RMB Imbalance Does Not Equal 21% Undervaluation; Sharp Revaluation Is Not the Cure

Overview

This Deutsche Bank research report provides a calm response to the fierce international debate surrounding the RMB exchange rate. The core conclusion is that the IMF's latest estimate of a 21.3% real effective exchange rate (REER) gap for China has been widely misread as implying the RMB is undervalued by 21%. In reality, this figure answers only a narrow counterfactual question: how much would the exchange rate need to appreciate if exchange rate movements alone were used to eliminate the current account surplus. The report advocates separating the issues of trade imbalance, exchange rate valuation, and exchange rate adjustment. It acknowledges that China's surplus is indeed large but argues the evidence falls far short of proving the RMB is significantly mispriced. True adjustment should rely more on domestic reflation, domestic demand absorption, and gradual appreciation, rather than a one-off sharp revaluation.

Core views

On the IMF's -21.3% Gap: The report points out that in its latest External Sector Report released on July 31, 2026, the IMF significantly widened China's 2025 REER gap from -8.5% the previous year to -21.3%. However, DB emphasizes that this figure is calculated by taking China's 2025 current account surplus as a percentage of GDP (3.8%), adjusting for cyclical factors to derive a potential surplus of 3.6%, and subtracting the IMF model's 'normal' value of 0.6% to get a 3.0 percentage point current account gap. This gap is then divided by China's sensitivity of the current account to the REER (0.14) to arrive at -21.3%. It answers the question: 'How much would the exchange rate need to appreciate to reduce the surplus from 3.6% to 0.6% relying solely on exchange rate appreciation?', rather than providing an independent estimate of whether the RMB market price is below fair value. The judgment from direct valuation models is much more moderate: The same IMF report contains two direct valuation models—the REER-Level model (comparing relative price levels across countries) shows the RMB is undervalued by only 2.7%; the REER-Index model (using China's economic fundamentals to explain exchange rate movements) suggests the RMB is 0.5% higher than historical benchmarks, essentially at fair value. The DB FX Research team reached similar conclusions using three independent methods: the RMB is undervalued against a basket of trade-weighted currencies by only 3-8%. Therefore, DB concludes that 'the degree to which China's surplus is clearly excessive is far higher than the degree to which the RMB is deeply undervalued.' Historical Experience of Surplus Reversal: Domestic adjustment is at least as important as exchange rate appreciation. A 2010 IMF study identified 28 policy-driven current account surplus reversals between 1960 and 2008. In these cases, private savings fell by an average of 3.3% of GDP, investment rose by 3.0% of GDP, while exchange rate appreciation was modest—nominal and real effective exchange rates appreciated by an average of only about 10%. More critically, adjustment occurred primarily through imports rather than exports: imports rose by an average of 4.2 percentage points of GDP, while the export share of GDP remained virtually unchanged. That is, surplus countries adjusted mainly by stimulating domestic demand to drive imports, rather than compressing exports via exchange rate appreciation. China's surplus reversal from 2007-2013 also confirms this: the surplus fell from about 10% to below 2%, accompanied by a five-year nominal appreciation of 27% and real appreciation of 31%. However, this adjustment was feasible because China relied heavily on domestic investment stimulus at the time, and weakening global demand also helped. Understanding China's Current Surplus Lies in the Divergence Between Nominal and Real Effective Exchange Rates: Between 2021 and 2025, BIS data shows China's broad nominal effective exchange rate (NEER) rose by approximately 1.2%, with another 5.5% rise in the first half of 2026, reaching historic highs. Meanwhile, the real effective exchange rate fell by 15.1%, rebounding only 3% in the first half of 2026. The reason is that China's domestic prices fell by 16.3% relative to trading partners—Chinese consumer prices remained almost flat, producer prices even fell, while the world experienced post-pandemic inflation spikes. This real depreciation, against the backdrop of nominal appreciation, enhanced the competitiveness of Chinese manufacturing, thereby expanding the trade surplus. Structural Factors Behind Real Depreciation: First, Chinese manufacturers have improved price competitiveness through technological upgrades, economies of scale, and increasingly complete supply chains. Taking electric vehicles (EVs) as an example, the MSRP of the base Model Y produced by Tesla's Shanghai factory in the Chinese market fell by 40% between 2020 and 2026, while it fell only 20% in the US market. China is the world's largest EV market (accounting for 60% share) and has the most EV manufacturers. The learning curve and economies of scale in manufacturing are the main drivers of real depreciation, applicable to many other products as well. Second, China's economic downturn has exacerbated price weakness: the property downturn reduces investment, dampens household confidence, and pushes up precautionary savings; local government balance sheet repairs suppress fiscal demand; producer competition compresses profit margins and prices; and household consumption has failed to fill the gap left by real estate and infrastructure. These forces have yet to show convincing turning points. Third, corporate foreign exchange hoarding has hindered the normal transmission of appreciation: From 2022 to the first half of 2025, Chinese private enterprises cumulatively held nearly $800 billion more in foreign exchange assets. This behavior weakened the normal link between rising trade surpluses and RMB appreciation. The RMB began to strengthen from the second half of 2025, coinciding precisely with the end of this hoarding behavior. Do Not Blame Surplus on FX Intervention: The report explicitly states that the People's Bank of China's core philosophy is to smooth exchange rate volatility, resulting in interventions during this period being more about 'fighting depreciation' than 'fighting appreciation'. For example, between 2022 and 2025, the daily USD/CNY central parity rate often stood on the stronger side of the spot rate; without PBOC intervention, the RMB would have been much weaker. The Correct Path for Adjustment: DB believes that China's rebalancing requires a stronger real effective exchange rate, which in turn requires stronger domestic demand and higher inflation. Stabilizing the real estate sector can reduce precautionary savings and revitalize household confidence; fiscal policies targeting household income, healthcare, pensions, and public services will support consumption more directly than another round of manufacturing investment; 'Six Networks' infrastructure investment will also boost domestic demand, having already driven up import growth rates this year. Crucially, a demand-driven recovery will not only allow the real effective exchange rate to appreciate but also push up domestic interest rates, supporting the RMB's nominal exchange rate. This does not mean RMB appreciation is ineffective, but that sequence matters: appreciation can promote rebalancing only when accompanied by stronger domestic demand and higher inflation, not as a substitute for them. Without domestic adjustment, nominal appreciation implemented to bridge the relative price gap would instead deepen the domestic forces causing the surplus—weak demand, deflation, and low profit margins—worsening domestic absorption capacity and pushing Chinese manufacturers further into export markets, ultimately offsetting the initial benefits of appreciation. Direct Responses to Two Popular Views: First, regarding Brad Setser's proposed undervaluation magnitude of 30-35%, DB points out that this methodology reconstructs customs data, excludes gold imports, and normalizes the investment income deficit, pushing the surplus up to 6-7% of GDP; this approach is inherently biased upwards. More fundamentally, even if the surplus were larger, it does not imply greater exchange rate mispricing—converting a current account gap into 'undervaluation' inherently assumes the exchange rate is responsible for adjustment, whereas historical evidence suggests the opposite. Second, regarding Michael Pettis's more nuanced view that 'appreciation achieves rebalancing by increasing household purchasing power,' DB believes this channel exists but is secondary: Household savings are determined mainly by the economic cycle, confidence, and income expectations. Chinese households are unlikely to significantly reduce savings simply because imported goods become cheaper; they may buy more imports but simultaneously reduce domestic purchases, potentially exacerbating deflationary pressures. In other words, externally driven nominal appreciation is unlikely to generate sufficient real appreciation to resolve China's excess savings problem. Adjustment at the Global Level: The report also emphasizes that if the rise in China's export share partly reflects genuine productivity advantages, global rebalancing requires not just China increasing imports but also other economies catching up in productivity and competitiveness. New technologies such as AI, robotics, renewable energy, and new materials are reshaping manufacturing, with Chinese manufacturers benefiting from early adoption and gaining competitive advantages; other economies must improve their productivity faster than Chinese competitors. Only by narrowing the productivity gap can global imbalances be finally resolved.

Analysis framework

DB's analysis follows a clear主线: first disentangling the concepts of 'imbalance, valuation, and adjustment' which are often conflated, and then testing each with data and historical evidence. Step 1 is deconstructing the numbers. The report does not directly deny the IMF's -21.3% but还原s it to its calculation formula—surplus gap divided by exchange rate sensitivity—allowing readers to see that the underlying assumption is 'exchange rate is the only adjustment tool', thereby revealing it is not an independent valuation. Step 2 is cross-referencing evidence from the same source: direct valuation models within the same IMF report, as well as DB's own three valuation methods, all point to the RMB being close to fair value, shaking the narrative of '21% undervaluation'. Step 3 introduces historical experience: using cross-country samples of 28 surplus reversals to prove that adjustment relies mainly on domestic absorption (declining savings, rising investment, growing imports), with the exchange rate playing a minor role. Step 4 explains the current mechanism: using the divergence between nominal and real effective exchange rates to show that the true cause of China's widening surplus is the decline in domestic relative prices, further breaking down the structural causes of price declines (learning curves, property downturn, FX hoarding). Finally, policy deduction: based on these mechanisms, deriving that 'gradual appreciation + domestic reflation' is the reasonable path, and responding one by one to two popular counter-arguments. This layer-by-layer progression of 'number reconstruction + cross-validation + historical analogy + mechanism breakdown' is a typical macro exchange rate analysis method.

Methodology notes

• Macroeconomic framework

  Current Account Identity: Current Account = Savings - Investment; a current account surplus essentially reflects domestic savings exceeding investment

  DB uses this identity to show that a surplus can narrow due to declining savings, rising investment, real exchange rate appreciation, or changes in external demand; exchange rate appreciation is just one channel, not the whole picture. This explains why the assumption of 'adjustment relying solely on exchange rates' is too narrow.

• Valuation methods

  Three Types of Exchange Rate Valuation Methods: Current Account Method (CA Gap Method), Relative Price Level Method (REER-Level), Fundamental Regression Method (REER-Index)

  DB compared three internal IMF calibers: the current account method, assuming single-adjustment via exchange rates, yields a '-21.3% large gap', while the relative price and fundamental methods point to only 2.7% undervaluation and near fair value respectively. Cross-validating different valuation calibers is key to judging whether an exchange rate is mispriced; conclusions from a single method are easily exaggerated.

• Valuation methods

  Divergence between Nominal Effective Exchange Rate (NEER) and Real Effective Exchange Rate (REER): REER = NEER adjusted for relative inflation

  DB emphasizes that nominal appreciation and improved real competitiveness can coexist—the RMB hit nominal records high, but China's domestic prices fell 16.3% relative to trading partners, causing the real effective exchange rate to fall instead of rise. Looking only at nominal exchange rates leads to the erroneous conclusion that 'the RMB is already very strong'; real competitiveness is the key driver of the widening surplus.

• Event Gaming and Behavioral Finance

  Exchange Rate Transmission Blockade by FX Hoarding Behavior: Enterprises holding foreign exchange assets instead of settling them weakens the transmission from trade surpluses to local currency appreciation

  DB points out that Chinese enterprises held approximately $800 billion more in foreign exchange assets from 2022 to the first half of 2025, severing the normal link between surpluses and appreciation; the end of hoarding in the second half of 2025 coincided with the RMB beginning to strengthen. This explains why the RMB did not appreciate synchronously with rising surpluses and hints at the source of subsequent appreciation pressure.

• Industry/Industrial Analysis FrameworkCost curve analysis

  Learning Curves and Economies of Scale Lead to Declines in Real Manufacturing Prices

  DB uses the example of the Tesla Shanghai factory Model Y dropping 40% in price (compared to only 20% in the US) to illustrate that technological upgrades, scale effects, and supply chain density are the main drivers of real depreciation in Chinese manufacturing. This is the supply-side logic understanding why China's surplus continues to expand—not suppressed by exchange rates, but driven by genuine productivity improvements.

• Cycle and Prosperity Framework

  Historical Sample Analysis of Surplus Reversal: In policy-driven current account surplus reversals, the contribution of domestic savings-investment adjustment exceeds that of exchange rates

  DB cites IMF 2010 research on 28 surplus reversals, proving that private savings fell by an average of 3.3 percentage points of GDP, investment rose by 3.0 percentage points, while exchange rates appreciated modestly by about 10%, and adjustment occurred mainly through imports rather than exports. This provides empirical basis for 'China should rely on domestic demand rather than significant appreciation'.

Key data

• IMF Latest REER Gap (2025)-21.3%Significantly widened from -8.5% the previous year, but is merely a counterfactual scenario of 'closing the surplus relying solely on exchange rates'
• IMF Current Account Method: China's Potential Surplus3.6% GDPActual surplus 3.8% GDP in 2025, adjusted to 3.6% after removing cyclical factors, corresponding to a normal value of 0.6%
• IMF Direct Valuation Model ResultsUndervalued 2.7% (REER-Level) / Overvalued 0.5% (REER-Index)Forms a stark contrast with the -21.3% gap, pointing to the exchange rate being close to fair value
• DB FX Valuation ConclusionRMB undervalued by 3-8% against trade-weighted basketDerived using three independent methods, far smaller than the 15-35% undervaluation magnitude discussed in the market
• Change in China's Broad NEER (2021-2025)+1.2%Rose another 5.5% in H1 2026, reaching historic highs
• Change in China's REER (2021-2025)-15.1%Rebounded only 3% in H1 2026; real competitiveness significantly improved
• Change in China's Relative Prices-16.3%Magnitude of decline in domestic prices relative to trading partners, the main cause of actual REER depreciation
• Scale of Corporate FX Hoarding in China (2022-2025H1)Approx. $800 BillionWeakened the transmission of surpluses to RMB appreciation; hoarding ended in H2 2025
• Historical Surplus Reversal Samples (28 Cases)Savings down 3.3% GDP, Investment up 3.0% GDP, Exchange Rate up approx. 10%Imports rose by an average of 4.2 percentage points of GDP, export share remained almost unchanged
• China's 2007-13 Surplus ReversalSurplus fell from approx. 10% GDP to below 2%Accompanied by nominal appreciation of 27% and real appreciation of 31%, but relied mainly on domestic investment stimulus
• Tesla Shanghai Model Y MSRP Drop (2020-2026)-40%US market dropped only 20% during the same period, reflecting learning curves and economies of scale in Chinese manufacturing

Impact & implications

Implications for RMB Trend: DB expects the RMB adjustment over the coming years to be steady and gradual. Trade and current account surpluses have been accumulating appreciation pressure—exporters have started settling more foreign exchange since the end of 2025; however, China's low inflation has created persistent interest rate differentials between the RMB and other currencies, imposing negative carry costs on holding the RMB and slowing the pace of appreciation. This combination points to gradual appreciation rather than a one-off jump. Policymakers will not accept significant appreciation, as that would completely contradict the goal of reflation. Implications for Policy Direction: The report argues that the debate focus should shift from 'whether the RMB is undervalued and by how much' to 'how China can advance a multi-year adjustment dominated by domestic demand absorption'. Stabilizing real estate, fiscal transfers to households, and 'Six Networks' infrastructure investment are more effective rebalancing tools; externally driven nominal appreciation alone cannot generate sufficient real appreciation to solve the excess savings problem and may instead exacerbate deflation. Implications for Global Imbalances: Market share expansion brought about by China's productivity improvement is an economic inevitability and cannot be prevented by exchange rate appreciation or domestic adjustment. RMB strengthening should be a result of China's genuine productivity improvement, not a force running ahead of economic fundamentals to reverse the rise in China's export share. Global rebalancing requires other economies to accelerate productivity catch-up in fields such as AI, robotics, new energy, and new materials, narrowing the productivity gap with Chinese manufacturers.

Risks

• If external pressure forces a one-off sharp RMB appreciation without matching domestic demand, it may deepen weak demand, deflation, and low profit margins, pushing manufacturers further into export markets and undermining rebalancing.
• If Chinese enterprises' FX hoarding behavior reverses again due to policy or market changes, it may alter the pace of gradual appreciation.
• If domestic real estate and demand-side adjustments fail to show turning points for a long time, real effective exchange rate appreciation may remain slow, prolonging the process of surplus contraction.

What to watch

• Recovery of China's domestic demand and inflation, particularly turning points in real estate stabilization and household consumption confidence.
• Whether Chinese enterprises' foreign exchange hoarding behavior continues to fade and settlement trends persist.
• Relative trends of the RMB's nominal and real effective exchange rates, and whether the gap in relative prices narrows.
• Specific implementation intensity of fiscal policies in household income, healthcare, pensions, public services, and 'Six Networks' infrastructure.
• Progress of other economies in catching up with China in productivity in fields such as AI, robotics, new energy, and new materials.