GPIF Japan 25/25/25/25

GPIF basic portfolio (Government Pension Investment Fund, Japan): 25/25/25/25

Government Pension Investment Fund (GPIF), Board of Governors; targets set under medium-term objectives from the Minister of Health, Labour and Welfare · Equal 25/25/25/25 mix effective 1 April 2020 (confirmed unchanged for 2025-2030); previous mixes 2014-2020 (35/15/25/25) and before 2013-2014 (bond-heavy)

Live account on this site. This portfolio is also tracked here with a virtual $100,000 since 21.10.2026. Live account on this site →

Where it stands today

Data through 2026-09. Updated every month.

Since January+3.5 %

Last 12 months+6.2 %

Last 5 years, per year+4.3 %

Now below the peak−2.6 %

Place by 12 months51 / 60

Our calculation, 1926-07 – 2026-09

Per year7.5 %

Real (after inflation)4.4 %

Volatility9.7 %

Worst drop−53.5 %

Worst year−21.5 % (1931)

Best year31.2 %

Worst 12 months−35.7 %

Positive years77 %

37 % of this history is calculated with stand-ins for asset classes that did not exist yet (see “How we model it” and the page about data). Without stand-ins the history starts in 1990.

Growth of $10,000

GPIF Japan 25/25/25/25: $14,704,197S&P 500: $196,488,890

Where it came from

GPIF is the public fund that holds the reserves of Japan's two public pension schemes (Employees' Pension Insurance and National Pension). Japan runs a pay-as-you-go system, and GPIF's own documents explain that the reserve exists to top up benefits for later generations because the falling birthrate and ageing population make it too heavy for working-age contributors to fund benefits alone. Government statements and commentary describe it as the world's largest pension fund. At end of fiscal 2024 (31 March 2025) it managed about 249.8 trillion yen. A secondary source from 2010 notes that the current governance structure took effect in 2006, when GPIF became an independent administrative agency; before that the reserves had been run under a single-manager arrangement.

Until the early 2010s the portfolio was extremely home-biased and bond-heavy. A 2010 article based on an OECD review put roughly 67 percent of holdings in Japanese bonds and 11 percent in domestic equities, and the OECD criticised the fund for a conservative, domestic allocation, for combining chairman, CEO and chief investment officer roles in one person, and for no visible link to pension liabilities. In 2013 a government-appointed expert panel led by Takatoshi Ito (the 'Ito paper', November 2013, as reported by news coverage) pushed for a less conservative strategy, and commentary at the time framed this as part of Prime Minister Abe's growth agenda. Reports also say a first reshuffle since 2006 lowered the bond target from 67 to 60 percent in 2013.

On 31 October 2014 GPIF revised its medium-term plan and basic portfolio. Reported targets before the revision were 60 percent domestic bonds, 11 percent foreign bonds, 12 percent domestic equities, 12 percent foreign equities (and about 5 percent short-term assets); after the revision they became 35, 15, 25 and 25. The same revision removed the separate short-term bucket and introduced alternatives (infrastructure, private equity, real estate) at up to 5 percent, counted inside the four main classes. This 2014 change roughly doubled the equity share to 50 percent and is the single most cited allocation shift in the fund's history.

In the fourth medium-term plan (fiscal 2020 to 2024) GPIF changed the mix again, to 25 percent in each of four classes: domestic bonds, foreign bonds, domestic equities, foreign equities. GPIF's own summary says this was the first policy-portfolio formulation since the integration of the employees' pension schemes and the creation of the Board of Governors; it took 13 deliberations of the Board of Governors, and a working sub-committee of economics and finance experts formed in February 2018 met 32 times. The portfolio was chosen to meet the medium-term objective of a real return (nominal return less nominal wage growth) of 1.7 percent with the lowest risk, using the 2019 actuarial valuation (released 27 August 2019). GPIF stated the reasons as historically low policy rates in developed countries and in Japan (Japan had negative interest rates from January 2016 and yield curve control from September 2016) and wrote that the result was more foreign bonds and fewer domestic bonds. New deviation limits for total bonds and total equities (both plus or minus 11 percent) were added to control equity risk. It took effect on 1 April 2020.

For the fifth medium-term plan (April 2025 to March 2030) the Board of Governors discussed six times (after twenty meetings of a project team set up in fiscal 2022) and kept the 25/25/25/25 targets. The stated reason: using expected returns based on the fiscal 2024 financial verification and updated risk and correlation data, a portfolio with a real return of 1.9 percent at minimum risk still had the same weights. What changed was the width of the bands, which were narrowed (for example domestic bonds from plus or minus 7 to 6 percent, bonds-as-a-whole and equities-as-a-whole from 11 to 9 percent). Alternative assets stay capped at 5 percent of the total and are managed inside the four main classes according to their risk and return profile.

What is in it

Domestic (Japanese) bonds	25%	Target from 1 April 2020 and again from 1 April 2025. Deviation limit plus or minus 6 percent in the fifth period (plus or minus 7 percent in the fourth).
Foreign bonds	25%	Deviation limit plus or minus 5 percent in the fifth period (plus or minus 6 percent in the fourth). Foreign bonds with currency hedge are classified as domestic bonds.
Domestic (Japanese) equities	25%	Deviation limit plus or minus 6 percent (fifth period); plus or minus 8 percent (fourth).
Foreign equities	25%	Deviation limit plus or minus 6 percent (fifth period); plus or minus 7 percent (fourth). Total bonds and total equities each have a band of plus or minus 9 percent (fifth period); plus or minus 11 percent (fourth).

Source of the shares: GPIF, Policy Asset Mix for the Fifth Medium-Term Objectives Period - Summary (2025); and Adoption of New Policy Portfolio (2020) — gpif.go.jp. The 2025 summary states 25 percent for each of the four classes and shows old and new deviation limits. The 2020 document gives the earlier 35/15/25/25 targets (limits 10, 4, 9, 8 percent) and the new 25/25/25/25 targets (limits 7, 6, 8, 7 percent).

How we model it

Developed markets outside the US	31.2%
Intermediate government bonds (10 y)	25%
US bond market (aggregate)	25%
US large companies (S&P 500)	16.2%
Emerging markets	2.5%

APPROXIMATION. Our series are in US dollars and contain no Japanese assets, no yen and no JGB series. Japanese equities (25 percent) are replaced by intl_dev (developed ex-US, of which Japan is a part). Foreign equities (25 percent) are split roughly 65/25/10 into us_large / intl_dev / em as a rough proxy for a global ex-Japan equity index; this split is our assumption and was not checked against GPIF's benchmark weights. Japanese government bonds (25 percent) are replaced by bond_mid (7-10 year government bonds); foreign bonds (25 percent) by bond_agg. GPIF's own benchmarks (per its 2020 document: Nomura BPI excl. ABS for domestic and a world government bond index excluding Japan for foreign bonds) are not in our data. Currency effects, yen hedging and the real GPIF liability (wage growth in yen) are not modelled, so the result shows how a 50/50 global stock-bond mix behaves in dollars, not how GPIF itself performed.

For a euro investor

The same portfolio bought with euros, without currency hedging: monthly returns converted at the end-of-month rate (German mark before 1999), inflation of the euro area. History in euros from 1953-10 to 2026-09.

Per year in euros6.8 %

After euro inflation4.0 %

Worst drop in euros−40.6 %

Worst year in euros−22.3 % (1973)

Safe withdrawal, 30 years, dollars3.78 % (1968-12)

Safe withdrawal, 30 years, euros2.77 % (1968-12)

The largest first-year withdrawal, then growing with inflation, that lasted 30 years from every monthly start. Test withdrawals for this portfolio →

How to build it with European funds

Examples of UCITS funds sold in the EU for each part of the portfolio. They track similar markets, not exactly the indices of our calculation.

Part	Share	Example funds (ISIN)
Developed markets outside the US (closest match)	31.2%	iShares Core MSCI Europe UCITS ETF EUR (Acc) `IE00B4K48X80` iShares Core MSCI Japan IMI UCITS ETF `IE00B4L5YX21`
Intermediate government bonds (10 y)	25%	iShares USD Treasury Bond 7-10yr UCITS ETF (Dist) `IE00B1FZS798`
US bond market (aggregate) (closest match)	25%	iShares Core Global Aggregate Bond UCITS ETF EUR Hedged (Acc) `IE00BDBRDM35` iShares Core Global Aggregate Bond UCITS ETF USD (Dist) `IE00B3F81409`
US large companies (S&P 500)	16.2%	iShares Core S&P 500 UCITS ETF USD (Acc) `IE00B5BMR087` Vanguard S&P 500 UCITS ETF (USD) Distributing `IE00B3XXRP09`
Emerging markets	2.5%	iShares Core MSCI Emerging Markets IMI UCITS ETF (Acc) `IE00BKM4GZ66`

These are examples, not recommendations. Check the fund's key information document, costs and taxes with your broker before buying. Names and ISINs checked on justETF on 4 November 2026; Udak receives nothing from these funds.

The idea

GPIF's stated aim is to secure, at the lowest possible risk, the investment return needed for long-term pension finance: a real return (nominal return minus nominal wage growth) of 1.7 percent for the 2020-2024 plan and 1.9 percent in the 2025 analysis. The downside risk of falling short of nominal wage growth may not exceed that of a portfolio of domestic bonds only. Portfolio optimisation used expected returns, risk and correlations for the four asset classes; the equal 25 percent split came out of this process, not from a rule of thumb. GPIF adds that the mix reflects diversification across both asset type and currency/region.

How people buy it

GPIF itself does not hold one product: it invests via passive and active mandates with external managers and index-tracking, under benchmarks for each of the four classes. Per its 2020 document, domestic bonds are benchmarked to a Nomura BPI-type index and foreign bonds to a world government bond index excluding Japan. For a private investor the closest replicas are a global equity index fund, a Japan equity ETF, a Japanese government bond fund and a global bond fund, but the fund's currency-risk treatment cannot be copied exactly.

Rebalancing as published

GPIF uses permitted ranges around each target rather than a fixed calendar. Fifth-period ranges: domestic bonds plus or minus 6 percent, foreign bonds plus or minus 5, domestic equities plus or minus 6, foreign equities plus or minus 6, and a combined band of plus or minus 9 percent for total bonds and for total equities. The policy portfolio is reviewed every five years with the medium-term plan, and GPIF says it will check in a timely way whether the environment has moved away from the one assumed and review if necessary.

Variants

Fiscal 2015-2019 basic portfolio: 35% domestic bonds, 15% foreign bonds, 25% domestic equities, 25% foreign equities (limits plus or minus 10, 4, 9, 8 percent).
Before October 2014 (reported): 60% domestic bonds, 11% foreign bonds, 12% domestic equities, 12% foreign equities, about 5% short-term assets.
Pre-2013 (secondary source, 2010): roughly 67% Japanese bonds and 11% domestic equities.

Criticism

Governance critique (OECD, via a 2010 article): domestic and conservative allocation, no clear written investment policy, concentration of chairman/CEO/CIO roles, and no clear link to liabilities.
Political pressure critique: when the equity share was raised in 2013-2014, commentators compared it with 1990s attempts to use pension money to support the stock market; GPIF's president at the time was reported as calling such demands unfair to an institution independent since 2006 (secondary press summaries, not verified against the originals).
The mandate's risk constraint itself notes that downside risk for equities may be larger than expected (GPIF 2020 medium-term objective), i.e. the shift towards 50 percent equities raises loss risk in a crisis.
An equal-weight four-way split is a simple result of optimisation and ignores the fact that Japanese equities and bonds are a tiny share of world markets, which leaves a large home bias (about half of the fund in Japan-linked assets before currency hedge classification).

Through the eras

Era	Dates	Total	Per year	Worst drop	S&P 500
Roaring Twenties boom	1926-07 – 1929-08	+73.2 %	+19.0 %	−2.8 %	+153.7 %
Wall Street crash and Great Contraction	1929-09 – 1932-07	−47.6 %	−19.8 %	−53.5 %	−78.2 %
New Deal recovery	1932-08 – 1937-02	+120.7 %	+18.8 %	−13.9 %	+268.7 %
Recession within the Depression	1937-03 – 1938-03	−25.1 %	−23.4 %	−25.1 %	−49.5 %
Rearmament and World War II	1938-04 – 1945-09	+96.8 %	+9.4 %	−14.0 %	+205.8 %
Postwar expansion	1945-10 – 1965-12	+396.1 %	+8.2 %	−12.4 %	+1212.9 %
Golden age of the 1950s	1950-01 – 1959-12	+160.8 %	+10.1 %	−7.6 %	+438.8 %
Go-go years and nifty-fifty	1966-01 – 1968-12	+19.7 %	+6.2 %	−8.7 %	+34.0 %
The Great Inflation (stagflation)	1969-01 – 1982-07	+101.7 %	+5.3 %	−24.2 %	+84.6 %
1973-74 oil shock and bear market	1973-01 – 1974-12	−18.9 %	−10.0 %	−24.2 %	−41.7 %
Second oil shock and the Volcker disinflation	1979-01 – 1982-08	+43.1 %	+10.3 %	−9.8 %	+59.2 %
Great bull market in stocks and bonds	1982-08 – 1999-12	+802.2 %	+13.5 %	−16.6 %	+2060.4 %
Black Monday	1987-08 – 1987-12	−11.3 %	—	−16.6 %	−21.8 %
Savings-and-loan crisis and 1990 recession	1990-07 – 1991-03	+3.5 %	—	−10.7 %	+7.4 %
1994 bond market massacre	1994-02 – 1994-12	−3.9 %	—	−4.9 %	−3.2 %
Asian and Russian financial crises	1997-07 – 1998-10	+11.2 %	+8.3 %	−5.7 %	+22.4 %
Dot-com bubble and crash	2000-03 – 2002-10	−8.1 %	−3.1 %	−12.1 %	−37.2 %
Recovery and housing boom	2002-11 – 2007-09	+84.1 %	+13.2 %	−3.3 %	+97.0 %
Global financial crisis	2007-10 – 2009-03	−18.8 %	−12.9 %	−24.3 %	−44.7 %
The decade after the crisis	2009-04 – 2019-12	+133.5 %	+8.2 %	−7.0 %	+411.2 %
Fourth-quarter 2018 sell-off	2018-09 – 2018-12	−5.6 %	—	−5.6 %	−14.4 %
Covid crash and stimulus boom	2020-02 – 2021-12	+18.4 %	+9.2 %	−7.9 %	+53.6 %
Inflation shock and rate hikes	2022-01 – 2022-12	−15.7 %	−15.7 %	−21.0 %	−19.9 %

Calendar years

Any N-year period

Any N-year period	Worst	10 % worse than	Median	10 % better than	Best
1 y (1192)	−35.7 %	−4.7 %	+8.5 %	+20.4 %	+64.5 %
3 y (1168)	−20.6 %	+0.3 %	+7.7 %	+14.0 %	+26.1 %
5 y (1144)	−6.7 %	+3.4 %	+7.3 %	+12.4 %	+23.3 %
10 y (1084)	+0.6 %	+4.6 %	+7.5 %	+12.2 %	+15.8 %
15 y (1024)	+2.5 %	+5.4 %	+7.6 %	+12.0 %	+14.1 %
20 y (964)	+3.3 %	+5.9 %	+7.8 %	+11.3 %	+13.1 %
30 y (844)	+5.7 %	+6.7 %	+8.0 %	+10.1 %	+11.0 %

How it behaved in history

Before 2013-2014. Bond-dominated: about 67 percent Japanese bonds and 11 percent domestic equities per a 2010 article summarising the OECD review. www.ai-cio.com
31 October 2014. Domestic bonds 60 to 35, domestic equities 12 to 25, foreign equities 12 to 25, foreign bonds 11 to 15 (percent). Equity share about doubled to 50 percent. gendai.media
2016. Bank of Japan negative interest rates from January 2016 and yield curve control from September 2016; GPIF cites falling yields on its benchmark bond indexes as one reason for cutting domestic bonds in 2020. www.gpif.go.jp
April 2020. 25/25/25/25 mix takes effect; in the same document GPIF says it will manage flexibly within bands after 'extreme' market volatility at the start of the COVID period. www.gpif.go.jp
April 2025. Fifth plan keeps the weights and narrows bands. gpif.go.jp

Published results (by others)

Time-weighted return of the whole fund in fiscal 2024 (FY2024 (April 2024 - March 2025)): +0.71 percent; asset size 249,782.1 billion yen www.gpif.go.jp
Annualised nominal return since fiscal 2001 (marked to market, includes unrealised gains) (FY2001-FY2024): 4.20 percent per year; cumulative gain 155,531.1 billion yen www.gpif.go.jp
Actual allocation at end of March 2025 versus the 25 percent targets (31 March 2025): domestic bonds 27.64, foreign bonds 24.37, domestic equities 23.94, foreign equities 24.05 percent www.gpif.go.jp
Return GPIF needs from the policy mix to meet its objective (FY2020 plan / FY2025 plan): real return (nominal minus wage growth) of 1.7 percent (2020) and 1.9 percent (2025 analysis) gpif.go.jp

What we could not verify

Exact pre-2014 weights (60/11/12/12 plus about 5 percent short-term) come from a Japanese press summary and a trade-press snippet; GPIF's own 2014 announcement was not opened.
The 2013 reduction of bonds from 67 to 60 percent is from a news snippet and was not cross-checked against GPIF documents; the 2014 old-versus-new split above may overlap with it.
The Ito panel details and the criticism of political pressure come from press summaries of FT/Reuters-type articles; originals were paywalled or not opened.
Our regional split of foreign equities (65/25/10) is an assumption, not taken from GPIF.
Whether GPIF treats foreign bonds with a currency hedge as domestic in practice, and how much of the foreign bond holding is hedged, was not checked.

Sources

gpif.go.jp https://gpif.go.jp/en/15324685gpif/policy_asset_mix_for%20the%205th-term_summary.pdf
www.gpif.go.jp https://www.gpif.go.jp/en/topics/Adoption_of_New_Policy_Portfolio_details_en.pdf
www.gpif.go.jp https://www.gpif.go.jp/en/performance/32821257gpif/annual_report_summary_2024_en.pdf
gendai.media https://gendai.media/articles/-/40988
www.top1000funds.com https://www.top1000funds.com/uncategorised/gpif-continues-equities-rampage/
www.ai-cio.com https://www.ai-cio.com/news/japans-national-pension-the-worlds-largest-faces-reform-pressure/
www.cnbc.com https://www.cnbc.com/id/100775786
business.columbia.edu https://business.columbia.edu/sites/default/files-efs/imce-uploads/CJEB/Events/2017-2018/Shiozaki%20final%20speech.pdf
www.business-standard.com https://www.business-standard.com/article/international/abe-s-reforms-take-aim-at-japan-s-1-trillion-pension-fund-113070300013_1.html
www.euro2day.gr https://www.euro2day.gr/ftcom_en/article-ft-en/1184618/japan-pension-fund-warns-of-unfair-abe-pressure.html

Open in the builder

Put this card on your site →

Calculated on historical data. Past results do not predict the future. Models before the 1990s use stand-ins for classes that had no index yet; fees, taxes and fund costs are not included. This is mathematics, not an investment recommendation.