies, population decline.
- The total regional population is projected to peak in 2054, according to United Nations population projections (United Nations, 2024).
- Demographic shifts are uneven across the region.
- Demographic forces (aging and population decline) will shape economies and financial markets, with specific consequences for financial systems and banks that have been less studied than macroeconomic impacts on public finances, pensions, interest rates, and productivity.

### Why the demographics–banking link matters for Asia‑Pacific
- The region is experiencing rapid yet uneven demographic shifts, with aging accelerating.
- Financial systems in many Asia‑Pacific emerging markets remain bank-dominated, with banks intermediating the majority of savings and credit.
- As populations age, household financial behavior is expected to change—shifting typically toward lower leverage and higher liquidity, including via greater deposit holdings.
- In bank-based financial systems, these household shifts will likely influence:
  - banks’ funding structures,
  - lending patterns,
  - profitability,
  - business models,
  - risk profiles.
- Demographic divergence across economies may prompt global and regional banks to reallocate assets away from older populations toward younger, faster‑growing ones, affecting capital flows, financial integration, and stability (including potential funding retrenchment or shifts in foreign subsidiaries’ product mix).

### Research questions
- How is demographic change (mainly aging and population growth) expected to affect banks’ balance sheets and risk‑taking in the Asia‑Pacific region?
- What are the potential implications for financial stability?

### Main contributions
- Three-fold contributions:
  1. Provide an integrated framework and fresh evidence—tracing the impact from households to banks and further to cross‑border channels—to assess the impact of aging on the banking sector.
  2. Focus on both global samples and Asia‑Pacific economies in more detail, covering a wide country sample including many emerging market economies.
  3. To the authors’ knowledge, offer the first exploration of changes to cross-border banking as a result of demographic change.

### Key findings (from this study)
- Aging is associated with:
  - (i) lower loan-to-deposit and loan-to-asset ratios,
  - (ii) a shift in bank assets toward non-loan assets, which are not necessarily safer,
  - (iii) changes in household behavior explain much of the funding and lending shifts,
  - (iv) the expansion of banks’ cross-border activities tends to be focused towards “early-dividend” (younger) markets.

### Structure of the paper
- Section 2: demographic developments in the Asia-Pacific region, focusing on aging and population growth.
- Section 3: literature review on the aging‑banking nexus.
- Section 4: description of the different data sets used.
- Section 5: stylized facts on household behavior from household surveys in the region and a simple stylized scenario.

*IMF Working Paper — Introduction*

### Section 6 presents the results from both country-level panel analysis and cross-border analysis, showing how

### Section 6: Results — How demographics impact bank balance sheets and global asset allocation

### Demographic developments
- Asia-Pacific exhibits rapid yet uneven demographic transition: some economies (Japan, Korea, China) are post-dividend, others late-dividend (Vietnam, Malaysia), and others early-dividend (Philippines, Bangladesh).
- Many economies in the region will become old before becoming rich (IMF, 2017b). UN projections imply that over the next two decades key aging metrics (share of old-age population and total population growth rate) will approach advanced‑economy levels.
- Two overarching patterns:
  - Timing: Europe and Japan entered aging earlier; before 2010 aging was largely confined to these economies.
  - Speed: since 2010 many Asian economies moved through late- to post-dividend stages in a single generation; by 2050 most are projected to reach severe aging (high old-age dependency ratios and population shrinkage).
- Aging in many Asian economies will occur at income levels below high‑income thresholds and with less comprehensive social safety nets.
- Even where population headcounts remain positive (outside post-dividend group), the share of older people rises, driving rapid deterioration in old-age dependency ratios via composition effects rather than universal population decline.
- Region-specific challenges arise for banks and regulators owing to speed and unevenness of transition (e.g., Japan or Korea vs. rapidly aging emerging markets such as China, Thailand, Sri Lanka).

### Linking demographics to the financial system
- Primary channel: shifts in household financial behavior (saving, borrowing, investing) as populations age; these household changes feed through to corporates, the public sector, and ultimately the banking sector.
- Conceptual impacts on banks include changes in funding structure, business model, risk-taking, overseas expansion, M&A, and interactions with the non-bank financial system.

### Literature summary
- Life-cycle hypothesis (Modigliani and Brumberg, 1954) and permanent income hypothesis (Friedman, 1957) imply age structure affects saving and asset accumulation (hump-shaped wealth accumulation).
- Empirical work links aging to lower risk appetite, higher deposit shares, lower leverage after retirement (e.g., Imam, 2013; Li et al., 2024); at bank level, aging tends to lower credit demand and compress margins (Ramlall, 2023; Berlemann et al., 2014).
- Recent studies point to possible risk-taking incentives as banks seek yield in aging economies (Doerr et al., 2024; Imam & Schmieder, 2025).
- Empirical evidence for Asia-Pacific, especially EMs and cross‑border banking, remains limited; this paper contributes by jointly analyzing (i) household behavior; (ii) bank balance‑sheet composition; and (iii) cross‑border asset and liability allocations using a global bank panel with a focus on Asia.

### Data description
- Household surveys: Australia, China, Indonesia, Japan, New Zealand, Philippines (primary); Korea, Malaysia, Nepal, Thailand (partial). Caveats: aggregation/age-grouping differences, mean-based skew, pension coverage differences, cohort effects, longevity-income correlations, mostly aggregated (non-longitudinal) data.
- Bank-level panel: Fitch Connect Database; global bank-level panel of 4,256 banks covering 1989–2023 (observations before 1989 limited). Unconsolidated financial statements used to avoid double-counting across countries.
- Country-level panel: aggregated averages of key ratios (loan-to-deposit ratio, loan-to-asset ratio, deposit-to-funding ratio) across 56 economies (unbalanced panel, main database). Balanced panel constructed for 2000–2023 for robustness. Low-income countries excluded in main analysis (findings broadly hold when included).
- Macroeconomic controls: GDP growth rate, GDP per capita (log), CPI inflation, trade openness (exports+imports to GDP), policy interest rate (WEO); credit-to-GDP from World Bank GFDD.
- Demographics: UN World Population Prospects — old-age share (65+), total population growth, life expectancy, active-age share (15-49).
- Cross‑border allocation dataset: Ultimate Ownership Dataset from Fitch Connect to map subsidiaries to 3,129 ultimate parents; of 4,256 banks, 3,751 matched to 3,129 ultimate parents. Excluding single-country groups yields final sample of 166 banking groups headquartered in 46 economies during 1989–2023, yielding 2,504 bank–country–year observations.

### Panel coverage and regional distribution (Table 1)
- Global bank-level panel: 4,256 banks (1989–2023).
- Country-level panel: 56 economies.
- Cross-border final sample: 166 banking groups; 2,504 bank–country–year observations.
- Region-specific counts (unbalanced panel):
  - Asia-Pacific: Number of Economies 16; Number of Banks 1,732; Demographic Status: Early Dividend 4, Late Dividend 6, Post Dividend 6.
  - Europe: Number of Economies 13; Number of Banks 1,412; Early Dividend 2, Late Dividend 3, Post Dividend 8.
  - Western Hemisphere: Number of Economies 11; Number of Banks 912; Early Dividend 3, Late Dividend 7, Post Dividend 1.
  - Middle East and Central Asia: Number of Economies 8; Number of Banks 99; Early Dividend 5, Late Dividend 3, Post Dividend 0.
  - Sub-Saharan Africa: Number of Economies 8; Number of Banks 101; Early Dividend 5, Late Dividend 2, Post Dividend 1.
- Note: panel unbalanced with maximum coverage 1989–2023.

### Household-level findings (stylized facts)
- Wealth accumulation:
  - Household financial assets and total assets peak around or before retirement; household liabilities peak slightly earlier (mid-life); net worth/net financial assets peak around retirement — consistent with hump-shaped accumulation (Figure 6).
- Debt participation:
  - Share of indebted households increases in mid-life and declines steadily with age; a non-negligible share of elderly households still hold debt.
- Portfolio composition:
  - Shift toward safer assets with age. Ratio of deposits to total financial assets follows a U-shaped pattern across ages — falls in mid-life and rises for elderly households, tilting toward bank deposits; level of bank deposits relatively stable across elderly groups.
- Financial leverage:
  - Average loan-to-asset ratio and average loan-to-deposit ratio by age group decline with age (Figure 9) — older cohorts show lower leverage and risk exposure.
- Interpretation:
  - Older households expect less labor income growth, have lower spending needs, lower mortgage/consumer loan demand; lenders may tighten credit to older borrowers due to repayment and longevity risks; medical/health risks incentivize safer liquid asset holdings (deposits).

### A simple scenario (illustrative)
- Method: hold within-age-group behavior constant; let population size and composition evolve to 2050 per UN projections. Formula summarized by:
  - x_t = sum_w_{i,t} x_{i,t}  with assumptions detailed in text; for ratios v_{t+n} = sum w_{i,t+n} l_i / sum w_{i,t+n} y_i (equations preserved in source).
- Illustrative country examples: Japan (post-dividend), New Zealand (late/younger), Philippines (early-dividend).
  - Findings: age-distribution shift (larger elderly share) alone would increase total deposits and decrease total loans (except in younger Philippines). When population size change is included, shrinking populations (Japan) imply both fewer deposits and fewer loans; growing populations (New Zealand, Philippines) imply increases in both aggregates.
  - Across examples, average loan-to-asset ratio (LAR) and average loan-to-deposit ratio (LDR) fall between 2025 and 2050.
- Key insight: aging plus positive population growth can raise aggregate levels (deposits, loans) while lowering leverage ratios; aging with population shrinkage reduces both levels and ratios.

### From household behavior to bank impact (conceptual)
- Household-level lower debt participation and safer-asset tilt imply:
  - Lower credit exposures for banks.
  - Lower bank profitability prospects (due to more low-yield liquid assets).
  - Slower household-loan growth and household-deposit growth in aging/shrinking economies.
  - Declining loan-to-deposit ratios, even where population growth remains positive.
- Banks likely to adapt business models toward non-loan assets, fee-based activities, and asset-allocation adjustments.

### Country-level panel regressions — specification
- Regression form:
  - Y_it = β A_it + α X_it + μ_i + μ_t + ε_it
  - Dependent variables Y_it: loan-to-deposit ratio, deposit-to-funding ratio, loan-to-asset ratio (country-year aggregates).
  - Key independent A_it: old-age share (baseline). Robustness includes population growth rate, life expectancy, active-age share (15-49).
  - Controls X_it: real GDP growth rate; ln(real GDP per capita); CPI inflation; policy interest rate; trade openness; credit-to-GDP ratio.
  - Country and year fixed effects included.

### Baseline country-level regression results (Table 2)
- Sample: Observations (country-year) = 1,038; R2 reported per column.
- Main coefficients (standard errors in parentheses):
  - Loan-to-Deposit (column 1):
    - Old-age Share (65+/Total): -5.162*** (0.764)
    - CPI Inflation: -0.283 (0.195)
    - Real GDP Growth Rate: -0.295** (0.142)
    - Ln Real GDP per Capita: 36.669*** (5.440)
    - Short-term Interest Rate: -4.233 (6.402)
    - Credit-to-GDP Ratio: 54.515*** (7.205)
    - Observations = 1,038; R2 = 0.579
  - Loan-to-Asset (column 2):
    - Old-age Share (65+/Total): -1.414*** (0.243)
    - CPI Inflation: -0.004 (0.067)
    - Real GDP Growth Rate: -0.163*** (0.047)
    - Ln Real GDP per Capita: 3.227 (2.041)
    - Short-term Interest Rate: 1.427 (1.806)
    - Credit-to-GDP Ratio: 21.759*** (2.565)
    - Observations = 1,038; R2 = 0.696
  - Deposit-to-Funding (column 3):
    - Old-age Share (65+/Total): 1.296*** (0.305)
    - CPI Inflation: 0.089 (0.056)
    - Real GDP Growth Rate: -0.025 (0.068)
    - Ln Real GDP per Capita: -12.596*** (2.014)
    - Short-term Interest Rate: -1.083 (2.365)
    - Credit-to-GDP Ratio: -15.921*** (2.397)
    - Observations = 1,038; R2 = 0.615
- Interpretation:
  - A 1-percentage point increase in the old-age share is associated with a 5.2-percentage point decline in loan-to-deposit ratio.
  - A 1-percentage point increase in the old-age share is associated with about a 1.4-percentage point decline in the loan-to-asset ratio.
  - A 1-percentage point increase in the old-age share is associated with about a 1.3-percentage point increase in the deposit-to-funding ratio.
- Robustness:
  - Including population growth, life expectancy, and active-age share preserves the relationships (Figure 12). Population growth often offsets aging effects (boosting credit and moderating deposit dominance).
  - Additional robustness checks: excluding low-income countries and using balanced panel (2000–2023) support results.

### Forward-looking application
- Method: apply estimated regression elasticities to UN demographic projections (assumes linear marginal associations hold forward).
- Illustration (Figure 13): Asia projected to experience larger declines in loan-to-deposit and loan-to-asset ratios than other regions due to rapid aging.
- Within Asia:
  - Australia and New Zealand (high initial base) would retain high loan-to-deposit and loan-to-asset ratios.
  - China, Korea, Thailand (faster aging, lower initial ratios) could see diversification of asset allocation and decreased loan-to-asset ratios.
- Note: linear interpretation; future research could explore non-linearities. Emphasis is on direction and speed rather than precise magnitudes.

### Cross-border channels — specification and findings
- Regression form (cross-border allocation):
  - Y_{b,i,t} / G_{b,t} = β1 A_{f,t} + β2 X_{i,t} + μ_{b,i} + γ_{h,t} + ε_{b,i,t}
  - Dependent variable: share of banking group b’s assets in host country f relative to group global total assets G_{b,t}.
  - Key independent: host-country old-age share (65+).
  - Controls: same macro variables X_{i,t} as country-level regression.
  - Fixed effects: bank-country pair μ_{b,i}; home-country*year γ_{h,t}. Observations in banking group’s home country excluded to focus on cross-border allocations.

- Main cross-border results (Table 3):
  - Old-age Share (65+/Total): -0.851** (0.418)
  - Real GDP Growth Rate: 0.170*** (0.059)
  - Ln Real GDP per Capita: -9.526*** (3.034)
  - Short-term Interest Rate: -0.025 (0.039)
  - Credit-to-GDP Ratio: 11.623*** (3.081)
  - CPI Inflation: 0.018 (0.044)
  - Trade-to-GDP Ratio: 2.159 (2.715)
  - Observations (bank-country-year) = 2,504; R2 = 0.827
- Interpretation:
  - A one-percentage-point increase in host-country old-age share is associated with an approximately 0.85-percentage-point decrease in the share of subsidiary banks’ assets relative to their ultimate parents’ total assets.
  - Linear projections imply projected changes in old-age shares between 2023 and 2050 correspond to decreases ranging from 8 to 32 percentage points in subsidiary asset shares relative to ultimate groups’ total assets.
  - Banks allocate more to countries with higher real GDP growth and lower ln(real GDP per capita); asset share also rises with credit-to-GDP ratio (financial market development attracts capital).

*Source: IMF staff analysis and calculations based on UN 2024 World Population Prospects, Fitch Connect Database, IMF World Economic Outlook Database, and World Bank Global Financial Development Database.*

### Conclusion and Discussion

### Conclusion and Discussion

### Main findings on demographics and bank behavior
- Demographics reshape bank behavior through household channels:
  - Household-level evidence shows aging is associated with more holdings of safer financial assets and reduced leverage—reflected in increased bank deposits and reduced loan participation.
- Aging shifts bank portfolios:
  - Panel regressions show that a rising old-age share is associated with lower loan-to-deposit and loan-to-asset ratios, indicating shrinking credit demand or supply and asset reallocation away from loans, not necessarily lower leverage.
- Demographics shape cross-border banking decisions:
  - A novel cross-border regression framework shows banks allocate fewer assets to aging host economies—or more assets to younger host economies—suggesting demographic dynamics influence global expansion strategies.

### Implications for bank balance sheets and risk
- Demographic transition in Asia‑Pacific is altering the plumbing of banking:
  - Lower household leverage and stronger preference for deposits translate into a lower share of loans in bank balance sheets.
  - As populations age and in some cases shrink, banks are likely to diversify away from loans toward other activities, which may introduce new risks not captured by traditional leverage metrics but reflected in adjusted balance sheet composition.
- Shift toward non-loan assets and fee-generating businesses:
  - Weaker loan demand and thinner lending margins in aging economies could push banks toward trading books, wealth or retirement services, and other fee-generating activities that better match a stable deposit base.
  - These activities raise exposure to market, duration, and fee-income volatility; aging can thus reduce loan concentration while reshaping, rather than eliminating, risk.
- Liquidity and regulatory buffers:
  - More liquidity in the banking system would benefit regulatory buffers, but risks could re-emerge in new, less regulated activities or in non-bank financial institutions, making regulatory agility critical.

### Cross-border and country-group considerations
- Systematic reallocation by banks:
  - Banks systematically allocate less balance-sheet capacity to demographically advanced host markets and more to demographically younger host markets, adding a cross‑border dimension to demographic transitions in the region.
- Opportunities for younger and lower-income host economies:
  - Younger and lower-income economies could tap foreign savings for domestic investment as banks reallocate assets toward younger host markets.
- Heterogeneity across country groups:
  - In more advanced aging economies with deeper capital markets, reallocation may primarily reflect portfolio diversification away from subdued domestic credit demand.
  - In emerging market economies—often aging with weaker social safety nets and more bank‑dependent financial systems—the findings underscore the importance of investing in deepening alternative sources of financing, including domestic capital markets, along with effective supervision and macro‑prudential frameworks.

### Policy recommendations and supervisory implications
- Strengthen regulatory agility:
  - Enhance regulatory and supervisory frameworks to keep pace with financial innovation and the migration of banks into non-loan and fee-generating activities.
- Deepen alternative financing sources:
  - Invest in domestic capital markets to provide non-bank financing channels, particularly in emerging market economies that remain bank-dependent.
- Bolster supervision of non-bank activities:
  - Monitor and, where appropriate, extend regulation or oversight to non-bank financial intermediation that may absorb bank reallocation and systemic risk.
- Tailor policies to country financial structures:
  - Recognize heterogeneity between bank-based and market-based economies: bank-based economies exhibit a sharper decline in loan-to-deposit ratios with aging, and deposit shares increase with aging in bank-based systems but not significantly in market-based systems—policy responses should account for these structural differences.

*Source: Conclusion and Discussion, "Graying Asia: How Aging Is Reshaping Banking", Working Paper No. WP/2026/150*

---


_Source: https://www.imf.org/-/media/files/publications/wp/2026/english/wpiea2026150-source-pdf.pdf_