The dollar’s share of global official foreign exchange reserves fell from 64 percent in 2015 to 56 percent in 2025. This downward trajectory is sometimes read as evidence that the dollar’s role in international financial markets is eroding. However, aggregate statistics obscure the composition of changes occurring at the country level. In this post, we show that the aggregate decline is not a systematic global shift away from dollar assets. Rather, the aggregate decline reflects the actions of a handful of large reserve holders, changing either their currency preferences or the size of their reserve portfolio. From the perspective of the cross section of countries holding dollar assets, the dollar’s status in official portfolios is largely intact.
Understanding the Aggregate Dollar Shares of Reserves
When economists calculate the dollar share of worldwide official foreign exchange reserves, countries with larger reserve holdings naturally exert disproportionate influence on the final number. As Goldberg and Hannaoui (2026) show, this seemingly straightforward calculation can mask two fundamentally different phenomena. Countries can actively reallocate their existing portfolios away from dollar assets and toward other currencies, which we term the “preferences channel.” Alternatively, countries can accumulate or decumulate new foreign exchange reserves at dollar shares different from the global average, which we call the “reserve change channel.” When a country with below-average dollar holdings expands its reserves, it mechanically pulls down the global aggregate, even without reducing its own allocation to dollars. From this lens, we can interpret the evolution observed in the chart below, showing the currency composition of global foreign exchange reserves as reported by the International Monetary Fund (IMF).
Shares of Dollars and Big Four Currencies in Global Foreign Exchange Reserves Have Declined
Two distinct periods, selected for availability of data on individual country composition of foreign exchange reserves, illustrate what drives the aggregates. From 2015 to 2019, the dollar share fell by 3 percentage points. From 2019 to 2023, the decline moderated to 2 percentage points. The central question is whether these aggregate movements reflect a large set of countries systematically reallocating away from the dollar, or whether they stem from the actions of a few large reserve holders making choices specific to their own circumstances.
Consider the most basic indicator: are countries moving dollar shares of reserves in the same direction? If the aggregate decline reflected a broad-based global shift away from the dollar, most countries would reduce their dollar allocations over these periods. Instead, examining directional changes in dollar share across countries during both four-year windows, we find that roughly equal numbers of countries increased and decreased their dollar holdings. This balance suggests no dominant cross-sectional shift away from dollars. Instead, examining directional changes in dollar share across countries during both four-year windows, we find that roughly equal numbers of countries increased and decreased their dollar holdings, as illustrated in the chart below.
There Are Similar Counts of Countries with Increased and Decreased Dollar Shares Over 4-Year Windows
If Countries Are Not All Moving Away from the Dollar, What Explains the Aggregate Decline?
During the 2015-19 period, we examine the reserve portfolios of seventy-nine countries, of which seventy-six have complete data for the beginning and end of the period. For these seventy-six countries, the aggregate decline was split almost evenly between the preferences and reserve change channels, which accounted for 1.2 and 1.5 percentage points of the decline, respectively, as seen in the chart below. These figures represent the net sum of positive and negative country-level contributions within each channel. Examining those underlying contributions reveals that the aggregate picture masks a high degree of concentration.
Changes in Preferences for USD Assets and in the Size of Official Reserves Drive Decline in USD Share from 2015-19
The preferences channel captures active reallocation decisions, such that a large reserve holder shifting away from dollars influences the global aggregate. During the 2015-19 period, China and Russia dominated this channel, accounting for most of the 1.2 percentage point decline attributed to active portfolio reallocation away from the dollar. Turkey, Peru, and Spain also contributed downward pressure, though their magnitudes were much smaller, at around negative 0.2 percentage point each, while Sweden and Switzerland increased their dollar portfolio shares.
The reserve change channel shows a similar degree of concentration. Switzerland emerged as the largest negative mechanical contributor to the dollar share of reserves through this channel over this specific period. Because Switzerland’s dollar share was below the global average at the start of the period, its large-scale reserve accumulation over 2015-19 mechanically dragged down the cross-country aggregate dollar share, even as Switzerland simultaneously increased its own dollar allocation! The Czech Republic and Russia also contributed negatively through reserve accumulation at below-average dollar shares. Hong Kong and China, by contrast, contributed positively to this channel, with China’s positive contribution reflecting reserve drawdowns over the period.
These results illustrate that the same country can contribute through opposing channels simultaneously, as was the case for China and Switzerland. Without our decomposition framework, the aggregate decline would appear to be a uniform phenomenon, when in reality it reflects potentially offsetting forces and possibly a small number of large players.
What Do Available Data Show for Recent Years?
Our data coverage for the 2019 to 2023 period presents both challenges and insights. Complete dollar-share data are available for sixty-two countries, but several major reserve holders that reported in 2019 did not disclose their 2023 dollar allocations. Among these countries with missing 2023 data, China, Russia, Mexico, and Morocco are the largest reserve holders. The omission of these four countries is highly consequential as they collectively hold reserves roughly equal in size to the entire remaining sixty-two-country sample combined. To address this data limitation, we consolidate China, Russia, Mexico, and Morocco into a single group and characterize their collective contribution under assumptions consistent with the observed aggregate decline of 2.3 percentage points in the IMF’s Currency Composition of Official Foreign Exchange Reserves (COFER) data. This approach allows us to understand what role this group must have played to generate the observed aggregate movement, given the behavior of the sixty-two countries with complete data.
For the sixty-two countries for which we have complete data, the preferences channel is actually slightly positive at 0.3 percentage point, meaning these countries collectively increased their dollar allocations over the period. The reserve change channel contributes a modest negative of 0.5 percentage point. Taken together, the known country sample contributes very little to the aggregate decline in dollar share. The picture changes dramatically when we consider the four countries for which we do not have complete data. Under assumptions matching the observed COFER decline, their implied contribution to the preferences channel is large and negative at 2.0 percentage points. This suggests that the aggregate dollar share decline over 2019 to 2023 is almost entirely attributable to the preferences of these four countries.
The contributions of individual countries reinforce this story of concentration (see chart below). Among countries with complete data, Brazil and Hong Kong reduced their dollar allocations, but their contributions are small in magnitude relative to the group for which we have incomplete data. Switzerland contributes positively to both channels in this period, a reversal from the earlier period that reflects a contraction in reserves. On the reserve change channel, Denmark emerges as the largest negative contributor to the overall 0.5 percentage point decline, followed by Bulgaria, Romania, and Poland, all of which expanded reserves from a below-average dollar share. Croatia and the United Kingdom contributed positively through reserve contraction. The slower pace of dollar-share decline reflects in part the absence of large-scale reserve accumulation over the 2015 to 2019 period.
Changes in Preferences for US Dollars Concentrated Among a Few Countries Drive The Decline in USD Share from 2019-23
What Will the Future Hold for Dollar Shares in Official Reserves?
The decline in the dollar’s share of foreign exchange reserves after 2019 was not a broad-based phenomenon reflecting widespread international reallocation away from dollar assets. Among the vast majority of countries we observe, contributions to both channels cluster around zero. The aggregate decline is instead driven by a small number of large reserve holders, with China and Russia leading through reduced portfolio allocations to dollars. The reserve change channel reflects a rotating group of countries responding to idiosyncratic reserve management needs rather than systematic dollar avoidance. Official reserve holdings traditionally are driven by country needs for dollar liquidity, currency management needs, and insurance against both funding shocks. These drivers still retain their strength.
For policymakers and market participants, our analysis indicates that there is little evidence of a widespread official diversification away from dollars, despite the decline in the dollar share of aggregate official reserves. Aggregate statistics can create misleading impressions of broad trends when they actually reflect the concentrated actions of a few large players. Understanding this concentration is essential for accurately interpreting the dollar’s continued importance in international financial markets.
Linda S. Goldberg is a financial research advisor in the Federal Reserve Bank of New York’s Research and Statistics Group.
Oliver Hannaoui is a former research analyst in the Federal Reserve Bank of New York’s Research and Statistics Group.
Sneha Parthasarathy is a research analyst in the Federal Reserve Bank of New York’s Research and Statistics Group.
How to cite this post:
Linda S. Goldberg, Oliver Hannaoui, and Sneha Parthasarathy, “Are Central Banks Moving Out of Dollar Assets?,” Federal Reserve Bank of New York Liberty Street Economics, September 2, 2026, https://doi.org/10.59576/lse.20260902
BibTeX: View |
Disclaimer
The views expressed in this post are those of the author(s) and do not necessarily reflect the position of the Federal Reserve Bank of New York or the Federal Reserve System. Any errors or omissions are the responsibility of the author(s).
You can follow this conversation by subscribing to the comment feed for this post.