Nostro account reconciliation process with daily steps, break types and break aging by days

A day, in order

The six differences, and what each one means

FX rate source

Same reference, same value date, amounts a fraction of a percent apart. The correspondent priced the conversion on their source and you priced it on yours. This is a variance to post, not a mystery to investigate, and it should be a configured tolerance with an approved percentage rather than a daily judgement call.

Timing and value date

Same amount, different date. Cut-off times and time zones mean an item posted on your Tuesday lands on their Wednesday. Set the window per correspondent. A single global window is either too tight for the far correspondents or too loose for the near ones.

Charges not accrued

Correspondent fees, lifting fees and periodic charges are deducted at source. If the accrual does not expect them, the amount will never agree. Every one of these that surfaces daily is an accrual that should be fixed once.

Duplicate posting

A re-presented instruction or a replayed batch posts twice on one side. The reversal is a maker-checker action with a reason code, and the original item should carry a note that a duplicate was found against it.

Partial settlement

One side settled short. The correct handling is to split, match the settled leg and leave the residual open against the same reference, so the outstanding exposure stays visible instead of being absorbed by a wide amount tolerance.

Missing reference

Something exists on one side with nothing to tie it to. This is the class that becomes a correspondent investigation, and it is the class that shrinks most when structured payment data survives the chain.

Why the aging number is the one to watch

Break counts fluctuate with volume and mean little in isolation. Age does not. An item open for four days on a nostro account is either a genuine dispute with a correspondent or a decision nobody has been willing to make. Both are worth surfacing on the first screen, which is why the resolver in this site puts the oldest break in the header strip next to the value at risk.

How many days should a nostro break stay open at a commercial bank?

No US regulator publishes a benchmark number of days. What examiners and auditors test is that your own policy defines aging thresholds and an escalation path, and that the open items show you follow it. A common design ages each break from the value date into buckets such as 0 to 2, 3 to 5, 6 to 30 and over 30 days, with treasury or finance management notified when an item crosses into the next bucket.

Two details decide whether that policy holds up. Age must count from the value date, not from the day someone first noticed the item, or the aging report flatters itself. And an item must not reset its age when it is re-referenced or split: a partial settlement keeps the age of the original instruction. Software that ages items automatically and shows the oldest break on the first screen removes most of the argument, which is the core of nostro reconciliation automation.

Nostro, vostro and loro, settled once

The vocabulary causes more confusion than it should, and it is worth fixing in one paragraph. A nostro account is our money held in their books, usually in their currency. A vostro account is their money held in our books, usually in our currency. Loro is the same relationship described from a third party position and appears mostly in older documentation. The account is one account; the name depends on who is speaking.

Operationally the distinction matters because the reconciliation is different in shape. For a nostro, the external record is a statement you receive and cannot control, and the internal record is your mirror. For a vostro, you are the record keeper: you produce the statement, and the reconciliation is between your ledger and what the account holder believes. Breaks on a vostro tend to arrive as queries rather than as unmatched rows, and the evidence you need is the evidence you already hold.

The mirror is where most problems begin

The internal side of a nostro reconciliation is often called the mirror, and how it is maintained determines how much work the reconciliation is. If the mirror is updated when a payment is instructed, it will differ from the statement by everything that has been instructed and not yet settled, which is a normal and predictable population. If the mirror is updated when settlement is assumed, it will differ by everything the assumption got wrong, which is not predictable at all.

The useful discipline is that the mirror should record the state a payment is in, not just its existence. An item that is instructed, an item that is confirmed by the correspondent and an item that has appeared on a statement are three states, and a reconciliation that can distinguish them turns most timing differences into expected items rather than into breaks.

Intraday, and why it is a funding question

End of day reconciliation proves what happened. Intraday reconciliation supports a decision that can still be made. The distinction is the entire argument for reading camt.052 and camt.054, or MT942, rather than waiting for the statement.

The funding argument is concrete. A treasury team that knows at eleven in the morning that a large expected receipt has not arrived can act. The same team learning it from a statement the next morning is describing history. Reconciliation timeliness on a nostro account is a liquidity capability before it is a control capability, and it is usually easier to fund on that basis.

Correspondent charges, in more detail than they usually get

Charges deserve their own section because they generate a disproportionate share of small breaks and because they are almost entirely predictable. There are three common shapes. A per item charge deducted from the amount, which makes what arrives smaller than what was sent. A lifting fee taken by an intermediary, which does the same thing in the middle of the chain. And a periodic charge posted separately, monthly or quarterly, which appears on the statement with nothing in the ledger to meet it.

Each has a correct handling and none of them is a wider amount tolerance. The per item and lifting charges should be predicted from the charge bearer instruction and the correspondent agreement, so the expected settlement amount is the net figure. The periodic charge should be accrued so that a posting exists to match. A charge that surfaces daily as a break is an accrual decision that has never been made, and it will keep surfacing until it is.

The correspondent conversation

Some breaks cannot be resolved internally, and the quality of the correspondent relationship decides how quickly they clear. Two habits make a large difference. The first is quoting a single identifier rather than describing a payment: a uetr or an end to end id turns a discussion into a lookup. The second is tracking investigations as items with an age, in the same worklist as everything else, rather than in an email folder. An investigation that is not aged is an investigation that closes when someone remembers it.

The liquidity argument

The control case for nostro reconciliation is accuracy. The commercial case is liquidity: unreconciled balances are cash you are holding because you cannot prove you can use it. Every day the reconciliation is timely, the treasury team is working with a real position instead of a cautious one.

Account structure, and a cost most banks under-examine

The number of nostro accounts a bank holds is a decision with a running cost, and it is frequently inherited rather than chosen. Each account carries a statement to ingest, a set of tolerances to maintain, a relationship to manage, charges to accrue and a balance that has to be funded. Consolidating currencies onto fewer correspondents reduces all of that, at the price of concentration and, sometimes, of losing a corridor where the remaining correspondent is weak.

The reconciliation view of the question is a useful input to it, because the data answers it directly. Rank accounts by break volume, by aged break value and by the proportion of items requiring an investigation. Accounts that are expensive to reconcile are usually expensive for structural reasons: a correspondent with poor reference handling, a corridor with heavy charge deductions, or a low volume relationship that nobody has time to keep clean. That ranking is a better basis for a rationalisation discussion than a balance sheet view alone.

What a good nostro control looks like on a Tuesday

• Intraday reports are read and matched against pending items during the day, not stored for later.
• The end of day statement arrives on a schedule and is ingested without a person pressing anything.
• Exact match rate is high and stable, and the tolerance band is small, per correspondent, and approved.
• Charges are accrued or predicted, so a fee difference is an expected item rather than an investigation.
• Every open item has a classification, an owner and an age counted from the value date.
• The oldest item is visible on the first screen anyone looks at, next to the value at risk.
• Evidence for the day exists because the work was recorded, not because someone assembled it.

None of that requires heroics. It requires the data to arrive reliably, the tolerances to be decisions rather than habits, and the residue to be treated as work with an owner instead of as a report with a distribution list.

Product detail is on nostro reconciliation automation, and the sample dataset in the resolver is exactly this reconciliation, including an FX rate source break and a charge nobody accrued.