Guan Tao: Don't Speak Hastily of a New Plaza Accord

On July 30 this year, the Japanese government once again entered the market to intervene against yen depreciation. On July 31, Japan and the United States went further, conducting high-profile joint intervention. The previous joint intervention in the yen exchange rate occurred in March 2011, when the Group of Seven (G7) acted together to curb excessive yen appreciation and prevent negative impacts on Japan's post-disaster economic reconstruction and the global economy. Earlier still, in June 1998, Japan and the US joined forces to support the yen, preventing its excessive depreciation from aggravating the contagion of the Asian financial crisis. The current round is likewise aimed at stemming yen depreciation. Amid these developments, talk of a new Plaza Accord has gained traction. The author, however, argues that this is merely a continuation of the international exchange rate policy coordination that has been ongoing since the original Plaza Accord of September 1985. Within the current international monetary system framework, non-intervention remains the principle, intervention is the exception, and joint intervention is the exception to the exception.

Five Rounds of Intervention: "More Failures Than Successes"

Since 2022, with the yen/dollar rate repeatedly hitting multi-decade lows, the Japanese government has intervened multiple times by selling dollars to buy yen. According to disclosures from Japan's Ministry of Finance, as of the second quarter of this year, Japan has conducted five rounds of intervention: on September 22, 2022; October 21 and 24, 2022; April 29 and May 1, 2024; July 11 and 12, 2024; and April 30, May 4 and 6 this year. The actions on July 30 and 31 this year constitute the sixth round (see Chart 1).

Measuring intervention effectiveness by how long it takes the yen to fall back below pre-intervention levels, the first, third and fifth rounds were less effective, lasting only 2, 39 (about 1.5 months) and 30 trading days (one month) respectively. The second and fourth rounds were more effective, lasting 264 days (12 months) and 510 trading days (21.5 months) respectively (see Chart 1). On closer examination, however, even these results are hard to attribute to FX intervention alone.

After the second round of intervention, between October 2022 and October 2023, the US dollar index trended weaker. In the first three quarters of 2022, the dollar index had risen 16.9% cumulatively, but in the fourth quarter, with US inflation peaking triggering expectations of a Federal Reserve rate hike slowdown, combined with China's pandemic optimization boosting global risk appetite, the dollar's interest rate and safe-haven support weakened and the index fell 7.7% that quarter. This reinforced the yen's rebound, which by January 13, 2023 had climbed to 127.85, a 17.4% recovery from the pre-second-round level. In 2023, the Fed's rate hike cycle peaked, and rapid inflation decline spurred rate cut expectations; the dollar index fell, rose, then fell again, closing the year down 2.0%. From October 20, 2022 to October 25, 2023, the dollar index dropped a cumulative 5.6%, while the yen exchange rate moved in a round trip, falling back below 150 to 1. At the start of 2024, unexpectedly strong US economic data and sticky inflation significantly postponed Fed rate cut expectations, driving the dollar index up 4.9% in the first four months, and the yen fell further toward 160, prompting the third round of intervention (see Chart 2).

After the fourth round of intervention, from July 2024 to April 2026, the dollar index fluctuated weaker. In Q3 2024, with the Fed kicking off its rate cut cycle as expected and policy expectations narrowing for non-US economies, the dollar index fell 4.8% that quarter, wiping out all its first-half gains. As a result, the yen climbed to 140.61 on September 16, a 15.0% rebound from pre-intervention levels. In Q4, the US election revived "Trump trades" and the dollar index strengthened, gaining 7.7% that quarter, with the yen falling back below 150. In 2025, the US government launched a global tariff storm, the "American exceptionalism" narrative collapsed and dollar credit cracks widened; in the first half, the dollar index fell 10.8% cumulatively, the yen broke back above 150, though its peak of 140.85 did not exceed the previous high. In the second half, easing tariff tensions and rate cut expectation revisions pushed the dollar index up 1.5% in stages, and the yen fell back below 150. In 2026, geopolitical conflicts boosted safe-haven demand and inflation expectations; the dollar index first fell then rose, climbing 2.4% from its intra-year low by end-April, with a maximum rebound of 4.9%, and the yen fell back below 160, triggering the fifth round of intervention (see Chart 2).

It can thus be seen that over the past four years, FX interventions mostly had a short-term supportive effect but could not reverse the yen's long-term weakness. Even the two relatively successful rounds were largely due to favorable changes in the external environment later on — what might be called "help from above." For the current round, the market generally believes that, with Japan and the US still facing wide negative interest rate differentials and significant difficulty in coordinating fiscal and monetary policy, the yen's rebound is likely to be short-lived. By August 14, the yen had closed at 159.33, a retreat of 2.11 yen from the August 3 pre-round high close (see Chart 2).

Decoding the Secrets of FX Intervention

On March 31, 2020, the Federal Reserve announced the launch of the Foreign and International Monetary Authorities Repo Facility (FIMA) as a temporary liquidity tool in response to the pandemic, and in July 2021 it became a standing facility. Its core function is to allow foreign central banks and international monetary authorities to use their holdings of US Treasuries as collateral to obtain short-term dollar financing from the New York Fed, avoiding direct selling of Treasuries that would disrupt the market. In the most recent round of intervention, US Treasury Secretary Bessent, while hinting at providing Japan with $5-10 billion in assistance, publicly called on the Fed to expand the FIMA limit for individual counterparties to back Japan's FX intervention efforts.

When FIMA is activated, the Fed's balance sheet records it as: an increase in the asset-side "Repurchase Agreements — Foreign Official" account, with a corresponding increase in the liability-side "Deposits of Foreign Official Accounts." From Fed disclosures, the weekly average balance and week-over-week changes in this "Repurchase Agreements — Foreign Official" account have been consistently small or even zero. Even during the fourth round of intervention, when the weekly average balance rose nearly $100 million week-on-week, that was still a drop in the bucket compared with Japan's intervention scale of over $30 billion at the time. To date, Japan has not confirmed actual use of the FIMA facility, only stating it is "planning/preparing to use" it.

Japan's底气 for not using FIMA lies in its large and highly liquid FX reserves. Japan is the world's second-largest foreign reserve holder after China, with reserves of $1.09 trillion as of end-July this year. By currency, the dollar dominates at over 90% of the total; by asset, reserves consist of securities and deposits, with deposits consistently accounting for more than 10%, reaching 14.9% (a balance of $162.3 billion) as of end-July, and a sizable share of the securities portion consists of short-term securities. During the first five rounds of intervention, the deposit component of Japan's FX reserves rose rather than fell, increasing by 2 million, $910 million, $1.25 billion and other amounts respectively.

The US Treasury's TIC (Treasury International Capital) report provides supporting evidence. Looking at the share of short-term Treasuries in Japanese investor holdings, after each of the first five rounds of intervention the share declined: after rounds 1 and 2, a cumulative decline of 1.2 percentage points; after round 3, a cumulative decline of 2.8 percentage points; after round 4, a quarter-on-quarter decline of 0.3 percentage points; and after round 5, a cumulative decline of 4.6 percentage points (see Chart 4). According to detailed TIC data disclosed since February 2023, during rounds 3 through 5, Japanese investors net reduced their short-term Treasury holdings by $36.1 billion and other amounts respectively.

Japan's intervention has had limited impact on short-term US Treasury market liquidity. Across the six rounds, comprising 12 trading days, in 3 of those days 3-month and 1-year Treasury yields rose week-on-week, with unchanged or declining in the remaining three-quarters. Across the six rounds, 3-month and 10-year Treasury yields rose week-on-week in 2 and 1 rounds respectively, with unchanged or declining in the remaining three-quarters (see Chart 5). Japan likely used short-term Treasuries held to maturity, or used deposits in its FX reserves as a "bridge" to spread out the selling of short-term Treasuries over time.

However, Japan's FX intervention has implications for long-term US Treasury market liquidity that should not be underestimated. Even when Japan does not directly sell long-term Treasuries, the need to use liquidity obtained from short-term Treasuries for FX intervention indirectly reduces funding demand for long-term Treasuries (see Chart 5). This year, 30-year US Treasury yields have surged, and "bond vigilantes" have made a comeback. This may be an important reason the US was compelled to join the coordinated yen depreciation intervention.

The "Plaza Accord" Narrative is Premature

After the Jamaica Accords in early 1976, floating exchange rates became legal and demonetized, the "double peg" Bretton Woods 1.0 system was completely dismantled, and the international monetary system entered the dollar credit-based Bretton Woods 2.0 era. However, Japan maintained a long-running dirty float on the yen until April 2004, when it announced a substantive end to routine FX intervention. Since then, Japan conducted phased interventions against yen appreciation in 2010 and 2011, and switched to intermittent interventions against yen depreciation from 2022 onwards (see Chart 6).

Japanese intervention is typically unilateral; joint intervention with other countries occurs only in rare circumstances. Whether unilateral or joint, Japan almost always informs the US in advance to obtain understanding and support. The New York Fed has even used rate checks multiple times to bolster Japan's intervention efforts. Unlike the verbal interventions of the past via rate checks, the US's high-profile market entry this time carries an extremely strong signaling effect, hence the largest short-term yen rebound in this round (see Chart 1). After official confirmation of joint intervention on August 3, the yen briefly rose to around 155 intraday, a maximum rebound of nearly 9 yen from the prior low.

But this is not a new Plaza Accord. First, while the Plaza Accord is considered to have kicked off the dollar depreciation cycle of the mid-1980s, this time around there is significant uncertainty. Even if it becomes a turning point from a strong to a weak dollar, it may not necessarily rescue the weak yen. Last year, with the dollar index plunging and the negative interest rate differential between Japan and the US narrowing, the yen rose only 0.3% against the dollar. Second, the Plaza Accord was a package agreement signed by five Western countries to correct dollar overvaluation, whereas this is merely exchange rate policy coordination on a case-by-case basis. Even if there are reports that Japan made a rate hike commitment to the US, that is just a gentlemen's agreement rather than a contract. Even in this joint intervention, the US operation of selling euros and buying yen.

This joint intervention is only a continuation of the Plaza Accord framework, not a new breakthrough, let alone a new normal. On one hand, Japan faces soft constraints. The International Monetary Fund (IMF) has set explicit quantitative criteria for "free float": within 6 months, no more than 3 FX interventions, each lasting no more than 3 trading days. Excess intervention would reclassify the country's exchange rate regime as ordinary "floating" rather than "free floating," weakening the yen's international credibility and Japan's voice in the G7 monetary system. Consequently, Japan's Ministry of Finance has long deliberately avoided exceeding intervention limits, even combining multiple days of intervention into a single count.