Where in-house bank account processes break, and how to spot each failure

An in-house bank usually breaks where money crosses a boundary. Receipts arrive that nobody can assign to a subsidiary, and intercompany positions get booked one way centrally and another way locally. The damage surfaces at period end, when the in-house bank statement and the subsidiary ledger refuse to agree.

Incoming receipts with no clear owner

Central collection works well when customers pay with a reference that points to one subsidiary and one invoice. Many do not. They pay several entities in a single transfer, short pay, or quote an old account number. Treasury receives the cash but cannot credit the right internal account, so the money sits in a suspense position while someone emails around for clues.

Returned items make it worse. A bounced or recalled payment has to be reversed against the same subsidiary that was credited. If the original credit went to the wrong place, the reversal lands wrong too.

How to tell: the unapplied cash balance never quite clears. Subsidiaries chase customers for invoices that were in fact paid. Treasury staff keep a private list of "probably belongs to" notes.

Outgoing payments released on stale instructions

When the in-house bank pays on behalf of a subsidiary, it depends on payment files built locally. The local team may have changed a supplier's bank details, cancelled an invoice, or approved an urgent run by phone. None of that reaches the central payment queue in time.

The result is duplicate payments, payments to closed accounts, and manual recalls that need cooperation from the external bank.

How to tell: a steady trickle of payment rejections and recall requests. Urgent payments routed outside the central file. Local controllers who keep a separate bank account "just in case."

Netting and intercompany loans out of step

Netting assumes both sides of every intercompany invoice agree on amount, currency and date before the cycle runs. They often do not. One entity has booked the invoice; the counterparty is still disputing it. The netting run settles a figure that only one ledger recognises.

Intercompany borrowing has a similar gap. Treasury records a drawdown on the internal loan, while the subsidiary books it as a current account movement, or not at all until the paperwork arrives.

How to tell: intercompany mismatches that reappear every cycle under new descriptions. Manual journal vouchers raised to force the two sides together. Disputes resolved by "netting it next time."

Interest and fee calculations that subsidiaries challenge

Interest on internal accounts depends on value dates, rate tables and the agreed charging method. Small differences in any of these produce figures the subsidiary cannot reproduce. If the method lives in a spreadsheet maintained by one analyst, nobody else can explain it.

How to tell: subsidiaries accrue their own estimate and then post a true-up when the real charge arrives. Queries about interest outnumber queries about payments. Rate changes are applied late or retrospectively.

Statements nobody relies on

An in-house bank statement should let a subsidiary reconcile its internal account the way it reconciles an external one. When the statement lacks transaction detail, uses central references the local team does not recognise, or arrives after the local close, people stop using it.

How to tell: local finance rebuilds the statement from emails and payment confirmations. Requests for "a breakdown of this line" are routine.

The period end collision

Every issue above converges when books close. The in-house bank posts its interest, fees and netting settlements. Subsidiaries post their own versions. If the account mapping between the two ledgers was set up inconsistently, even correct transactions land in different places, and elimination entries fail.

The tell here is simple. Consolidation teams spend the close clearing intercompany differences that trace back to in-house bank activity, and the same accounts cause trouble every period.

Workarounds that quietly become the process

Most in-house banks run on a layer of informal fixes. A suspense account that holds unidentified cash indefinitely. A shared mailbox where subsidiaries request urgent payments. A manual override on the interest calculation for one awkward entity. Each fix made sense once. Together they mean the documented process describes something that no longer happens.

Before redesigning anything, find these. They show exactly where the formal design failed.

Questions to ask the people who run it

• When cash arrives and the reference is unclear, what do you actually do with it, and who decides where it goes?
• How do you hear about a payment that a subsidiary wants stopped or changed after the file has been sent?
• Which intercompany differences do you already expect to see before the netting run starts?
• If a subsidiary disputes its interest charge, who can explain the calculation, and where does that explanation live?
• What do local teams use to reconcile their internal account if they do not trust the statement?
• Which manual journals do you post every period that nobody has formally approved as standard?
• Is there any payment or receipt that still bypasses the in-house bank, and why?
• At close, which accounts do you check first because they are usually wrong?

APQC's Process Classification Framework® (PCF) is an open standard developed by APQC, a nonprofit that promotes benchmarking and best practices worldwide. To download the full PCF or to view definitions and measures, please visit www.apqc.org/pcf.