Reserve Demand Under the Microscope: Eight Basic Facts and Two Key Questions
Summary
The quantity of reserves associated with “ample” reserve conditions has increased substantially over time.
At the aggregate level, ample reserve levels appear to exhibit a fairly stable relationship with total bank assets. However, among domestic banks, most banks have “ratcheted up” their reserve holdings relative to total assets.
Reserve management practices vary widely across individual institutions.
Reserve management practices of branches and agencies of foreign banks appear to be driven by global liquidity management considerations of the parent organization.
The wide range of reserve management practices across banks and the changes in reserve management behavior of individual banks over time pose challenges for proposals aimed at reducing aggregate reserve demand.
Introduction
After about three and a half years of gradual balance sheet reduction, the Federal Open Market Committee (FOMC) last December judged that reserves had reached an ample level. Currently, aggregate reserve balances stand at about $3 trillion and, based on past experience, the supply of reserves might need to grow at roughly the pace of nominal GDP going forward. The sizable quantity of reserves necessary to maintain ample reserves conditions has garnered a great deal of attention over time from various quarters. Concerns about the size of the Fed’s balance sheet have been exacerbated by perceptions that reserve demand has been “ratcheting” higher over time (Acharya and Rajan, 2024).
Many observers have suggested that the Federal Reserve should downsize its balance sheet to reduce distortions in financial markets, lower the risk of future periods of negative net income for the Fed, diminish the interactions between the Federal Reserve’s balance sheet and issues of federal finance, and put the Fed in a better position to address future emergencies with quantitative easing. Indeed, prior to assuming his new role, Chairman Warsh often voiced concerns about the size of the Fed’s balance sheet. It comes as no surprise then that one of the five high profile task forces chartered by Chairman Warsh is devoted entirely to matters related to the Fed’s balance sheet. Recent analyses by Federal Reserve economists and policymakers may provide some indication of the range of issues this task force could address. Included among the proposals discussed in these papers are various mechanisms that might work to lower bank demand for reserves.[1] Against this backdrop, this note takes the perspective of a fact-finding mission with the goal of summarizing important details of bank reserve management behavior over time. To that end, section 2 below describes a simple accounting framework that is useful in thinking about the connections between individual and aggregate reserve demand. Section 3 describes three basic data sets employed on our fact-finding mission. Sections 4, 5, and 6 mine these data sets in detail and identify “8 basic facts” that may be helpful in thinking about policy proposals aimed at reducing reserve demand. Section 7 concludes with some general observations and two key questions that policymakers will need to wrestle with in efforts to significantly reduce aggregate reserve demand.
Accounting for Reserve Demand Growth
In the discussion below, it’s helpful to posit a simple accounting framework linking individual and aggregate reserve demand. We assume the individual bank demand for reserves takes the form:
Ri(t)
Ai(t)
=
Bi(t) ⋅ F(s(t))
Where
Ri(t) = reserves of entity i in period t
Ai(t) = assets of entity i in period t
Bi(t) = reserve intensity factor for entity i in period t
s(t) = spread of the Interest On Reserve Balances (IORB) rate over the federal funds rate
F(•) = a function describing the response of reserve demand to the spread
With this assumption, the percent increase in the individual bank reserve-asset ratios between dates with the same interest rate spread is equal to the percent increase in the reserve intensity factors, % ( Ri Ai ) = %Bi . The behavior of the reserve intensity factor thus should be independent of the funds rate spread and should reflect factors driving reserve demand over and above trend growth in bank assets. For example, positive changes in the reserve intensity factor over time could be a sign of so-called ratchet effects on reserve demand. Changes in the reserve intensity factors could stem from many other influences including changes in bank regulation and supervision as well as changes in bank preferences for the size and composition of liquidity.
The dollar change in reserves for individual institutions over the reference period, ΔRi , is a function of the percent changes in assets over the period, the percent changes in the reserve intensity factor, and an interaction term:
ΔRi = Ri(t) ⋅ (%Ai + %Bi + %Ai ⋅ %Bi)
With the assumed form of the individual bank reserve demand curves, the percent change in the aggregate reserve-asset ratio is equal to the percent change in the aggregate weighted average reserve intensity, θ(t) = ∑i wi(t) ⋅ Bi(t) with the weights representing each bank’s reserve holdings as a share of total reserves, wi(t) = Ri(t) / R(t) . The dollar change in aggregate reserves over the reference period is analogous to the expression for individual banks, except that the term for percent changes in the individual reserve intensity factor is replaced by percent changes in the weighted average reserve intensity factor:
ΔR = R(t) ⋅ (%A + %θ + %A ⋅ %θ)
Much of the discussion of domestically chartered banks below makes use of these accounting identities. The most important point is that in this framework, changes in individual reserve-asset ratios are important in understanding changes in “reserve demand” over time.
Data
The analysis below draws on three basic data sources—the H.8 statistical release published by the Federal Reserve, Call Reports filed by insured depository institutions (FFIEC 031/041), and Call Reports filed by U.S. branches and agencies of foreign banking organizations (FFIEC 002).
H.8 Statistical Release: The H.8 statistical release is published weekly by the Federal Reserve based on a sample of banks. Each report provides an aggregate snapshot of key balance sheet items for large and small domestic banks and for branches of foreign banks collected on the FR2644. Large banks are defined as the top 25 banks ranked by asset size as of the most recent benchmark date. Small banks are those outside of the top 25.[2]
Sample Based on Call Reports for Insured Depositories (FFIEC 031/041): For insured depository institutions, we also employ data drawn from individual bank call reports (FFIEC 031 and 041). For the purposes of this note, we focus on a sample of 570 domestically chartered banks. The sample includes all domestically chartered banks that maintained reserve balances at the Federal Reserve and for which call report data is available on both reference dates.[3] These dates were selected as points with ample reserve conditions prevailing and with a funds rate spread on these two dates at 1 basis point. Under our assumptions about individual bank reserve demand curves discussed above, percent changes in the reserve-asset ratios between these two dates should correspond to percent increases in the reserve intensity factor. Banks are categorized in four groups by size based on their average asset levels reported in 2018Q3 and 2025Q4. Group 1 includes banks with average assets less than $1 billion; category 2 includes banks with average assets greater than or equal to $1 billion and less than $100 billion; category 3 includes banks with average assets greater than or equal to $100 billion and less than $250 billion; category 4 includes banks with average assets greater than or equal to $250 billion.[4]
Sample Based on Call Reports for Branches and Agencies of Foreign Banks: For branches of foreign banks operating in the United States, we again rely on Call Report information collected in the FFIEC 002. This includes basic balance sheet variables for foreign branches. Information for this source was collected for all U.S. branches with reserve balances exceeding $1 billion. The sample accounts for essentially all reserve balances held by branches and agencies of foreign banks.
Evolution of Reserve-Asset Ratios Over Time
The most striking feature of Figure 1 is the series for branches and agencies of foreign banks. Both the level and the swings in the aggregate cash-asset ratio for these entities stand well apart from those of small and large domestic banks. We return to the behavior of foreign branches in detail below.
Reserve-Asset Ratios at Domestically-Chartered Banks
The H.8 data reviewed above pointed to some interesting differences in the pattern of reserve-asset ratios across different classes of depository institutions. Here we employ Call Report data to delve more deeply into the patterns of reserve management behavior at domestically-chartered depository institutions. The analysis is based on the sample of banks described above in section 3. The banks in the sample accounted for nearly all reserves and assets held by domestically chartered banks on the two reference dates, 2018Q3 and 2025Q4. Not surprisingly, as shown in Table 1, reserves and assets are heavily concentrated in the relatively small number of banks in size categories 3 and 4.
Table 1
| Size Group | Reserves 2018:Q3 |
Assets 2018:Q3 |
Reserves 2025:Q4 |
Assets 2025:Q4 |
|---|---|---|---|---|
| 1 | $2.1 | $57.9 | $5.6 | $91.0 |
| 2 | $110.2 | $2,718.1 | $283.4 | $4,771.7 |
| 3 | $110.1 | $1,518.4 | $241.2 | $2,589.0 |
| 4 | $692.0 | $9,166.0 | $921.0 | $13,684.8 |
| All | $914.4 | $13,460.3 | $1,451.2 | $21,136.4 |
Note: Data reported in $ Billions
Fact 2: Among Domestically-Chartered Banks, Reserve-Asset Ratios Generally Increase with Asset Size
Table 2 shows the evolution of the aggregate reserve-asset ratio computed on the two reference dates and similar ratios for four bank subgroups. As noted in the first two columns of the table, reserve-asset ratios generally are higher for the larger banks (those in categories 3 and 4). Higher reserve-asset ratios for these banks could reflect a number of factors including the effects of liquidity regulation and supervision, liquidity-intensive business models, and cautious attitudes toward liquidity risk management.
Fact 3: Reserve-Asset Ratios Generally Have Increased Over Time
Column 3 of Table 2 reports the percent change in the reserve-asset ratio between the two reference dates. Recall that under our assumptions, this percent increase should correspond to the increase in the weighted average reserve intensity factor for each group. Referring to the last row of the table, in aggregate, consistent with findings in the research literature, the aggregate reserve-asset ratio for all banks in the sample was remarkably steady at about 0.068 between the reference dates.[5] However, the picture varies substantially for banks in different size groups. The aggregate reserve-asset ratios for banks in size categories 1 to 3 increased notably while the reserve-asset ratio for the largest banks declined somewhat.[6] The stability in the aggregate reserve-asset ratio in the sample might be taken as evidence suggesting that reserve demand has not been ratcheting up over time. However, as shown in column 5 of the table, although the aggregate reserve-asset ratio was little changed over the reference period, more than 70 percent of the banks in the sample recorded an increase in their reserve-asset ratio. Moreover, this pattern is present among banks in all size classes.
Table 2
| Size Group | Reserves/Assets 2018:Q3 (1) |
Reserves/Assets 2025:Q4 (2) |
% Change Reserves/Asset (3) |
Count (4) |
MEMO: Count of Banks % Change Reserves/Asset > 0 (5) |
|---|---|---|---|---|---|
| 1 | 0.0359 | 0.0620 | 72.8% | 120 | 80 |
| 2 | 0.0405 | 0.0594 | 46.5% | 426 | 305 |
| 3 | 0.0725 | 0.0932 | 28.4% | 13 | 11 |
| 4 | 0.0755 | 0.0673 | -10.9% | 11 | 8 |
| All | 0.0679 | 0.0687 | 1.1% | 570 | 404 |
Fact 4: The Evolution of Reserve-Asset Ratios at the Largest Banks Varies Widely Across Banks
Table 3 reports the reserve-asset ratios for the largest banks on our two reference dates and the percentage change in the ratio over the reference period. There is a remarkable degree of variation across these banks, all of which are subject to the same regulatory requirements. The reserve-asset ratios for many of the largest banks increased by 100 percent or more over the period. Among this group, it’s interesting to note that the increases for the two major clearing and custody banks—Bank of New York and State Street—resulted in discretely higher reserve-asset ratios than for the other banks at the end date for the reference period. In stark contrast, three large banks—JP Morgan Chase, Goldman Sachs Bank, and Wells Fargo—registered sharp declines in their reserve-asset ratios over the period. Notably, the steep decline in the reserve-asset ratio for JP Morgan was accompanied by a sharp increase in the asset share of its holdings of Treasury securities. The behavior of these three banks played a very important role in the evolution of the aggregate reserve-asset ratio for the full sample over the period. Had these banks simply maintained a constant reserve-asset ratio over the reference period, the aggregate reserve-to-asset ratio (and the quantity of reserves deemed consistent with ample reserves) would have been notably higher.
Table 3
| Bank Name | Reserves/Assets 2018:Q3 |
Reserves/Assets 2025:Q4 |
% Change |
|---|---|---|---|
| BANK OF AMERICA | 0.0470 | 0.0513 | 9.11% |
| CAPITAL ONE | 0.0182 | 0.0798 | 337.58% |
| CITIBANK, NATIONAL ASSOCIATION | 0.0453 | 0.1068 | 135.90% |
| GOLDMAN SACHS BANK USA | 0.2974 | 0.1335 | -55.12% |
| JPMORGAN CHASE BANK | 0.1249 | 0.0250 | -80.02% |
| PNC BANK | 0.0527 | 0.0572 | 8.45% |
| STATE STREET BANK AND TRUST COMPANY | 0.0828 | 0.2188 | 164.06% |
| TD BANK | 0.0132 | 0.0697 | 428.21% |
| THE BANK OF NEW YORK MELLON | 0.1011 | 0.2100 | 107.72% |
| U.S. BANK | 0.0151 | 0.0508 | 235.55% |
| WELLS FARGO BANK | 0.0802 | 0.0588 | -26.63% |
Fact 5: There is Substantial Within-Group Heterogeneity of Reserve-Asset Ratios
The substantial variation in reserve management practices among the largest banks is also present in the data for banks in other size groups. Figure 2 below provides a visual representation of this heterogeneity focusing on the reserve-asset ratios for individual banks within each size group. The distributions of the individual bank reserve-asset ratios for each size group tend to have a “fat tail” toward higher levels of the reserve-asset ratio. As a result, the mean value of the reserves-asset ratio for each group (marked by the X in each bar) is above the median value for each group denoted by the horizontal line in each bar.
Fact 6: Reserve-Asset Ratios Exhibit a Degree of Persistence Over Time
The notable changes in the reserve-asset ratios noted for some banks above raises the broader question of the extent to which bank approaches to reserve management change over time. Some insight on that topic can be gleaned by examining the transition matrix shown in Table 4 below. Each row of the matrix shows the proportion of banks in each quintile for the reserve-asset ratios as of 2018:Q3 that migrates to the reserve-asset quintiles as of 2025:Q4. For example, row 1 indicates that among the banks in the lowest quintile of reserve-asset ratios in 2018:Q3, about 39 percent remained in the lowest quintile by reserve-asset ratio in 2025:Q4 (the upper left cell). At the other end of the scale, as shown in the last row of the table, a little more than 60 percent of banks in the highest quintile by reserve-asset ratio in 2018:Q3 remained in the top quintile in 2025:Q4 (the bottom right cell). More generally, the fractions along the diagonal of the matrix, the blue shaded cells, tend to be noticeably higher than the off-diagonal elements suggesting that there is a tendency for banks to remain in the same reserve-asset quintile over time. That said, there is a fair bit of migration over the period from one quintile to an adjacent quintile, shown by the grey shaded cells. The change in the reserve-asset ratio for JP Morgan noted above was associated with a transition from the top quintile in 2018:Q3 to the second lowest quintile in 2025:Q4—the orange shaded cell—a very unusual transition.
Table 4
| Quintiles 2018:Q3 |
Quintiles 2025:Q4 | ||||
|---|---|---|---|---|---|
| 1 | 2 | 3 | 4 | 5 | |
| 1 | 0.39 | 0.28 | 0.17 | 0.11 | 0.05 |
| 2 | 0.33 | 0.30 | 0.15 | 0.13 | 0.09 |
| 3 | 0.14 | 0.25 | 0.30 | 0.22 | 0.10 |
| 4 | 0.11 | 0.14 | 0.28 | 0.31 | 0.16 |
| 5 | 0.03 | 0.04 | 0.11 | 0.23 | 0.61 |
Reserve-Asset Ratios for Branches and Agencies of Foreign Banks
The evidence above suggests that reserve demand varies widely across domestic banks with different business models. The branches and agencies of foreign banks represent the most extreme case of this variation. As noted in the discussion of Figure 1, both the level and the swings in the reserve-asset ratio for foreign branches are much larger than those for small and large domestic banks. For these institutions, reserve demand may be best understood as a function of the global liquidity strategy set by the parent organization.
Fact 7: Foreign Branches Appear to Operate as Dollar Warehouses for their Parent Organization
One important clue in understanding the behavior of the reserve-asset ratios for foreign branches is shown in Figure 3 below. The chart plots the ratio of the net due to related foreign offices to total U.S. assets for foreign branches. The “due to” parent offices effectively represents an intracompany deposit placed by the parent bank at its U.S. branch. Beginning in about 2011, foreign branches in aggregate transitioned from acting as a source of dollar funding for their parent offices to a structure with a large positive “due to” position relative to their parent offices. The chart suggests that the contour of the “due to” position is strongly reflected in the time-series behavior of the reserve-asset ratio for foreign branches. That juxtaposition in turn suggests that foreign banks may have increasingly come to view their U.S. branches as “warehouses” for dollar liquidity for the global banking organization. This may have been particularly relevant for European banks during the intensification of the Euro area debt crisis and the extreme liquidity pressures that they faced during that period. Under this interpretation, dollar funding raised in global markets would be parked at the U.S. branch that in turn would hold much of the associated “due to” the parent office in the form of balances held at the Federal Reserve. In practice, a sizable share of global dollar funding is likely issued through affiliates in international tax havens headquartered in the Caribbean with the proceeds then placed in the branch office in New York.
Note: Data shown are taken from the H.8 statistical release; the reserve/asset ratio is proxied by the cash/asset ratio.
The trend toward foreign branches acting as dollar warehouses may have been further amplified by the extreme liquidity pressures experienced at the outset of the COVID crisis and an increased focus in regulation and supervision on maintaining a currency composition of high quality liquid assets that aligns with potential funding needs across different currencies. For example, paragraph 30.27 of Basel Committee on Bank Supervision LCR 30 document notes that “While the LCR must be met and reported in a single currency, banks should be able to meet their liquidity needs in each currency and maintain HQLA consistent with the distribution of their liquidity needs by currency. The bank should be able to use the stock to generate liquidity in the currency and jurisdiction in which the net cash outflows arise. As such, the LCR by currency should be monitored and reported to allow the bank and its supervisor to track any potential currency mismatch issues that could arise, as outlined in SRP50. In managing foreign exchange liquidity risk, the bank should take into account the risk that its ability to swap currencies and access the relevant foreign exchange markets may erode rapidly under stressed conditions. It should be aware that sudden, adverse exchange rate movements could sharply widen existing mismatched positions and alter the effectiveness of any foreign exchange hedges in place.”
Fact 8: There is a Wide Degree of Variation in Reserve Management Practices Across Branches and Agencies of Foreign Banks.
The aggregate reserve-asset ratio for branches and agencies of foreign banks is about 30 percent, much higher than the corresponding figures for domestically chartered banks. Similar to the case for domestic banks, there is substantial variation in reserve-asset ratios across individual branches and agencies. Indeed, many large foreign branches operate with reserve-asset ratios much higher than 30 percent. As shown in Table 5 below, branches of foreign banks display a wide range of behavior in their reserve management practices. Reserve-asset ratios (reserve balances at the Fed divided by assets of the U.S. branch) as of 2025:Q4 (column 3) were well above 50 percent for many of these entities and a number of branches had reserve-asset ratios of 70 percent or more. As shown in column 5, most of the branches operate with a large “due to” position relative to their head office. In many cases, the ratio of the “due to” position to total U.S. assets is also very high (column 5).
Table 5
| Nation | Name | Reserves (1) |
Total Assets (2) |
Reserves / Assets (3) |
Due To (4) |
Due To / Assets (5) |
|---|---|---|---|---|---|---|
| UK | Barclays | $93.6 | $154.1 | 0.61 | $86.1 | 0.56 |
| Japan | MUFG | $67.3 | $244.8 | 0.28 | $69.6 | 0.28 |
| Japan | Mizuho | $53.6 | $173.7 | 0.31 | $38.8 | 0.22 |
| UAE | First Abu Dhabi | $47.4 | $48.5 | 0.98 | $47.8 | 0.99 |
| Germany | Deutsche Bank | $46.9 | $178.5 | 0.26 | $0.0 | 0.00 |
| Japan | Sumitomo | $44.3 | $226.7 | 0.20 | $87.0 | 0.38 |
| France | Soc Gen | $38.6 | $129.5 | 0.30 | $67.9 | 0.52 |
| China | Bank of China | $37.6 | $61.9 | 0.61 | $36.4 | 0.59 |
| Switzerland | UBS Branch (Stamford) | $35.0 | $70.7 | 0.49 | $56.5 | 0.80 |
| UK | Standard Chartered | $29.3 | $63.6 | 0.46 | $10.7 | 0.17 |
| Canada | Toronto Dominion | $28.9 | $97.4 | 0.30 | $0.0 | 0.00 |
| China | Commercial Industrial Bank of China | $26.3 | $36.3 | 0.72 | $22.8 | 0.63 |
| Spain | Santander | $25.3 | $86.6 | 0.29 | $21.5 | 0.25 |
| Germany | Commerzbank | $23.7 | $37.3 | 0.64 | $27.4 | 0.74 |
| Canada | RBC Three World Center | $22.0 | $189.8 | 0.12 | $0.0 | 0.00 |
| France | BNP Paribas | $17.0 | $119.6 | 0.14 | $0.0 | 0.00 |
| Canada | Bank of Montreal (Chicago) | $15.4 | $134.1 | 0.11 | $35.3 | 0.26 |
| Germany | Landesbank Baden | $13.0 | $29.9 | 0.43 | $5.2 | 0.17 |
| Argentina | Banco Bilbao Vizcaya Argentina | $12.5 | $43.0 | 0.29 | $26.5 | 0.62 |
| Sweden | Svenska | $12.3 | $22.8 | 0.54 | $0.0 | 0.00 |
| France | Groupe BPCE – Natixis | $10.8 | $65.1 | 0.17 | $1.3 | 0.02 |
| France | Credit Agricole | $10.8 | $94.3 | 0.11 | $4.0 | 0.04 |
| Finland | Nordea | $10.5 | $28.6 | 0.37 | $0.0 | 0.00 |
| Canada | Canadian Imperial Bank | $9.9 | $87.0 | 0.11 | $0.0 | 0.00 |
| Netherlands | Rabobank | $9.9 | $67.8 | 0.15 | $27.9 | 0.41 |
| China | China Construction Bank | $9.7 | $17.4 | 0.56 | $4.4 | 0.25 |
| Canada | National Bank of Canada | $9.1 | $11.5 | 0.79 | $3.5 | 0.30 |
| UK | Lloyds Bank | $8.7 | $18.2 | 0.48 | $8.8 | 0.49 |
| Germany | DZ Bank | $8.1 | $15.9 | 0.51 | $2.8 | 0.18 |
| Japan | Sumitomo Trust | $7.6 | $41.3 | 0.18 | $0.0 | 0.00 |
| Sweden | Skandinaviska | $7.1 | $7.7 | 0.92 | $0.4 | 0.05 |
| Bahrain | Gulf Intl UK | $7.0 | $7.0 | 1.00 | $7.0 | 1.00 |
| India | State Bank of India | $6.4 | $22.6 | 0.29 | $8.8 | 0.39 |
| Australia | National Bank of Australia | $6.1 | $32.9 | 0.19 | $12.7 | 0.38 |
| Sweden | Swedbank | $6.1 | $7.1 | 0.86 | $0.0 | 0.00 |
| Taiwan | Mega | $5.6 | $8.5 | 0.66 | $0.4 | 0.05 |
| Italy | Intesa | $5.5 | $20.3 | 0.27 | $16.1 | 0.79 |
| India | Bank of Baroda | $5.5 | $15.9 | 0.35 | $5.9 | 0.37 |
| China | Agricultural Bank of China | $5.4 | $10.1 | 0.53 | $1.6 | 0.15 |
| Brazil | Banco Do Brasil | $5.3 | $7.8 | 0.68 | $1.5 | 0.19 |
| Australia | Westpac | $5.1 | $17.0 | 0.30 | $7.5 | 0.44 |
| Israel | Mizrahi Tefahot (Los Angeles) | $5.1 | $8.7 | 0.58 | $4.6 | 0.53 |
| Korea | Woori Bnk (Los Angeles) | $4.9 | $5.6 | 0.88 | $3.3 | 0.59 |
| Germany | Landesbank Hessen Thuringen | $4.7 | $14.5 | 0.32 | $8.2 | 0.57 |
| Italy | Unicredit SPA | $4.2 | $5.3 | 0.79 | $4.8 | 0.91 |
| India | Canara | $3.8 | $5.4 | 0.70 | $2.5 | 0.47 |
| Korea | Shinhan | $3.7 | $5.9 | 0.62 | $3.2 | 0.55 |
| India | ICICI | $3.5 | $3.9 | 0.88 | $3.7 | 0.93 |
| France | Credit Industrial Commercial Bnk | $3.4 | $28.2 | 0.12 | $4.3 | 0.15 |
| Brazil | ITAU Unibanco (Miami) | $3.0 | $5.5 | 0.55 | $3.9 | 0.70 |
| Japan | Norinchukin | $3.0 | $23.1 | 0.13 | $0.0 | 0.00 |
| Kuwait | National Bank Kuwait | $2.8 | $8.6 | 0.33 | $0.0 | 0.00 |
| Bahrain | Arab Banking Corp | $2.3 | $4.0 | 0.59 | $3.4 | 0.87 |
| Malaysia | Malayan Bank Berhad | $2.1 | $2.5 | 0.87 | $2.1 | 0.87 |
| Egypt | National Bank of Egypt | $2.1 | $2.6 | 0.82 | $2.0 | 0.77 |
| Norway | DNB | $1.8 | $2.5 | 0.72 | $0.0 | 0.00 |
| Netherlands | Adyen NV (San Francisco) | $1.4 | $1.8 | 0.78 | $1.2 | 0.65 |
| Germany | Bayerische Landesbank | $1.1 | $8.9 | 0.12 | $1.8 | 0.20 |
Note: Dollar values reported in $ Billions
Barclays, shown in row 1, is an interesting case. Reserve holdings for Barclays were the largest among all foreign branches on this reporting date (2025Q4) and their reserves to asset ratio exceeded 60 percent. The U.K. reportedly is one of the strictest countries regarding the currency composition of HQLA. That could be a factor contributing to Barclay’s outsized holdings of Fed balances at their New York branch.
Another interesting case among the largest holders of balances at the Fed is First Abu Dhabi. Headquartered in the UAE, First Abu Dhabi is the largest bank in the Mideast and its U.S. branch maintains nearly all of its assets in the form of balances held at the Federal Reserve (a reserves-asset ratio of 98 percent). Moreover, nearly all of its assets at the U.S. branch are financed by passthrough funding from its head offices (a due to-asset ratio of 99 percent). Funding for the U.S. branch appears to be associated with a subsidiary in the Netherlands Antilles, likely issuing Eurodollar deposits to global investors.
Concluding Observations and Two Key Questions
The basic reserve demand facts described above illustrate many underlying layers of complexity in the relationship between the “aggregate demand for reserves” and the reserve management behavior of individual institutions. Perhaps most importantly, the data point to a very wide range of reserve management practices across banks of different types and also considerable variation of reserve management practices over time among individual banks. In addition, although the aggregate reserve-asset ratio for domestic banks over our sample period was remarkably stable, that stability appears to reflect the offsetting effects of a large number of banks recording increases in their reserve-asset ratio and a small number of very large banks reducing their reserve-asset ratios. The data also suggest that the reserve management practices of branches of foreign banks are motivated by much different considerations than those applicable for most domestic banks. In particular, foreign banking organizations appear to rely on reserves held by their U.S. branches as warehouses for dollar liquidity for the global organization. Notably, based on Pillar 3 disclosures for the largest foreign banking organizations, aggregate holdings of central bank balances as a share of total worldwide assets for foreign banks are comparable to the reserve-asset ratios for domestic banks reported above. For example, although the New York branch of Barclays maintains about 60 percent of its total U.S. assets in reserve balances at the Fed, the global bank maintains only about 15 percent of its global total assets in the form of cash and balances at central banks. Similarly, while First Abu Dhabi maintains essentially all of its assets in the United States in the form of balances at the Fed, cash and central bank balances are only about 19 percent of its total global assets.[7] For Barclays, First Abu Dhabi and other large global entities, then, the outsized holdings of balances at the Fed seem to reflect the global firm’s desired composition of liquidity by currency and geographic location rather than unusually large levels of liquidity aggregated across all of their operating subsidiaries.
The basic facts uncovered in our fact-finding mission raise at least two key questions for policymakers. One key question is whether changes in important structural factors such as liquidity regulations or payments mechanisms could be adjusted to meaningfully reduce aggregate reserve demand. It certainly seems sensible that adjustments along these lines could reduce reserve demand at the margin. That said, the very substantial variation in reserve-asset ratios across banks and even among banks with similar characteristics may suggest that aggregate reserve demand is strongly affected by a range of idiosyncratic factors across banks. For foreign branches in particular, liquidity regulation and supervisory practices of their home-country authorities may be especially important factors driving their behavior.
A second key question concerns the longer-run trajectory of reserve demand over time under an ample reserves regime. The apparent stability of the aggregate reserve-asset ratio between our two reference dates might suggest that maintaining a stable aggregate reserve-asset ratio could be a reliable guide for policymakers seeking to maintain ample reserve conditions over time. However, as noted above, in our sample, most domestic banks registered significant increases in their reserve-asset ratios, perhaps pointing to a “ratcheting” of reserve demand. Over the 2018Q3 to 2025Q4 period, a steep decline in reserve-asset ratios at a handful of very large banks offset any ratcheting effect at the aggregate level of reserves held by domestically-chartered banks. On the one hand, assuming that there are no similar large outliers among the largest banks going forward, perhaps the trend increase in the reserve-asset ratios observed for most banks over the reference period would begin to register more clearly in the aggregate data. On the other hand, reserve-asset ratios cannot trend higher forever, so perhaps the upward drift in the reserve-asset ratios observed for most banks over the reference period will not be repeated going forward.
In addressing these types of questions, our fact-finding mission highlights one important consideration. While economists (present company included) often turn to “representative bank” models as a first cut in analyzing policy proposals, the implications of the great diversity of individual bank reserve manage practices noted above warrants close attention for any proposals aimed at shaping the aggregate demand for reserves.
Footnotes
[1] See Anderson et. al. (2026) and Logan and Schulhofer-Wohl (2026).
[2] See Federal Reserve Board – Assets and Liabilities of Commercial Banks in the United States – H.8 – About for additional information about the H.8 database and data construction.
[3] We exclude a handful of insured branches of foreign banks that report on the FFIEC 031/041.
[4] The sample of banks used here is not adjusted for mergers.
[5] Afonso et. al. (2025) analyze time variation in reserve demand using assets as a scale factor. They find that aggregate reserve-asset ratios of 12 to 13 percent generally correspond to ample reserve conditions.
[6] Note, the reserve-asset ratios in Table 2 are computed as the sum of reserves for all banks in each size group divided by the sum of assets for all banks in each size group.
[7] See page 320 of Barclays Annual Report for details regarding its aggregate holdings of balances at central banks and the currency composition of its liquidity pool. See page 243 of the Annual Report for First Abu Dhabi for information regarding its global holdings of cash and central bank balances.
References
Acharya, Viral and Raghuram Rajan (2024). “Liquidity, Liquidity Everywhere, Not a Drop to Use: Why Flooding Banks with Central Bank Reserves May Not Expand Liquidity,” Journal of Finance, 79(5), July.
Afonso, Gara, Domenico Giannone, Gabriele La Spada, and John C. Williams (2025). “Scarce, Abundant, or Ample? A Time-Varying Model of the Reserve Demand Curve,” Staff Reports, Number 1019, Federal Reserve Bank of New York.
Anderson, Alyssa G., Alessandro Barbarino, Anthony M. Diercks, and Stephen Miran (2026). “A User’s Guide to Reducing the Federal Reserve’s Balance Sheet,” Finance and Economics Discussion Series 2026-019. Washington: Board of Governors of the Federal Reserve System,.
Logan, Lorie and Sam Schulhofer-Wohl (2026). “Options for reducing the size of the Fed’s balance sheet,” Federal Reserve Bank of Dallas, In Depth, April.
