MAS holds its currency stance: how period averaging shapes bank cash management

The currency stance held steady while required cash increased

On October 14, 2026, the Monetary Authority of Singapore (MAS) maintained its policy of gradual Singapore-dollar effective appreciation.¹ Its new review reports approximately S$2.4 billion of additional baseline bank cash demand in FY2025/26 as liabilities expanded.² The amount banks need to arrange can change even when the exchange-rate stance stays the same.

The starting point is to avoid treating S$2.4 billion as an additional injection announced that day. Growth in demand during a past financial year and the October 14 decision to maintain policy concern different periods and different objects. Combining them into a single story of easing or tightening obscures why banks needed more cash and what the central bank was adjusting.

Use the two numbers to answer different questions

This article focuses on the Minimum Cash Balance (MCB) framework: an end-of-day floor of 2%, a two-week average of at least 3%, and a 4% counting cap, measured against the liability base.⁴ The cap is not a prohibition on holding more cash.

Together, these conditions mean that accumulating cash on the final day may not be enough. Banks must consider what to hold early in the period and what to leave for later. When required balances increase, maintaining the same ratios requires larger amounts. We examine this operational constraint, which a policy announcement alone can leave out of view.

SG Group’s conclusion is to read the increase in required cash as a question of allocating funds that support financial transactions in Singapore. The figure alone cannot establish that bank lending will contract, overseas investment will stop or the Singapore dollar will move to a particular level. Start with the liability base, the pattern of balances across the period and the transactions that move cash in and out. These narrow down where the cost of meeting the framework could emerge.

An October 14 decision and a review of the past financial year

The decision concerned the policy band for the Singapore dollar nominal effective exchange rate, or S$NEER, which combines the currencies of trading partners. MAS uses the band’s slope, width and midpoint to operate policy. This did not fix a single market exchange rate or raise a policy interest rate.¹³

Maintaining an appreciating path is different from a currency standing still. The exchange rate can fluctuate within the band, and movements against the US dollar need not match movements in the effective rate. An importer’s payments also depend on the contract currency and settlement date. Reading exchange rates and rate differentials covers those foundations; this article does not try to explain growth in banks’ cash requirements through one exchange-rate quote.

What remained in place was an appreciating path

Reading the decision alongside the operating record shows why monetary policy cannot be understood solely through changes announced on a decision day. A bank’s liability base and cash flows can change even when the framework’s ratios remain constant. On a day labelled “maintain” in the policy statement, banks still have to manage that day’s payments and their balances over the full maintenance period.

This article’s reference date is October 14, 2026. It examines the policy judgment and the newly published operating review on that date. The basic cash-balance conditions are not presented as newly introduced rules. What deserves attention is how longstanding conditions interact with bank liabilities that have changed in scale.

The same 3% can require more cash when the liability base grows

Cash amounts alone do not explain the MCB calculation. There is also a denominator: a percentage of what? The framework defines the applicable liability base.⁴ A bank’s reported total liabilities should therefore not simply be treated as that denominator: the included categories must match. Nor can growth in total assets be substituted for the base.

Consider a bank with an illustrative liability base of 1,000 units. An average of 3% requires 30 units. If the base increases to 1,200 while the ratio stays the same, the required average becomes 36 units. The additional six units are not the result of increasing the regulatory ratio. The amount needed to meet the same condition has grown because the relevant base is larger.

Comparing ratios alone misses changes in amounts

These are calculations to distinguish ratios from amounts, not observations of an actual bank. An increase of six units in required cash does not establish a six-unit funding shortfall. A bank that acquires additional cash at the same time may retain its previous margin. If funds associated with new liabilities are allocated elsewhere, it may need to rebuild balances. Assessing the burden also requires examining cash in the numerator.

Linking growth in the liability base to deposit stability requires further care. How that growth is divided by item, timing and bank affects how cash can be arranged. Funds that remain in place for a period and funds that move frequently create different preparation needs, even if the increase in amount is identical. The approximately S$2.4 billion figure cannot reconstruct those details.

The cost of arranging cash need not rise at the same rate as the amount required. A bank with existing spare balances faces a different marginal cost from one that needs additional market funding. Market rates, counterparty credit limits and room to use collateral also matter. Separating the arithmetic of a larger requirement from the calculation of its effect on earnings helps avoid overstating or understating the review’s figure.

Conversely, an unchanged ratio does not show that nothing changed for banks. Once the liability base expands, they have to recalculate whether the same cash amount remains sufficient. Treasury teams need to examine the applicable denominator and the balances arranged against it, alongside the label used for the policy decision.

The end-of-day minimum, period average and counting cap are different conditions

Remembering only the average leaves out the end-of-day condition. The floor concerns the minimum balance at the end of a day,⁴ a different question from the period average. Holding 2% every day produces a 2% average, below the required 3%. Meeting the end-of-day floor does not guarantee the average.

With an illustrative liability base of 100 units, holding 4.6 units on a day contributes only four units to the average. The remaining 0.6 units may still be useful cash for payments that day, but it does not contribute in the same way to making up lower balances on other days.

Distinguish a daily margin from a margin across the period

This distinction qualifies the intuition that more cash is always sufficient. A margin for making payments and a margin for meeting the period average overlap, but they are not identical. A bank may rationally hold more cash on a day with large expected payments. That does not mean balances above the counting cap can be freely carried forward as an additional contribution to the average for subsequent days.

Likewise, having room above the average requirement does not remove the end-of-day minimum. Even if larger early balances reduce the average contribution needed later, the end-of-day condition still applies. Collapsing both conditions into a single target balance hides the difference between a comfortable expected average with thin cash today and comfortable cash today with an insufficient period average.

MCB is not the whole of a bank’s liquidity management. Transfers and debits also occur during the day, requiring preparations for payment times as well as a plan for meeting balance conditions. The examples here illustrate the arithmetic of an average and a counting cap, rather than a procedure for judging an individual bank’s compliance. Actual transactions and reporting must be assessed under the applicable rules and the bank’s records.

Reaching the same 3% average with different early and late balances

Fix an illustrative liability base at 100 units and consider an average balance of three units across 14 days. Counted balances must add up to 42 unit-days: three multiplied by 14. A unit-day is an accounting quantity formed by adding each day’s balance, not evidence that a bank received 42 units of cash at once. The same cash remaining for several days contributes according to the number of days it stays.

The chart uses counted balances of 3.8, 3.6, 3.6, 3.6, 3.4, 3.2, 3.0, 2.8, 2.6, 2.6, 2.6, 2.4, 2.4 and 2.4 units. They total 42 and average three. Every value lies between two and four. This is neither an observed bank’s record nor a reconstruction of actual holiday accounting. It shows how a two-week average can combine larger early balances with smaller later ones.

What is gained by building the average early

Holding three units each day would also total 42. The two patterns are equivalent for the average requirement but provide different daily cash margins. Larger early balances could match the operational needs of a bank expecting more payments early in the period. The same pattern need not suit a bank expecting large payments later.

Building a contribution to the average early can leave more room to respond if later cash movements differ from expectations. It also requires more cash at that point, raising a comparison with other uses and funding costs. Meeting the average requirement early is not necessarily the cheapest allocation for the bank as a whole.

The example shows how managing a period average relates to a bank’s choices. It does not establish that an individual bank followed the chart. A possible pattern and a bank’s actual choices remain distinct.

A final-day increase cannot always make up the difference

The counting cap is clearest in an example where the average is still short on the final day. Suppose the first 13 days’ counted balances total 37 against the same liability base of 100. Reaching the target of 42 would require five on the remaining day. Yet no more than four can count on that day. Arranging five or more in actual cash would leave the counted total no higher than 41.

The highest possible average would then be 41 divided by 14, or approximately 2.93, below three. The issue is not necessarily an inability to acquire cash on the final day. Even if the bank acquires enough cash, it cannot make up the contribution of past days in a single day. Balances having stayed above the daily minimum would not eliminate this average shortfall.

Fewer remaining days leave less room to adjust

Identifying the issue with two days remaining produces a different set of possible patterns. If the first 12 days total 34, four on each of the remaining days would bring the total to 42. That arithmetic assumes the bank can actually arrange the cash for those days. Earlier identification does not guarantee funding, but knowing the number of remaining days has concrete value.

If the first 12 days instead total 35, seven is needed over the last two days. Either 3.5 on each day or a combination of three and four reaches 42. The case with an accumulated 34 required both remaining days to contribute the maximum; the case with 35 leaves room to choose the allocation. One additional unit of past accumulation changes the patterns available late in the period.

The choice depends on the end-of-day margins the bank needs and the terms on which it can arrange cash. Equal amounts on both days are not always optimal. Assigning four to a day when a larger margin is desired is another option, assuming the next day’s needs can also be met. The average calculation describes a feasible range; choosing within it requires payment and funding information.

Room in the calculation and an executable pattern

A treasury team needs to examine today’s balance, the contributions already accumulated and the days remaining. A thin accumulated contribution may require a future daily contribution above the counting cap. If so, a plan simply to raise final-day cash cannot be sufficient. This translates the period average into a daily operational question.

The example shows how the amount of cash and its pattern through the period work together. Actual counting methods, eligible accounts, holidays and reporting are subject to the applicable conditions. This article assumes neither specific weekdays nor a bank’s actual counted balances. Establishing a real shortfall would require the formal calculation and account records for that period.

System-wide growth in demand does not reveal each bank’s margin

Growth in cash demand provides an entry point for understanding the banking system. A larger system-wide requirement is different from every bank’s burden rising by the same proportion. Different growth in liability bases means different additional amounts. One bank’s demand cannot be inferred directly from another bank’s margin.

Consider two illustrative banks, each with a liability base of 100. Over 14 days, suppose A has counted balances of 2.8 every day and B has 3.2. Combined cash is six against a combined base of 200, giving 3%. Yet A averages 2.8% and B 3.2%, so A is below the 3% average requirement. These are calculations comparing aggregate and individual results, not actual bank records.

Examine scale, funding sources and concentrated cash movements

If each bank instead holds three daily, the same combined total of six meets both averages. Allocation alone changes the assessment at an unchanged aggregate amount. A’s shortfall cannot simply be netted on paper against B’s margin. Moving funds between banks requires actual transactions and changes to their respective balances.

The example also highlights the unit at which the average is assessed. Before taking reassurance from aggregate balances, establish who must meet the requirement. Even aggregate statistics with aligned periods and objects cannot establish a particular bank’s average without its counted amounts. Examining allocation complements knowledge of the aggregate amount. When banks have different bases, a simple average of ratios is not the aggregate ratio: cash and liability bases must each be summed. The same 0.2-percentage-point margin represents more cash at a larger base. Comparing conditions by ratio is different from comparing allocation by amount.

At least three dimensions need comparison: the bank’s liability base and cash amount; the days on which cash moves in and out; and the counterparties and instruments available for additional funding. Banks with similar balance margins may need different preparations depending on whether payments are concentrated or spread out. Narrow counterparty limits can also require time to find alternatives.

Changing allocation requires an actual transaction

Even within one bank, operational demand and cash management do not sit entirely in a single department. Teams that know clients’ payment schedules, central treasury and market funding teams need to share amounts and timing. A completed plan for the average requirement cannot produce a reliable daily margin if information about a particular day’s expected outflows is not updated.

Bank size alone is insufficient to assess the distribution of costs. A larger bank may face a larger additional amount while also having more instruments and counterparties available. A smaller bank may need less cash but rely more heavily on a particular funding source. The actual structure determines the advantage; size by itself does not.

Central-bank transactions addressing aggregate demand still leave the question of how cash is distributed among banks. Instruments used, transaction terms and subsequent interbank transfers can change the result. The financial-year review does not establish an individual bank’s shortage or a skewed allocation. Avoiding a jump from the aggregate figure to bank-specific earnings or credit concerns is essential when reading markets.

Public-sector cash movements affect balances separately from requirements

In FY2025/26, autonomous factors were net contractionary, mainly through public-sector flows.² Forecasting requirements alone is insufficient: changes in these inflows and outflows can call for a revised cash allocation.

It helps to distinguish the requirement calculated from the liability base from cash actually entering and leaving accounts. A larger base raises required amounts. Funds moving to the public sector, other things equal, reduce cash on the bank side. These developments can increase the pressure to arrange cash when they act in the same direction, but other inflows and market transactions can offset them.

Bring changing demand and cash movements into the same analysis

For example, a concentration of payments to the public sector on one day may be followed by funds returning to banks several days later. Banks still need to arrange cash for the intervening days. Knowing the total amount that will return does not prepare them for the earlier payment day. Where the movement falls within the maintenance period can also change its contribution to the average. Balance planning is more than reconciling total inflows and outflows at period-end.

Net withdrawal over a financial year also does not mean flows moved in the same direction every day. Withdrawals could have been larger at one point and returning funds larger at another. Inferring a shortage on the announcement day or the following day’s market rate requires daily amounts and transaction outcomes that the annual direction does not supply. The review informs structural understanding, rather than serving as a short-term price forecasting equation.

The practical question for businesses and investors is how far banks anticipate public-sector flows and how many days of required cash they arrange. Expected movements can be incorporated into a funding pattern; unexpected amounts or timing may require additional funding. Examining growth in requirements and changes in cash flows separately before combining them makes the conditions for a burden more concrete.

What the mix of cash-adjustment instruments can tell us

MAS Bills and clean borrowings together dominated the reported operations.² This describes instrument use, not the abandonment of an appreciating exchange-rate policy path.

A swap exchanges two currencies and provides for a reverse exchange later. A repo combines a securities-and-cash transaction with an agreement to reverse it. Clean borrowing does not use collateral, while MAS Bills are short-term securities issued by MAS.⁵ Contract structure, maturity and the assets or counterparty limits used by banks differ. Counting instrument names alone cannot compare banks’ costs.

Align shares, totals and periods before interpreting the mix

When reading the mix, distinguish percentages from absolute amounts. An instrument’s share can fall while its amount rises if total activity grows. It also matters whether a figure is an outstanding balance or the total of transactions over a period. Composition alone is insufficient to conclude that cash adjustment weakened or that liquidity provision increased.

Maturity also matters. Repeated short-dated transactions and a balance maintained over a longer term can support different spans of time even if their amounts are identical. A bank planning its average requirement needs to consider where a transaction begins and ends within the maintenance period. An annual composition chart describes instrument use, rather than directly showing daily funding terms or period-end cash.

International market participants may assume that an exchange-rate-centred central bank adjusts cash only through foreign-exchange transactions. In practice, several instruments address regulatory demand and autonomous flows. RBI’s dollar window and FX hedging rules distinguishes foreign-exchange transactions from domestic cash in the Indian case. Applying that distinction here directs attention to MAS’s newly reported operations and their relationship to averaging requirements and maturity.

Strong growth and banks’ required cash are different measures

Singapore’s economic growth forms part of the background to the October 14 decision. Advance estimates released that day show real GDP rising 5.4% year on year and 1.1% quarter on quarter on a seasonally adjusted basis in Q3 2026. The same release gives Q2 figures of 6.2% and 1.5%, respectively. Year-on-year and quarter-on-quarter rates compare different periods and should not be added or treated as changes in each other.⁶

MAS expects the exceptionally rapid pace of AI-related activity to moderate in the latter half of 2027.¹ A future outlook is distinct from evidence that investment has already fallen sharply. An AI slowdown and chip and memory demand examines differences in the associated types of demand.

AI provides growth context, not a direct breakdown of cash demand

Growth and cash demand can be related, but GDP cannot be converted one for one into banks’ liability bases. Even when companies’ receipts increase, the banks and currencies in which they hold funds and the timing of payments can differ. Where financial transactions increase bank liabilities, it remains necessary to establish whether those liabilities enter the framework’s base. Substituting the GDP growth rate directly for cash-demand growth is inappropriate.

The reverse inference also fails. A confirmed increase in required cash does not establish that companies made new capital investments of the same amount. Cash balances describe one aspect of banks’ allocation of funds; they do not directly measure equipment orders, construction or operation. Connecting S$2.4 billion to the volume of global AI investment would require other evidence about borrowers and contracts.

The economic context is included neither to isolate MCB completely from the business cycle nor to explain it through growth rates alone. In a firm economy, the transactions and liabilities banks handle can change. Separate evidence is needed to establish how the baseline cash requirement grew. Advance estimates can be revised, so later detailed GDP releases must also be compared on a consistent vintage.⁶

SG Group View: the burden appears in room to allocate, not just the increase

SG Group reads the operating review as a question of changing room to allocate cash, rather than cash volume alone. With unchanged ratios, a larger liability base increases the amount required. As a maintenance period proceeds, fewer days remain in which the average can be adjusted. Together, these conditions call for treasury management of both amounts and the daily pattern, rather than a single instruction to hold more cash.

What can be understated is how the counting cap limits the ability to make up past thin balances later. Funding capacity alone does not show that sufficient contributions can still be accumulated in the remaining period. The final-day example illustrates the need to review a plan early, separately from the speed at which cash can be obtained.

The days that cannot be recovered deserve attention

Benefits and costs are not distributed uniformly. Larger early balances can help a bank prepare for uncertain later movements, while requiring funding costs or foregoing other opportunities. Counterparties providing additional funding may gain business, but prices, credit limits and collateral conditions can narrow a bank’s options. Who benefits depends on balances and funding terms together.

The international financial connection lies in managing Singapore-dollar funds

The connection to the world economy is not a power to move every overseas interest rate. It lies in how businesses, financial institutions and treasury teams using Singapore-dollar payments and funding encounter banks’ cash requirements and market funding conditions. Even a globally active company’s local-currency payments depend on the banks and market handling that currency.

If larger demand affects short-term market conditions through banks’ additional funding, businesses need to examine how deposit or short-term funding quotes change. Greater MCB requirements are not the same as corporate borrowing rates rising at the same moment by the same amount. Credit, tenor, product and alternative sources also enter pricing. How rates reach households and businesses explains the foundations of this transmission.

Reaching businesses requires a change in banks’ transaction terms

If banks can arrange required cash reliably, businesses’ payment terms may change little. The larger regulatory requirement may be absorbed by treasury operations. Whether additional costs are passed to customers or borne within the bank also depends on competition and contracts. The increase in required cash should not simply be converted into a rise in household or business costs.

For businesses using several banks, their own banks’ terms are closer to operational needs than a system-wide direction. When planned cash transfers, payroll, supplier payments and loan renewals coincide, it is useful to confirm where and when funds will be used. This is a routine alignment of payment plans with funding conditions, rather than a judgment that a particular bank has a problem.

Investors likewise cannot determine banks’ margins or the direction of the currency from demand growth alone. They need to know whether other income offsets costs and how a bank’s funding terms change relative to competitors. The exchange-rate policy decision and the cash-balance operating record are useful when applied to the different questions each illuminates.

What should the cost of holding more cash be compared with?

Arranging required cash early means retaining funds that could have another use or raising additional funding. The first involves a choice of uses; the second a choice of funding terms. Grouping both under the cost of holding cash makes the source of an increased burden less clear. Separating balances for the average requirement from balances for unexpected payments provides a starting point for comparison.

Funding quotes need to be compared at consistent maturities, not just by rate. Cheap funding repaid midway through a maintenance period leaves later balances to be arranged elsewhere. Longer-term funds can support the whole period more readily, but may continue to incur costs after the need has fallen. The pattern of the average and daily payment schedules determine suitability.

Cheap funding and flexible funding are not always the same

A collateralized transaction also involves allocating the securities used as collateral. If those securities are needed elsewhere, their uses have to be coordinated. In an uncollateralized transaction, counterparty limits can constrain the amount available. Comparing prices across instruments does not mean a bank can switch freely if usable transaction limits differ.

Assessing earnings effects therefore requires aligning average balances, funding tenors, prices and instruments. We do not multiply S$2.4 billion by a single market rate and call it the system’s cost. The period for which the increase remained outstanding and the prices at which it was funded are unknown. Identifying missing conditions is more useful than compressing earnings effects into a plausible-looking amount.

How to assess the argument that this is merely routine operations

A credible alternative reading is that greater demand and early accumulation of cash are routine matters for banks and do not reveal a special problem. This argument has merit. Where a longstanding framework is anticipated and banks arrange the required amounts, an increase in demand need not destabilize funding.

This interpretation and our analysis do not need to disagree over demand growth itself. The question is how funding instruments, costs and room through the period changed in responding. If banks can fund the larger amount with the same cost and flexibility, the case for emphasizing operational burdens becomes weaker. The increase provides a reason to investigate rather than proving higher costs.

Separate adjustment to the framework from evidence of a problem

Another interpretation is that larger early balances show strong funding anxiety. Yet accumulating the average early has an ordinary rationale under a framework with a counting cap. Establishing increased anxiety would require additional comparable evidence, such as earlier accumulation than before, different margins or changing funding prices. An action being present is not the same as its motivation having changed.

Net annual withdrawals might also be described as policy tightening. Public-sector movements affect bank cash, but their direction alone does not show that the exchange-rate policy band changed. Reading evidence of policy decisions separately from evidence of transactions that moved cash allows this claim to be tested. UK quantitative tightening and asset maturities provides a related guide to interpreting central-bank balance-sheet changes.

The most effective way to test objections is to add observations rather than choose stronger language. Preserved balance margins and stable transaction terms support the routine-operations explanation. Greater requirements coinciding with changed terms for certain banks or periods would reveal where a simple routine explanation falls short. The annual total alone settles neither case.

Three conditions would change how the evidence is read

The first condition is that banks continue to arrange required cash reliably as the liability base expands. Preserved room against the average requirement and broadly stable funding terms would make the increase easier to understand as operations accommodating larger business volumes. Markets would see a larger requirement that banks successfully absorbed.

The second is that larger requirements coincide with large outflows and a need for additional funding late in the period. Counted contributions still possible over the remaining days, available transaction limits and funding costs then matter. Earlier identification of a thin average can provide more options, but maturities and counterparty terms determine which plans can be executed.

Requirements and the capacity to meet them work together

The third is that liability-base growth settles, required amounts stabilize and funds return to banks. The accumulated average margin and expected future payments may leave room to reconsider spare cash. Stable requirements alone do not guarantee more available cash, however; other outflows can produce the opposite outcome.

The diagram branches into evidence to examine, rather than forecasts with assigned probabilities. Identify the currency, tenor and product where change appears, then compare aligned figures. Stable capacity to meet requirements can coexist with currency movements driven by overseas rates, growth and risk appetite; the branches are not trading instructions.

What the available evidence cannot establish

The available figure does not show how approximately S$2.4 billion of greater demand was divided by bank, liability category or period. It also does not describe cash margins or the prices of additional funding. The increase is important, but bank-specific shortages, missed average requirements and earnings effects require other records.

The next evidence to examine: requirements, balance patterns and terms

A useful sequence is to establish changes in the liability base, calculate required amounts, and then examine balance patterns and funding terms. Starting with a price move and attributing it to larger cash demand risks confusing overseas rates or other flows. Establishing the amount the framework requires helps narrow the possible causes.

When using bank disclosures, check whether the published deposit or liability categories match the MCB base. A detailed published total need not be the regulatory denominator. If the objects cannot be aligned, a self-calculated requirement has limits. Keep observed figures separate from figures supplied through assumptions.

Align comparison periods and update the hypothesis

Testing the balance pattern requires the relevant period and daily counted amounts as well as an average. The same average can arise from different early and late allocations. Without daily records in public materials, an estimate of compliance should not be presented as an observed outcome. Using the illustrative chart to understand the arithmetic is different from judging an actual bank.

Comparisons of funding terms should align amount, currency, maturity and collateral treatment. A single short-term rate does not show that a bank could fund its required amount for the relevant time. Examining prices alongside usable limits helps test whether greater demand transmitted into cost. Examining bank lending and the credit cycle provides further guidance on reading evidence of bank credit provision.

Detailed GDP releases and revisions should be used to assess the breadth of growth on the same basis as these advance estimates. Operating evidence about bank cash and economic evidence should each answer the appropriate question.

Final assessment: unchanged ratios can mean different amounts and days

The news brings an unchanged exchange-rate stance and increased annual cash demand into view on the same publication day. Unchanged policy parameters do not mean banks’ required amounts were unchanged. A larger liability base requires more cash to meet the same 3% average.

Averaging leaves room to allocate balances through the period, but the daily minimum and counting cap qualify that room. Examining early contributions and the amounts that can still count over remaining days shows why final-day funding cannot always resolve a shortfall. This is a specific constraint on managing balances, rather than a diagnosis of anxiety at banks.

For businesses and financial institutions using Singapore-dollar funds internationally, their funding terms and payment-day margins are closer to practical needs than the aggregate increase alone. Instead of reducing the figure to a crisis or easing label, examine the liability base, balance pattern and funding terms. These show where a burden could emerge and what evidence should change the assessment.