Inside the pooled reserve claim
This chapter opens the Synthetic Reserve Deposit (SRD) at the level where its promise becomes testable: the balance sheet. The focus is not whether pooling sounds diversified, but what happens to value and cash when a participant contributes reserves, redeems units, or borrows against them. Tracing those interfaces makes the central judgment precise: the SRD can change who manages liquidity and when assets are sold, but the market risk of the underlying portfolio remains with the claim holder.
A floating claim backed by two kinds of pool assets
The proposed arrangement connects three balance sheets. A participating central bank contributes reserve assets and receives units issued by the pooled vehicle. The vehicle holds two broad categories of assets: cash and near-cash instruments for settlement, and an investment portfolio chosen to pursue a better return-liquidity trade-off. If a standing borrowing window exists, a third balance sheet—the liquidity provider—advances foreign currency against pledged units. Each actor has a different responsibility. The participant owns the economic exposure; the pool manages the common assets and the rules for issuing and redeeming units; the liquidity provider supplies short-term cash under collateral terms.
The benchmark vehicle is deliberately simple. It has no external leverage, no junior capital layer, and no promise to return a fixed principal amount. Its net asset value is the contemporaneous value of the assets it holds. The value of one unit is that net asset value divided by the number of units outstanding. If the investment portfolio rises, the unit value rises; if bonds lose value after an interest-rate move or equities fall during a risk-off episode, the unit value falls. Pooling changes how the assets are selected and managed. It does not turn their market value into a government-guaranteed amount.
That distinction matters because “deposit” can suggest a fixed-value liability. The SRD paper uses the term for a reserve claim, not for a conventional bank deposit protected from loss. A participant that exchanges directly held securities for SRD units has changed the form of ownership, but not escaped exposure to the pool’s portfolio. The participant’s claim is floating-value precisely so that gains and losses are recognized rather than hidden in an implied guarantee.
The settlement buffer has a narrower role. It keeps some cash or near-cash assets inside the pool so that ordinary withdrawals can be settled without selling investment assets every time a participant needs funds. That can reduce trading frequency, execution costs, and the risk that small, unrelated flows trigger unnecessary asset sales. But the buffer does not create a senior redemption claim. It does not make the first dollars withdrawn safer than the remaining units, and it does not promise that the pool can meet every synchronized withdrawal from cash on hand.
The redemption rule determines how losses are shared. When a participant redeems, the relevant value is the pool’s contemporaneous executable value after the transactions required to meet the aggregate redemption have taken place. If the cash buffer is enough, the pool pays from cash. If it is not, the pool sells assets and the prices obtained in those sales enter the value used for redemption. Units that leave and units that remain are valued on the same basis.
The buffer is therefore not a hard ceiling on total redemptions. The benchmark rules allow aggregate redemptions up to the number of units outstanding; the constraint is not that cash on hand must equal every possible claim, but that the pool must execute the necessary asset sales at current market prices. This is an important interface between the pool and the market. A larger buffer can make ordinary settlement smoother and reduce expected trading, but it cannot make a full or synchronized withdrawal costless.
The accounting after a redemption follows from the same principle. The pool’s post-transaction value is what remains after paying the redeemers, and the remaining units represent that residual value. If current prices have fallen, both the redeemed position and the residual pool reflect the effect of selling at those prices. There is no separate senior class whose value is protected by silently diluting the units that stay behind. The design is intended to make the allocation of market risk transparent even when the timing of cash needs differs across participants.
A simple example makes the rule visible. Imagine a pool with a cash buffer and a longer-term portfolio, and a participant whose withdrawal is larger than the cash available. If the pool must sell part of the portfolio after prices have fallen, the participant receives the value that can actually be realized, not the value recorded before the sale. The remaining holders do not receive a hidden subsidy, and the withdrawing participant does not receive a protected price. The exact numbers in such an example are illustrative; the institutional point is that the valuation rule refuses to turn liquidity into a promise of price stability.
This also prevents a misleading comparison between the SRD and a fixed-value reserve asset. A short-duration government bill may often be sold with little price movement, but the SRD’s broader portfolio is expected to have more exposure to duration, credit, and equities. Its justification must therefore come from a combination of portfolio construction and a credible cash route. The pool can reduce duplicated management and make a diversified claim easier to monitor, but it cannot make a claim worth more than the assets and obligations that stand behind it.
The same rule applies whether withdrawals are isolated or widespread. In a small, ordinary redemption, the buffer may absorb the payment and leave the investment portfolio untouched. In a large or synchronized redemption, the buffer may be only the first layer before asset sales or secured borrowing become necessary. The pool’s size, the liquidity of its assets, the timing of withdrawals, and the execution prices all affect the result. None of those variables changes the basic allocation: the holder of the floating claim remains exposed to market value.
This makes the SRD different from a promise that cash will always be available at par. Its liquidity is a rule for converting a claim into current value, supported in ordinary states by cash and in difficult states by executable sales or a separate secured advance. If markets are deep and withdrawals are not synchronized, the arrangement may avoid much of the friction of separate country portfolios. If markets are stressed and many participants redeem together, the same design exposes the constraint that no balance sheet can escape: cash must come from assets, borrowing, or a provider willing to accept collateral. The pool can coordinate those sources; it cannot create purchasing power by declaration.
Borrowing against the claim without transferring its market risk
The optional standing borrowing window addresses a different problem from redemption. A participant may need foreign-currency cash for a short period but prefer not to sell SRD units while the underlying assets are temporarily depressed. Instead, the participant pledges some units and receives an advance. The claim remains outstanding, so the participant retains exposure to any later recovery as well as to any further decline. The transaction changes the timing of liquidation; it does not change who owns the investment risk.
The lender does not advance the full market value of the collateral. A haircut creates a cushion between the pledged units and the cash advanced, and a participant-specific borrowing limit can restrict concentration or exposure. The advance also has a short maturity and a stated repayment obligation. These details are not bureaucratic decoration. They determine whether a temporary loan can be supplied without asking the liquidity provider to absorb the long-term volatility of the pooled portfolio.
The distinction between the two cash routes is therefore sharp. Redemption ends the participant’s position, or part of it, at the pool’s contemporaneous executable value. Borrowing leaves the position in place but creates debt secured by that position. In the first route, a market sale may crystallize a loss immediately. In the second, the participant postpones that sale while paying the cost and accepting the conditions of a collateralized advance. Neither route guarantees that the underlying assets will recover.
The borrowing window also has a lifecycle that redemption does not. At origination, the advance is set below the value of the pledged units by the haircut. During the loan, the collateral is revalued and the participant remains responsible for keeping coverage above the required level. At maturity, the advance must be repaid according to its posted rate and term, or the collateral may have to be realized under the facility’s rules. The participant has gained time, but not a free option on the pool. If the market decline continues, the borrower’s equity in the claim is still reduced and the margin requirement may become harder to meet.
That lifecycle is why the facility can be useful without being a second reserve portfolio. Its job is to bridge a short period in which selling may be especially costly. It is not designed to finance a permanent withdrawal, cover an unbounded loss, or allow a participant to keep using pledged units as if they were freely available reserves. The contractual limits are the boundary between liquidity support and leverage against a risky claim.
Margining is the feedback mechanism that keeps the borrowing route from becoming an unsecured promise. At the next valuation, the lender compares the outstanding advance with the haircut-adjusted value of the pledged units. If the collateral gap is positive, the participant can post additional eligible collateral, repay part of the advance, or reduce or substitute the pledged position. Only if those corrections fail does forced liquidation arise within the lending facility.
This design places the public or quasi-public layer in a limited role. It can provide a pre-positioned, short-maturity advance against collateral when private funding capacity or private-dealer incentives are least reliable. It does not sustain the unit value, guarantee principal, or make redemption possible; the pool’s redemption rule already provides a route to cash at contemporaneous value. The provider is protected through haircuts, access caps, and margining, while ordinary market losses remain with SRD holders.
The paper gives two reasons a public layer might be considered even when private lenders exist. A private dealer active in related sovereign or funding markets may learn something from a participant’s request for cash, creating an information friction around the act of borrowing. Private funding capacity can also contract in precisely the states in which reserve liquidity is most valuable. A rules-based public backstop can bypass those particular vulnerabilities, but only by accepting a narrower task: lend against pre-positioned collateral for a short maturity. It is not asked to underwrite the investment portfolio or to socialize its losses.
The distinction protects incentives as well as the lender. If a guarantee absorbed the portfolio’s ordinary losses, participants could receive the upside of diversification while shifting the downside to the public provider. The benchmark architecture avoids that arrangement. Participants own the floating claim and its market outcomes; the liquidity provider receives compensation and collateral for bridging timing; and the pool’s operating rules determine how redemptions are valued. Whether those incentives would remain credible in a real facility would depend on legal enforceability and governance, which the paper leaves open.
That allocation produces an important but easily missed consequence for reserve accounting. Units that a participant has pledged are encumbered. The participant may receive immediately usable foreign currency from the advance, but the pledged units are not freely available reserves while they secure the loan. Borrowing against an asset does not make the asset unencumbered, and the existence of a liquidity facility does not by itself establish reserve-asset status.
Reserve recognition is a separate institutional test. The paper says that an SRD claim would need to be an external foreign-currency asset controlled by the monetary authority and redeemable or liquidatable at contemporaneous executable value within the time required for ready availability. Domestic law may restrict fund shares, equity exposure, or particular accounting classifications even when the pooled portfolio looks economically suitable. A claim can therefore be useful as a financial instrument and still fail to count as official reserves for a particular participant.
The distinction is practical because the legal form can change the accounting treatment without changing the underlying portfolio. A mutual-fund-like share may be recorded as an investment-fund share or equity security even when the vehicle owns fixed-income instruments. That classification may not automatically rule out reserve treatment, but it can interact with domestic rules that limit fund-share or equity holdings. Recognition must therefore be evaluated for the claim itself, the authority that controls it, and the time and conditions under which it can be liquidated—not inferred from the ratings or liquidity of the assets held inside the pool.
The same is true of governance. Someone must define the eligible asset universe, maintain the settlement buffer, publish or monitor the portfolio rule, determine executable value, enforce collateral terms, and resolve a margin shortfall. The IMF paper does not present a completed operating charter for those functions. It describes an analytical architecture whose legal form, custody arrangements, currency composition, and liquidity-provider commitments remain implementation questions. Applying the logic to a multicurrency or SDR-basket portfolio would add another design choice rather than remove the need to answer them.
The SRD’s balance sheet can now be stated in one sentence. It centralizes diversification and can create a separate route to short-term cash, but it leaves the value of the pooled portfolio floating and leaves its market gains or losses with the participants. The next test is consequently empirical: given these rules, can a diversified reserve-style portfolio improve the return profile without violating the downside and usability constraints that made reserves valuable in the first place?