## _wp16194

## Source details

**Canonical URL:** [_wp16194](https://www.imf.org/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2016/_wp16194.pdf)

## Other formats

- [Markdown version](/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2016/_wp16194.pdf.md)
- [Structured JSON version](/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2016/_wp16194.pdf.json)

---

### I. Introduction — fiscal context and research focus
- Fiscal policy re-emerged to support demand while monetary policy in many advanced economies is constrained by the zero lower bound (ZLB).
- Japan context:
  - BoJ engaged in unprecedented government bond purchases and introduced a negative interest rate on marginal excess reserves.
  - Private consumption and investment subdued, especially since the consumption tax hike in April 2014.
  - Government target: achieve a primary surplus by FY2020 and reduce the debt-to-GDP ratio afterwards.
  - IMF (2016) estimate: potential growth will decline from 0.5 percent to 0.1 percent by 2030.
- Research focus:
  - How Ricardian is Japan? Two structural breakers examined: myopia (discount wedge) and liquidity constraints.
- Main previewed findings:
  - The discount wedge (myopia) has increased.
  - Liquidity constraints are assessed to be rising.
  - Illustrative simulations of fiscal consolidation are presented under different assumptions on these two factors.

### II. Theory and literature on Ricardian equivalence — mechanisms and prior evidence
- Core concept:
  - Ricardian equivalence: deficit financing is equivalent to taxation if individuals anticipate future tax increases and fully offset debt issuance by saving.
- Key conditions breaking Ricardian equivalence:
  - Imperfect private capital markets.
  - Uncertainty regarding future tax liabilities.
  - Liquidity constraints (households cannot smooth consumption).
  - Myopia (consumers discount future more than market rate).
- Prior empirical evidence summarized:
  - Feldstein (1974), Buchanan (1976): private saving fell with social security introduction.
  - Poterba and Summers (1987): sustained budget deficits in the US in the 1980s coincided with reduced saving and increased consumption.
  - Bhattacharya (1999): for high net debt countries, negative relationship between public debt and propensity to consume; no clear relationship in Japan in 1995 (net debt 23 percent of GDP in 1995).
  - Walker (2002): in Japan (1980–2000) timing of taxation has little impact; spending multiplier falls with deficits larger than 7 percent of GDP.
  - Bayoumi and Sgherri: model with myopic consumers fits the US; discount wedge in the US declining, making the US more Ricardian.

### III. Stylized facts — demographics, assets, and public finances
- Growth and demographics:
  - Average real GDP growth: 4.6 percent in the 1980s, 1.4 percent in the 1990s, 0.5 percent in the 2000s.
  - Median age: 33 years in 1980; 47 years in 2015.
  - Share of people aged 60 or above: 13 percent in 1980; 33 percent in 2015.
  - Population began declining around 2010 and is projected to be smaller by more than 30 percent by 2060 compared to the peak.
- Household and government balance-sheet co-movement:
  - Net financial assets of households increased by 130 percent of GDP between 1991 and 2014.
  - Net financial liabilities of the general government increased by 125 percent of GDP between 1991 and 2014.
  - Total household asset-to-GDP ratio stable in the 1990s and after, reflecting a large valuation loss on non-financial assets after the bubble burst.
  - Per capita total household assets broadly flat since 1990.
  - Family Income and Expenditure Survey: no noticeable increase in the saving ratio except for households headed by people younger than 30, which occupies 4 percent of the total sample in 2015Q4.
- Fiscal position note:
  - Fiscal deficit was 6.2 percent of GDP in 2014; expected to continue declining if government pursues FY2020 primary surplus goal, but net debt ratio expected to be stable for next few years and then start rising (IMF, 2016).

### IV. Evidence on myopia — discount wedge estimation
- Methodology:
  - Follows Bayoumi and Sgherri (2006) using a system of three equations (consumption, income, net tax rate) estimated by seemingly unrelated regression on annual data from 1980; model incorporates r (real interest rate) and λ (discount wedge/myopia).
- Key unrestricted regression results (JPN 1981–2014):
  - Consumption equation: αC = -.021 (.004)** ; βY = .80 (.05)** ; βT = -.10 (.04)** ; βe = .23 (.04)** ; R2 = 0.89.
  - Income equation: αY = .38 (.15)* ; θY = .067 (.03)* ; γ = -.001 (.0002)** ; R2 = 0.55.
  - Net tax rate equation: αT = -.02 (.002)** ; θT = .16 (.03)** ; δ = 1.4 (.21)** ; R2 = 0.50.
  - Note: * (**) denotes statistical significance at a 5 (1) percent level. Numbers in parentheses are standard deviation.
- Discount wedge estimates and sensitivity:
  - Discount wedge in Japan is around 0.1 over the full sample period (1980–2014).
  - Sensitivity values: 0.11 and 0.095 when assuming a real interest rate of 4 percent and 0 percent, respectively; both statistically significant at the 1 percent level.
  - The average real interest rate for 1981–2014 was 1.5 percent.
  - Rolling 20-year window estimates suggest the additional discount rate has increased over time.
  - Average elasticity of consumption to net tax rate changes: 0.15 (first window) and 0.22 (last 5 windows), suggesting an increased tax multiplier.
- Interpretation:
  - Increasing discount wedge indicates Japan has become more myopic and thus less Ricardian.
  - Demographic driver: life expectancy minus median age (an indicator of national planning horizon) has been rapidly declining in Japan; IPSS projections suggest further narrowing.

### V. Evidence on liquidity constraints — recent developments and program evidence
- Macro and micro indicators pointing to stronger liquidity constraints:
  - Household saving rate has been on a declining trend (contrast with US and Germany); sparsely timed rises in 1998, 2009, and 2011 may reflect downturns or stimulus.
  - Increase in non-regular workers among young males implies more workers with lower wages and less job security, raising exposure to liquidity constraints.
  - CAO (2015): younger households with lower income reduced consumption after the 2014 consumption tax hike more than other households.
  - Per person financial assets have declined for younger generations.
  - Pension benefits per person have been cut in nominal terms recently and even in real terms in the past few years; Gini coefficient among the elderly is larger compared to other generations—inequality in asset stock among the elderly is large.
- Evidence from past cash transfer programs:
  - 1999 shopping coupon program:
    - Coupon: 20 thousand yen per child aged 15 or below and eligible elderly; valid six months.
    - Marginal propensity to consume (MPC) of the coupon: 10 percent overall; higher for liquidity-constrained households.
  - 2009 cash benefit program (CAO, 2012):
    - Payment: one-time 20 thousand yen to each person aged 18 or below and 65 or above; 12 thousand yen to other cohorts.
    - MPC: 0.25 for the whole sample; 0.40 for households with children; 0.37 for households with elderly.
  - Interpretation:
    - Japan is not fully Ricardian.
    - Liquidity constraints exist and appear binding for targeted groups.
    - The 1999 program (better targeted/time-limited) had a lower impact than 2009, suggesting the multiplier of cash transfers might have risen — consistent with Japan becoming less Ricardian.
- Public assistance trend:
  - Number of people on public assistance rose from less than a million to an all-time high of 2.2 million in the past twenty years.
  - Public expenditure on the program more than doubled from a bottom of 1.3 trillion yen (0.27 percent of GDP) before the collapse of the asset bubble to 3.6 trillion.

### VI. Evidence of rising liquidity constraints and demographic drivers (additional points)
- Pension and public assistance indicators:
  - Total pension benefits and monthly basic pension benefits per person series highlight rising pension outlays (figures shown for 1999–2016 and projections).
  - The ratio of people on public assistance is empirically explained by a regression of the cohort-level ratio on its lagged value, the non-regular worker ratio, and the job-to-applicant ratio.
  - Government estimate: 13.4 percent of workers employed at close to the minimum wage (less than 15 percent more than the minimum wage) in 2014, up from 9.2 percent in 2009.
  - An increase in the non-regular worker ratio has a statistically significant impact on public assistance participation, suggesting rising liquidity constraints.
  - People aged 65 or above have a higher chance of being on the public assistance program than other cohorts.
  - Even assuming a flattening of the non-regular worker ratio going forward, the number of people on public assistance is estimated to continue to rise, especially in terms of the ratio to the total population (the number in 2030 is calculated from the regression explained in footnote 9).
- Conclusion from these results:
  - A larger share of households will become liquidity constrained, making Japan “even less Ricardian.”

### VII. Multipliers and consumption tax experience
- Theoretical and model-based estimates:
  - A smaller Ricardian offset implies a higher fiscal multiplier.
  - Using the IMF’s Global Integrated Monetary and Fiscal (GIMF) Model:
    - Consumption tax multiplier estimated at around 0.4 using assumptions broadly common in advanced economies.
    - Calibrating parameters to reflect a higher ratio of liquidity constrained consumers raises the multiplier to around 0.6.
- Empirical estimate for the 2014 consumption tax hike (quarterly data for 1998Q1–2013Q3 using CAO (2015) methodology):
  - Implied multiplier amounts to 0.9 (national accounts data) or 0.6 (the BoJ’s new consumption indicator).
  - The 0.6 estimate is close to the tailored GIMF simulation result.
- Regression specification used to estimate private consumption:
  - Ct = α1 Yt + α2 OLDt + α3 Yt*OLDt + α4 FAt-1 + α5 CCIt + α6 E1 + α7 E2 + εt
  - Variables: C = private consumption; Y = compensation; FA = household’s net financial assets (all in real and logarithm); OLD = dependency ratio (ratio of people aged 60 or above to the total population); CCI = consumer confidence index; E1 and E2 = dummies for 2011Q1 and Q2.

### VIII. Timing and planning-horizon scenarios for consolidation
- Illustrative consolidation simulation:
  - A 10 percentage point consumption tax increase distributed over 10 years (1 point each year).
- Two timing scenarios compared:
  - (i) Start the 10 percentage point consumption tax increase over 10 years 10 years from now; assumed planning horizon of 6 years.
  - (ii) Start the same adjustment 5 years from now; assumed planning horizon of 10 years.
- Simulation result:
  - It is advisable to start fiscal consolidation sooner rather than later because longer planning horizons lead people to better prepare (save more), making the negative impact of the consumption tax increase more benign.
- Impact of liquidity constraints on consolidation:
  - Two liquidity-constrained consumer ratios assumed: 30 percent and 40 percent.
  - The negative impact of fiscal consolidation is larger when liquidity constraints are more binding; the first-year multiplier is estimated at -¾ in the higher-constraint case.
  - Given the possibility of rising liquidity constraints, these results argue for starting consolidation soon.

### IX. Financing Japanese Government Bonds (JGBs) and demand-supply outlook
- Becoming less Ricardian implies less household savings, reducing the principal domestic source of public debt financing.
- Supply calculation approach:
  - Supply of JGBs = current outstanding amount (including issuance by the Fiscal Investment and Loan Program (FILP)) + projected fiscal deficits.
- Key demand assumptions:
  - BoJ will stop increasing its JGB holdings by the end of 2017 and keep the outstanding amount at that time for the future.
  - Total savings at depositary institutions by the household and non-financial private sectors will continue to increase at the same pace as in the past 5 years; assumptions on their JGB holdings-to-deposit ratio come from Han (forthcoming).
  - Household assets managed by pension funds and insurers will continue to increase at the same pace as in the past 5 years; portfolio allocation assumed to follow the Government Pension Investment Fund’s target in the next few years.
- Illustrative outlook and risks:
  - Illustrative calculation suggests the supply of JGBs exceeds demand as early as 2018.
  - At that moment, interest rates may need to rise to attract enough buyers, including domestic depository institutions holding excess reserves at the BoJ’s current account.
  - Private pension funds and other insurers may reverse portfolio rebalancing towards riskier assets.
  - In the long run, the savings rate is expected to decline further due to ageing, myopic behavior, and rising liquidity constraints, resulting in a slower increase or decline in household financial assets.
  - Implication: pressure on government bond yields may start to be felt possibly in the near future as financing needs are increasingly met from external sources; efforts to restore fiscal sustainability should start soon and be sustained.

### X. Policy recommendations and conclusions
- Key findings:
  - Japan has likely become, and may continue to become, less Ricardian with ageing and weak economic prospects.
  - The estimate of the discount wedge suggests Japan may have become more myopic; a declining average remaining life expectancy plausibly shortens planning horizons.
  - Several indicators suggest liquidity constraints have become more binding in Japan.
- Policy implications and recommendations:
  - Stronger fiscal multipliers raise the effectiveness of fiscal policy as a counter-cyclical tool.
  - Reorienting expenditure toward liquidity constrained households is expected to enhance fiscal policy impact.
  - Fiscal consolidation should start soon in a gradual, yet steady, manner.
  - A potential reduction in savings poses a challenge for JGB financing in the not-too-distant future.
  - Given the large impact of the previous consumption tax hike, more gradual increases seem warranted.
  - Structural reforms to address ageing and rising liquidity constraints are indispensable.

*Source: _wp16194 - References (content excerpt).*

### References .............................................................................................................

### _wp16194 - References

### I. Introduction — fiscal context and research focus
- Fiscal policy has re-emerged as a tool to support demand while monetary policy in many advanced economies is constrained by the zero lower bound (ZLB).
- Japan context:
  - BoJ has engaged in unprecedented government bond purchases and introduced a negative interest rate on marginal excess reserves.
  - Private consumption and investment subdued, especially since the consumption tax hike in April 2014.
  - Government target: achieve a primary surplus by FY2020 and reduce the debt-to-GDP ratio afterwards.
  - IMF (2016) estimate: potential growth will decline from 0.5 percent to 0.1 percent by 2030.
- Key research question: How Ricardian is Japan? Focus on two structural breakers of Ricardian equivalence examined here:
  - Myopia (discount wedge)
  - Liquidity constraints
- Main findings previewed:
  - The discount wedge (myopia) has increased.
  - Liquidity constraints are assessed to be rising.
  - Illustrative simulations of fiscal consolidation are presented under different assumptions on these two factors.

### II. Theory and literature on Ricardian equivalence — mechanisms and prior evidence
- Ricardian equivalence (Ricardo, 1821; Barro, 1974): deficit financing is equivalent to taxation if individuals anticipate future tax increases and fully offset debt issuance by saving.
- Conditions breaking Ricardian equivalence highlighted:
  - Imperfect private capital markets
  - Uncertainty regarding future tax liabilities
  - Liquidity constraints (households cannot smooth consumption)
  - Myopia (consumers discount future more than market rate)
- Empirical evidence against full Ricardian equivalence:
  - Feldstein (1974) and Buchanan (1976): private saving fell with introduction of social security.
  - Poterba and Summers (1987): sustained budget deficits in the US in the 1980s coincided with reduced saving and increased consumption.
  - Bhattacharya (1999): for high net debt countries, negative relationship between public debt and propensity to consume consistent with Ricardian equivalence; no clear relationship in Japan in 1995 (net debt 23 percent of GDP in 1995).
  - Walker (2002): in Japan (1980–2000) timing of taxation has little impact; spending multiplier falls with deficits larger than 7 percent of GDP.
- Bayoumi and Sgherri findings:
  - A model with myopic consumers fits the US better than one with rule-of-thumb consumers.
  - Discount wedge in the US has been on a declining trend, making the US more Ricardian.

### III. Stylized facts — demographics, assets, and public finances
- Growth and demographics:
  - Average real GDP growth: 4.6 percent in the 1980s, 1.4 percent in the 1990s, 0.5 percent in the 2000s.
  - Median age: 33 years in 1980; 47 years in 2015.
  - Share of people aged 60 or above: 13 percent in 1980; 33 percent in 2015.
  - Population began declining around 2010 and is projected to be smaller by more than 30 percent by 2060 compared to the peak.
- Household and government balance-sheet co-movement:
  - Net financial assets of households increased by 130 percent of GDP between 1991 and 2014.
  - Net financial liabilities of the general government increased by 125 percent of GDP between 1991 and 2014.
  - Total household asset-to-GDP ratio has been stable in the 1990s and after, reflecting a large valuation loss on non-financial assets after the bubble burst.
  - Per capita total household assets broadly flat since 1990.
  - Family Income and Expenditure Survey: no noticeable increase in the saving ratio except for households headed by people younger than 30, which occupies 4 percent of the total sample in 2015Q4.
- Fiscal position note:
  - Fiscal deficit was 6.2 percent of GDP in 2014; expected to continue declining if government pursues FY2020 primary surplus goal, but net debt ratio expected to be stable for next few years and then start rising (IMF, 2016).

### IV. Evidence on myopia — discount wedge estimation
- Methodology:
  - Follows Bayoumi and Sgherri (2006) using a system of three equations (consumption, income, net tax rate) estimated by seemingly unrelated regression on annual data from 1980.
  - Model incorporates r (real interest rate) and λ (discount wedge/myopia).
- Key regression results (unrestricted model, JPN 1981–2014):
  - Consumption equation: αC = -.021 (.004)** ; βY = .80 (.05)** ; βT = -.10 (.04)** ; βe = .23 (.04)** ; R2 = 0.89.
  - Income equation: αY = .38 (.15)* ; θY = .067 (.03)* ; γ = -.001 (.0002)** ; R2 = 0.55.
  - Net tax rate equation: αT = -.02 (.002)** ; θT = .16 (.03)** ; δ = 1.4 (.21)** ; R2 = 0.50.
  - Note: * (**) denotes statistical significance at a 5 (1) percent level. Numbers in parentheses are standard deviation.
- Discount wedge estimates:
  - Discount wedge in Japan is around 0.1 over the full sample period (1980–2014).
  - Sensitivity: 0.11 and 0.095 when assuming a real interest rate of 4 percent and 0 percent, respectively. Both are statistically significant at the 1 percent level.
  - The average real interest rate for 1981–2014 was 1.5 percent.
  - Rolling 20-year window estimates suggest the additional discount rate has increased over time.
  - Average elasticity of consumption to net tax rate changes: 0.15 (first window) and 0.22 (last 5 windows), suggesting an increased tax multiplier.
- Interpretation:
  - Increasing discount wedge indicates Japan has become more myopic and thus less Ricardian.
  - Demographic driver: life expectancy minus median age (an indicator of national planning horizon) has been rapidly declining in Japan, implying shorter national planning horizons; IPSS projections suggest further narrowing.

### V. Evidence on liquidity constraints — recent developments and program evidence
- Macro and micro indicators pointing to stronger liquidity constraints:
  - Household saving rate has been on a declining trend (contrast with US and Germany). Sparsely timed rises in 1998, 2009, and 2011 may reflect downturns or stimulus.
  - Increase in non-regular workers among young males implies more workers with lower wages and less job security, raising exposure to liquidity constraints.
  - CAO (2015): younger households with lower income reduced consumption after the 2014 consumption tax hike more than other households.
  - Per person financial assets have declined for younger generations.
  - Pension benefits per person have been cut in nominal terms recently and even in real terms in the past few years; Gini coefficient among the elderly is larger compared to other generations—inequality in asset stock among the elderly is large.
- Evidence from past cash transfer programs:
  - 1999 shopping coupon program:
    - Coupon: 20 thousand yen per child aged 15 or below and eligible elderly; valid six months; marginal propensity to consume (MPC) of the coupon was 10 percent overall; higher for liquidity-constrained households.
  - 2009 cash benefit program (CAO, 2012):
    - Payment: one-time 20 thousand yen to each person aged 18 or below and 65 or above; 12 thousand yen to other cohorts.
    - MPC: 0.25 for the whole sample; 0.40 for households with children; 0.37 for households with elderly.
  - Interpretation of both episodes:
    - Japan is not fully Ricardian.
    - Liquidity constraints exist and appear binding for targeted groups.
    - The 1999 program (better targeted/time-limited) had a lower impact than 2009, suggesting the multiplier of cash transfers might have risen — consistent with Japan becoming less Ricardian.
- Public assistance trend:
  - Number of people on public assistance rose from less than a million to an all-time high of 2.2 million in the past twenty years.
  - Public expenditure on the program more than doubled from a bottom of (text truncated in source).

### VI. Policy implications (as discussed in the source)
- Structural increases in myopia and liquidity constraints imply:
  - Fiscal stimulus can be more effective than under full Ricardian equivalence assumptions, especially when targeted to liquidity-constrained households.
  - Consolidation effects may be mitigated by non-Ricardian behavior, but fiscal sustainability concerns remain given high public debt.
- Importance of demographic and distributional considerations:
  - Ageing and rising inequality among the elderly increase the relevance of liquidity constraints and short planning horizons in shaping fiscal multipliers.
- Need for policy design to account for changing degrees of Ricardian behavior over time when assessing stimulus vs. consolidation trade-offs.

*Source: _wp16194 - References (content excerpt)._

### 1.3 trillion yen (0.27 percent of GDP) before the collapse of the asset bubble to 3.6 trillion

### _wp16194 - 1.3 trillion yen (0.27 percent of GDP) before the collapse of the asset bubble to 3.6 trillion

### Evidence of rising liquidity constraints and demographic drivers
- Total pension benefits and monthly basic pension benefits per person series highlight rising pension outlays (figures shown for 1999–2016 and projections).
- The ratio of people on public assistance is empirically explained by a regression of the cohort-level ratio on its lagged value, the non-regular worker ratio, and the job-to-applicant ratio.
- Government estimate: 13.4 percent of workers are employed at close to the minimum wage (less than 15 percent more than the minimum wage) in 2014, up from 9.2 percent in 2009.
- An increase in the non-regular worker ratio has a statistically significant impact on public assistance participation, suggesting rising liquidity constraints.
- People aged 65 or above have a higher chance of being on the public assistance program than other cohorts, demonstrating the role of ageing.
- Even assuming a flattening of the non-regular worker ratio going forward, the number of people on public assistance is estimated to continue to rise, especially in terms of the ratio to the total population (the number in 2030 is calculated from the regression explained in footnote 9).
- These results support the argument that a larger share of households will become liquidity constrained, making Japan “even less Ricardian.”

### Multipliers and consumption tax experience
- A smaller Ricardian offset implies a higher fiscal multiplier.
- Using the IMF’s Global Integrated Monetary and Fiscal (GIMF) Model:
  - The consumption tax multiplier is estimated at around 0.4 using assumptions broadly common in advanced economies.
  - Calibrating parameters to reflect a higher ratio of liquidity constrained consumers raises the multiplier to around 0.6.
- Empirical estimate for the 2014 consumption tax hike using quarterly data for 1998Q1–2013Q3 and the CAO (2015) methodology:
  - The implied multiplier amounts to 0.9 (national accounts data) or 0.6 (the BoJ’s new consumption indicator).
  - The 0.6 estimate is close to the tailored GIMF simulation result.
- Regression specification used to estimate private consumption:
  - Ct = α1 Yt + α2 OLDt + α3 Yt*OLDt + α4 FAt-1 + α5 CCIt + α6 E1 + α7 E2 + εt
  - C: private consumption; Y: compensation; FA: household’s net financial assets (all in real and logarithm); OLD: dependency ratio (ratio of people aged 60 or above to the total population); CCI: consumer confidence index; E1 and E2: dummies for 2011Q1 and Q2.

### Timing and planning-horizon scenarios for consolidation
- Illustrative consolidation simulation: a 10 percentage point consumption tax increase distributed over 10 years (1 point each year).
- Two timing scenarios compared:
  - (i) Start a 10 percentage point consumption tax increase over 10 years 10 years from now; assumed planning horizon of 6 years.
  - (ii) Start the same adjustment 5 years from now; assumed planning horizon of 10 years.
- Simulation result: it is advisable to start fiscal consolidation sooner rather than later because longer planning horizons lead people to better prepare (save more), making the negative impact of the consumption tax increase more benign.
- Impact of liquidity constraints on consolidation:
  - Two liquidity-constrained consumer ratios assumed: 30 percent and 40 percent.
  - The negative impact of fiscal consolidation is larger when liquidity constraints are more binding; the first-year multiplier is estimated at -¾ in the higher-constraint case.
  - Given the possibility of rising liquidity constraints, these results argue for starting consolidation soon.

### Financing Japanese Government Bonds (JGBs) and demand-supply outlook
- Becoming less Ricardian implies less household savings, reducing the principal domestic source of public debt financing.
- Supply of JGBs calculation: current outstanding amount (including issuance by the Fiscal Investment and Loan Program (FILP)) plus projected fiscal deficits.
- Key demand assumptions:
  - The BoJ will stop increasing its JGB holdings by the end of 2017 and keep the outstanding amount at that time for the future.
  - Total savings at depositary institutions by the household and non-financial private sectors will continue to increase at the same pace as in the past 5 years; assumptions on their JGB holdings-to-deposit ratio come from Han (forthcoming).
  - Household assets managed by pension funds and insurers will continue to increase at the same pace as in the past 5 years; portfolio allocation assumed to follow the Government Pension Investment Fund’s target in the next few years.
- Illustrative calculation suggests the supply of JGBs exceeds demand as early as 2018.
  - At that moment, interest rates may need to rise to attract enough buyers, including domestic depository institutions holding excess reserves at the BoJ’s current account (white dotted area in the chart).
  - Private pension funds and other insurers may reverse portfolio rebalancing towards riskier assets.
- In the long run, the savings rate is expected to decline further due to ageing, myopic behavior, and rising liquidity constraints, resulting in a slower increase or decline in household financial assets.
- Implication: pressure on government bond yields may start to be felt possibly in the near future as financing needs are increasingly met from external sources; efforts to restore fiscal sustainability should start soon and be sustained.

### Policy recommendations and conclusions
- Findings:
  - Japan has likely become, and may continue to become, less Ricardian with ageing and weak economic prospects.
  - The estimate of the discount wedge suggests Japan may have become more myopic; a declining average remaining life expectancy plausibly shortens planning horizons.
  - Several indicators suggest liquidity constraints have become more binding in Japan.
- Policy implications:
  - Stronger fiscal multipliers raise the effectiveness of fiscal policy as a counter-cyclical tool.
  - Reorienting expenditure toward liquidity constrained households is expected to enhance fiscal policy impact.
  - Fiscal consolidation should start soon in a gradual, yet steady, manner.
  - A potential reduction in savings poses a challenge for JGB financing in the not-too-distant future.
  - Given the large impact of the previous consumption tax hike, more gradual increases seem warranted.
  - Structural reforms to address ageing and rising liquidity constraints are indispensable.

*Source: IMF working paper content provided in the supplied text.*

---


_Source: https://www.imf.org/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2016/_wp16194.pdf_