From March tenth, two thousand twenty-five to March eleventh, two thousand twenty-five

I. Foundation 

resistance to change and hence greater dependability for the results shown in the past.

Understanding the Mar-a-Lago Accord in one article

The market is entering a generational transformation of the global trade and financial system, and Trump’s trade and monetary policies may become a new Market Paradigm Shift. Essentially, the dollar’s long-term overvaluation stems from inelastic demand for dollar reserve assets, while the manufacturing sector’s difficulties are attributable to the Triffin dilemma. The United States continually exports Treasury debt to supply global reserves. Although this twin-deficit model gives the United States low borrowing costs (CIP close to zero deviation), it substantially weakens export competitiveness and exacerbates the trade deficit.

The market’s central misconception is equating tariffs with an inflation shock, but 2018-2019 data show that CPI & PCE inflation remained stable or even declined during the tariff war. And if tariff measures continue to be accompanied by exchange-rate adjustments, the country being taxed may still bear the ultimate tax burden, rather than U.S. consumers.

Dollar appreciation offsets inflationary pressure, followed by the next monetary-policy adjustment to correct the impact on manufacturing.

Against the backdrop of the so-called Mar-a-Lago Accord, the strategic path for Trump’s term is to lead with tariffs and use trade policy to reshape global competition rules, while using monetary (FX) tools to offset the structural flaws of the dollar’s long-term overvaluation.

Between 2018 and 2019, the Trump administration sharply raised tariffs, and the effective tariff rate on trade with China rose significantly, but the dollar appreciated by almost as much as the tariffs. As a result, dollar-denominated import prices rose only slightly, and overall inflation remained stable. This shows that if dollar appreciation fully offsets tariffs, the country being taxed will be the true bearer of the tax burden, while the United States will achieve Cost-Sharing in Reserve Asset Supply through increased fiscal revenue, Through exchange-rate adjustments, the tariff burden is ultimately passed on mainly through currency depreciation in exporting countries, putting very little pressure on consumer prices, while the U.S. Treasury achieves a cost-sharing effect by increasing tax revenue.

During the 2018-2019 U.S.-China trade war, the macroeconomic impact of tariffs was smaller than the market expected. The key reason was that exchange-rate adjustments directly absorbed most of the price shock. The depreciation of the renminbi by 13.7% offset the increase in the effective tariff rate of 17.9%, so the final increase in dollar-denominated prices of Chinese imports was only 4.1%.This mechanism strengthened the Trump administration’s understanding of how tariffs are transmitted through real purchasing power, which may lead it to pursue even more aggressive tariff policies.

The market’s biggest misjudgment of Trump’s policies is underestimating the government’s room for maneuver in exchange-rate policy.

The policy tools the Trump administration could use to drive dollar adjustments are severely underestimated by the market. The prevailing consensus is that dollar policy depends on Federal Reserve interest rates, but the Treasury and the Federal Reserve are not completely independent, and the Treasury can still impose reserve asset charges under IEEPA (International Emergency Economic Powers Act) to suppress inelastic demand for U.S. Treasuries, or directly use the Exchange Stabilization Fund (ESF) to intervene in foreign-exchange markets.

The market generally believes that “dollar policy cannot change,” but from the Plaza Accord (1985) to the Louvre Accord (1987), the United States has proactively adjusted the dollar’s valuation many times throughout its history. Against this backdrop, the Trump team not only values tariffs as a tool but also expects to adjust the dollar in stages through monetary policy.

Specifically, there are two possible paths:

First, a multilateral coordination plan, similar to the Plaza Accord of 1985 or the Louvre Accord of 1987. The Trump administration may try to promote a “Mar-a-Lago Accord” through coordinated adjustments by trading partners to lower the dollar’s valuation. The practical problem is that economic growth in Europe, Japan, and Country Z is weak, and they have no intention of proactively pushing their currencies higher.

If the Trump administration uses tariffs + the issue of the military umbrella to apply negotiating pressure, it may force some allies to accept a framework for dollar adjustment, but this would be far more difficult than in the 80s.

Second, unilateral action: use tools such as IEEPA and ESF to impose a “usage fee” on U.S. Treasury securities held by foreign official entities, forcing them to adjust their reserve behavior. At the same time, coordinate with the Federal Reserve using SOMA to smoothly manage short-term interest rates and control the long-term yield curve, preventing large-scale foreign capital outflows and a sharp rise in long-term yields caused by dollar depreciation.

Through this operation, the government can not only lower the dollar’s “excessively high” value in international trade but also shift part of the reserve costs to reserve buyers, thereby maintaining the dollar’s reserve-currency status while reducing its negative impact on U.S. exports.

At the same time, tariffs themselves, as a negotiating lever, serve not only to increase fiscal revenue, but also to force policy concessions from rival countries by raising the cost of imports. For example, during his campaign, Trump proposed raising tariffs on major countries to 60%, and on other countries to 10% or higher. Although this extreme plan carries a risk of triggering market volatility, costs can be passed on gradually under a framework of phased implementation and “forward guidance”.

The market generally assumes that tariff policy and dollar policy are independent, but the Trump administration’s core strategy is tariffs first, followed by monetary adjustment. Market pricing needs to account for the following path:

1. Short-term dollar strength (Tariff-Driven USD Strength)

2. Improved tariff revenue, increasing U.S. Treasury policy flexibility

3. If global capital flows are constrained, the Treasury launches FX intervention

4. The path of dollar adjustment depends on the degree of Federal Reserve cooperation

Overall, against the backdrop of the Mar-a-Lago Accord, Trump’s strategy will first use tariffs to rebalance trade terms, use the fiscal revenue generated by tariffs to improve cost-sharing for the U.S. supply of external reserves, then rely on exchange-rate policy to strengthen the dollar in the short term, and subsequently use FX Intervention and usage fees on external debt to adjust reserve behavior, ultimately achieving a comprehensive restructuring of the trade and financial systems through a series of gradual measures.

The pace of policy adjustments will determine the market’s main source of volatility over the next 6-12 months.

2025.03.10

Five 1990 down cycles since USDJPY

USDJPY has fallen 7% from its 7 peak; looking back at the five USDJPY downturns since 1990, the currency pair typically falls 18% to 47% from its peak within 16 to 60 months:

The current macroeconomic backdrop resembles several historical periods of regime shifts in yen trading, These periods are typically accompanied by Federal Reserve rate-cutting cycles, slowing U.S. economic growth, and foreign-exchange intervention The difference this time is that the U.S. economy remains resilient for now, but how long that will last remains to be seen.

Japan’s Ministry of Finance data show that, as of 1, Japanese investors were net sellers of overseas bonds over the past three months, mainly through banks and life insurers. Although the amount remains modest compared with 2022-2023, it could have an important impact on further yen appreciation in the future. This also means that related currency-pair trades such as EURJPY, CHFJPY, and GBPJPY still have room to move lower.

Public disclosures show that Japan’s 1.7 trillion-dollar pension fund has maintained a 25% target allocation to domestic bonds since 2020, when yields on 10-year Japanese government bonds (JGB) were near 0%. This target allocation is part of a five-year plan and is expected to be reviewed in 4 this year. Yields on 10-year JGB have now risen to 1.4%, a level last reached in 2011, when the pension fund’s domestic bond allocation exceeded 60%.

The current allocation to foreign bonds is 24%, compared with just 9% in 2011. Given that the yield spread between U.S. Treasuries and JGB remains elevated at around 3%, a high allocation to foreign bonds is still reasonable. But as the Bank of Japan (BoJ) raises rates and the Federal Reserve cuts them, the spread may narrow, This would prompt pension funds to shift from foreign bonds back to domestic bonds.

Since 7 last year, when USDJPY peaked, the Bank of Japan’s policy rate has risen by 40bp, while the yield on 10-year JGB has risen by 28bp. The subsequent the JGB yield curve will flatten further, as the Japanese government’s heavy debt burden (255% of GDP) is putting downward pressure on long-end yields.

According to Japan’s Ministry of Finance, the government’s current annual interest payments amount to 9.7 trillion yen, equivalent to 1.6% of GDP, a sharp increase from 1.2% in 2023. As JGB yields rise, pressure on the fiscal deficit may intensify further. The average maturity of Japanese government debt is currently 9 years and 7 months (the 2024 Ministry of Finance debt management report).

Until new incremental buyers are found (such as domestic pension funds, life insurers, and banks), the BoJ faces constraints in reducing its JGB purchases and will find it difficult to tighten policy significantly.

Black: Japanese government interest payments as a share of GDP; blue: JGB yield on 10-year bonds

Historically, yen regime shifts have often been accompanied by a flattening of the JGB yield curve, and this time will be no different.

2025.03.11

AI Spending vs. GDP: Why is the market euphoric while the macro data lag behind?

Over the past two years, annualized revenue at listed companies related to AI Infra has grown by more than $340B, and it could approach $580B by the end of 2025.But in U.S. GDP accounts, actual annualized investment in AI-related categories has increased by only $42B, a striking gap. On the surface, AI is a high-growth market theme, and its share price is rising, but the macro data show a negligible boost to actual GDP.

Breaking it down, much of listed companies’ nominal revenue growth reflects rising costs (especially semiconductors) and overseas revenue; neither boosts actual U.S. GDP. Margin expansion ($30B) and overseas revenue ($130B) account for about half of the surge in AI spending.

The BEA (Bureau of Economic Analysis) calculation method for GDP is also a key factor behind the underestimation. Semiconductor purchases are counted in GDP accounts classified as intermediate inputs (Intermediate Input), because they are typically embedded in final products and resold, rather than counted as capital investment. Over the past two years, the U.S. domestic semiconductor supply chain has grown by $35B, but this has barely appeared in GDP statistics. Similarly, spending on Cloud Services for AI training and inference also is mostly counted as intermediate inputs rather than fixed investment (Fixed Investment), meaning that even rapid growth among cloud computing companies will not show up clearly in GDP growth.

This explains why AI spending is booming at the market level while macro data lag behind: the key lies in statistical definitions and how capital spending is classified.With a more reasonable adjustment, the actual contribution of AI investment to GDP may have been underestimated by $100B.

In 25, as AI investment expands into Data Centers, Servers & Networking, and Utility Infrastructure, more of this spending is likely to count as actual investment, potentially providing some support to GDP growth.

But a more realistic issue is that unless the way the U.S. national accounts are calculated changes, semiconductor and Cloud-related investment will continue to be treated as intermediate inputs and thus “ignored” by the statistical system. In other words, Even if enthusiasm for the AI industry continues to build, the response in GDP data may still be “below expectations,” potentially skewing future policy discussions and market interpretations.

2025.03.11

II. Insights 

Our Viewpoints and ideas on market.

III. Market Pulse 

Our latest Comments on the economic and market environment

Note: All content in this article is for reference only, does not constitute investment advice, and is not a basis for any laws, regulations, or regulatory policies. Investors should not use or rely on this information as a basis for decision-making.