Japanese savings may support JGBs, but the key effect is more about policy signaling than direct buying

Overview

This report discusses two channels through which Japanese policymakers could support JGBs by directing domestic savings into Japanese assets: first, adjusting GPIF asset allocation to raise the weight of domestic bonds; second, allowing JGBs to be included in NISA tax-advantaged savings accounts. The report argues that both measures could increase JGB demand, but the scale of direct buying is limited, with greater importance attached to policy signaling, market expectations, and positive feedback from follow-on buying by domestic investors.

Core views

The core view is that Japanese domestic savings do have the potential to provide marginal support for JGBs, especially at the 20y and 30y long end and the corresponding swap spreads; however, current cheapening in the JGB curve is mainly driven by macro factors such as fiscal policy, monetary policy, and the inflation path. Therefore, without support from fiscal or monetary policy, supply-demand measures alone are unlikely to deliver a sustainable decline in yields and volatility.

Analysis framework

The report uses a flow-of-funds and relative-value framework to estimate GPIF's room to add JGBs within its current asset allocation range, its historical rebalancing behavior, its duration management approach, and the potential impact of NISA accounts on direct household holdings of JGBs, and then maps these demand shocks to 30y swap spreads and long-end JGB curve performance.

Methodology notes

• Flow of Funds AnalysisGPIF Asset Allocation Capacity Estimation

  Under GPIF's current 25% domestic bond benchmark and +/-6ppt range, estimate the potential room for domestic bond allocation to rise from about 27% to 31%.

  Based on this, the report estimates about USD75bn of additional JGB purchase capacity, and emphasizes that while this size is not particularly large relative to routine rebalancing flows, the policy signal value of active reallocation is greater.

• Spread SensitivitySensitivity of 30y Swap Spreads to Changes in Free-Float Supply

  In UST, Bund, and Gilt frameworks, a 1ppt change in the share of debt held by the private sector corresponds to about 1bp in 30y swap spreads; the impact could be greater if buying is concentrated in the 30y sector or triggers more domestic buyers to follow.

  The report maps a rise in GPIF's domestic bond allocation from 27% to 31% to an estimated 3-6bp impact on 30y swap spreads.

• Retail Flow BehaviorNISA Tax-Advantaged Account Demand Assessment

  Assess whether allowing direct JGB holdings in NISA can materially increase household allocations to JGBs.

  The report argues that low household allocation is more likely due to historically low yields; the Italian experience shows that rising retail holdings usually coincides with rising yields, not falling yields.

Asset mapping & comparison

Structured mapping from thesis to named assets (strengths, weaknesses, peers, risks).

• JGBs

  Core asset under discussion

  Strengths

  Both GPIF and NISA measures could marginally increase demand, and policy signaling could stabilize market sentiment.

  Weaknesses

  The scale of direct purchases is limited, foreign ownership is already low, and curve cheapening is mainly driven by macro policy.

  Comparison

  Compared with the 10y sector and the belly of the curve, the 20y and 30y sectors are more likely to benefit from a repair in supply-related cheapening.

  Risks

  If fiscal policy, monetary policy, and the inflation path do not improve, demand measures are unlikely to deliver sustained yield declines.

• 20y/30y JGBs and Swap Spreads

  Curve positions the report relatively favors

  Strengths

  The long end is more sensitive to targeted buying and policy signaling, and the report believes the 20s and 30s have more room to strengthen.

  Weaknesses

  The adjustment pace may be slow, and the market has already experienced sharp volatility and pullbacks.

  Comparison

  More likely than the 10s and the belly of the curve to receive support from the supply-demand side.

  Risks

  If the market interprets policy communication as more equity-focused than bond-focused, long-end support could reverse.

• USTs, Bunds, Gilts and other G10 sovereign bonds

  Cross-market reference points and potential funding source

  Strengths

  Useful for estimating swap spread sensitivity to changes in free-float supply.

  Weaknesses

  GPIF holds less than 1pp of major overseas bond markets, so a modest shift toward JGBs would have little direct impact.

  Comparison

  The report does not view this factor as a key driver across markets or for swap spreads.

  Risks

  Cross-market effects would likely broaden only if domestic investors more broadly rotate from overseas bonds into Japanese assets.

• Japanese Equities

  Potential domestic asset direction in policy communication

  Strengths

  Recent communication appears more tilted toward supporting domestic equities, which could attract domestic savings.

  Weaknesses

  If policy focus shifts from bonds to equities, support for JGBs would weaken.

  Comparison

  Relative to JGBs, recent communication has been interpreted by the market as more supportive of domestic equities.

  Risks

  Unclear asset allocation policy objectives could lead to repeated swings in bond market expectations.

Key data

• GPIF Asset SizeAbout USD2.8tnAs of end-March 2026, GPIF managed about USD2.8tn in assets.
• GPIF Current Share of the JGB MarketAbout 5%The report states that GPIF currently holds about 5% of the JGB market.
• GPIF Room to Increase JGB AllocationAbout USD75bnWithin the currently permitted strategy range, domestic bond allocation still has about USD75bn of room to increase into JGBs.
• Domestic Bond Allocation Scenario27% to 31%This scenario is used to estimate the impact on 30y swap spreads.
• Estimated Impact on 30y Swap SpreadsAbout 3-6bpThe report believes this impact is comparable to this week's roughly 8bp round-trip move in 30y swap spreads.
• Historical Fluctuation in GPIF Domestic Bond AllocationAbout 1-2ppt per yearTherefore, even if reallocation occurs, the pace of portfolio adjustment may be relatively slow.
• Foreign Ownership Share of Japanese Government BondsAbout 14%Foreign ownership of JGBs is already at a relatively low level within G10, limiting further scope to replace foreign capital with domestic savings.
• NISA Account ScaleMore than 20mn accounts, with transaction value above JPY71tn as of end-2025This shows broad NISA usage, but the current growth allowance can invest in JGB funds, not directly in JGBs.
• Household Direct Holding Share of JGBsAbout 4%Excluding NISA accounts, households directly hold JGBs at about 4% of portfolio assets.

Impact & implications

In investment terms, the report is more supportive of focusing on relative strengthening at the long end of the Japanese rates curve and widening in 20y/30y swap spreads, rather than betting on a sustained decline in overall JGB yields driven by repatriated savings. The impact on other G10 bond markets is considered more limited unless policy expands further into a systematic shift by domestic investors away from overseas bonds and toward Japanese assets.

Risks

• GPIF reallocation size is about USD75bn, which is not large relative to existing rebalancing flows, so direct buying may be insufficient to keep yields sustainably lower.
• Historically, GPIF's domestic bond allocation has fluctuated by only about 1-2ppt per year, so the actual implementation pace may be slow.
• Including JGBs in NISA may not materially change household allocation, because the underweight in JGBs may be mainly explained by historically low yields.
• JGB curve cheapening is mainly driven by fiscal policy, monetary policy, and the inflation path; supply-demand measures may only damp volatility rather than solve the root cause.
• If policy communication tilts more toward domestic equities than bonds, the initial gains in JGBs could continue to retrace.
• If the market over-extrapolates GPIF reallocation to the broader pension industry, short-term price reactions may overstate actual fund flows.

What to watch

• Whether GPIF adjusts its domestic bond allocation range or actual allocation pace before 2030.
• Whether the Japanese government formally allows direct JGB investment through NISA, and the design of investable scope and limits.
• Performance of 20y and 30y JGBs relative to the 10y sector and the belly of the curve.
• Whether 20y and 30y swap spreads continue to widen.
• Whether Japan's fiscal policy, monetary policy, and inflation path show substantive changes that can reduce rates volatility.
• Whether domestic investors more broadly reduce overseas bond holdings and rotate into Japanese assets.