This page provides a detailed description of the assumptions on which the value model is based and the way in which the calculations of equivalent FTEs are calculated. This description is broken down into the key activities of a typical clearing Bank’s systems.

  • Cash Withdrawals
  • Entries and Daily Balance Calculations
  • Daily Reconciliations
  • Charges Calculations
  • Interest Calculations
  • Payment Processing
  • Standing Order and Brought Forward File Processing
  • Statement Production

Cash Withdrawals

An average UK Clearing Bank Group processes around 1million cash withdraws through its ATM’s per day. Let us assume that branch cash withdrawals are negligible volume. In our manual bank these cash withdrawals would be made at a branch teller.

Let us assume that the customer fills out a cheque or a branch cash withdrawal form in their own time. The cost to the bank of the transaction is:

Teller staff checking the voucher (10 seconds).

Then either:

  1. If the customer has his account with the branch –
    • Going to the hard copy ledger to check the available funds (1 minute) .
    • Updating the ledger with the entry if the customer has his account with the bank (30 seconds).
    • Return ledger to its home
  2. If the Customer has his account with another branch –
    • Phoning the customer’s branch to see if the customer has funds (20 seconds).
    • Someone at the customer’s branch going to the customer ledger to check if the customer has available funds (1 minute).
    • Someone at the customer’s branch updating the ledger to show the funds removed (30 seconds) and then
    • Return ledger to its home (1 minute).

Back at the paying branch, count the money (twice to check) and give to customer (30 seconds).

Let us assume 30% of cash transactions are withdrawals from the customer’s own branch.

A calculation based on these assumptions shows that, for cash withdrawals a bank would require an additional 8,800 FTE’s. This is assuming an 8 hour day with 80% efficiency. Clearly this does not meet the 24 X 7 availability of ATM networks, nor their geographic convenience. On a peak day (e.g. near Christmas the daily volume and required FTE’s probably increases by 50%).

Entries and Daily Balance Calculations

An average UK Clearing Bank group has around 15m customer accounts and processes approximately 7m non-cash entries onto those accounts. Let us assume that apart from the cash entries which are entered direct into the ledgers, all the other entries are put into a provisional ledger (let us call this the AM ledger). A provisional balance is calculated using the previous nights closing balance and these entries at the start of the banking day. After the bank has closed, these provisional entries are transposed to the customer ledger and the closing balance is calculated. (This two-stage process is used to enable branch staff to decide whether to bounce cheques, standing order payments, etc).

Let us assume that on any given day three quarters of accounts do not receive an entry so the 7m entries are spread over about 4m accounts. We have the following activities for each account.

  • Obtain the entries (vouchers) and ledger for the customer for that day (1 minute).
  • For an account with entries write in the AM ledger with the new entries, the date, the overnight balance (1 minute).
  • Calculate the new balance and write in the AM balance (1 minute).
  • Pass the customer account ledger to a checker. (10 seconds).
  • The checker checks the correctness of the entries and the calculation (1.5 minutes).
  • Checker returns the ledger to its home. (1 minute).

In the afternoon there is a similar process for transposing the AM entries into the customer ledger and calculating/checking the closing balance.

A calculation based on these assumptions shows that for entry processing and balance calculations a bank would need an additional 110,700 FTE’s assuming an 8 hour day and 80% efficiency (in this case probably evening or night shift work). The volume of month ends and quarter ends can double or even triple in terms of numbers of entries and accounts affected with a corresponding peak in FTE’s.

Daily Reconciliations

The fundamental unit of control in a typical UK Clearing Banking model is the branch. Let us assume there are 2,000 branches in this case. For each branch there are a number of control totals and/or General Ledger aggregation points for financial control and reporting purposes. (E.g. Total Cash Withdrawals, Total Debit Interest Posted, Total Cheques awaiting clearing etc.). Let us assume there are 1,000 such totals (a not unreasonable number for a modern UK clearing bank). Further, every balance and entry has to be allocated to one or more of these control totals which are kept in separate ledgers.

Thus the activity would be:

  • During the close of business entries and daily balance calculation process (see “Entries and Daily Balance Calculation” section) copy the entries and balances into a “category sheet”. There is one category sheet for each control total I.e. an interest sheet, loans drawdown sheet etc. (20 seconds). As described in other sections, it is assumed that there are 7m non cash entries per day, 1m cash entries and 4m balance movements per day.
  • For those accounts with no entries, copy the previous night’s category sheet balance into today’s sheet. (10 seconds). Typically 11m balances per day.
  • Sort all the category sheets from each customer account into one of 1,000 piles (30 seconds per sheet for 23m sheets).
  • For each category copy the data off the category sheet into the category ledger. (20 seconds per sheet, for 23m sheets).
  • For each category add up the entries and write total in the ledger. (1 minute per total, 1,000 totals per branch, 2,000 branches).
  • A checker checks the total. (1 minute per total, 1,000 totals per branch, 2,000 branches).

Let us assume that this process is no more inaccurate that the current computerised reconciliation process so no extra labour is generated by extra mistakes. (A big assumption). A calculation based on the above shows that a bank would need an extra 59,400 FTE’s to perform financial reconciliation. This assumes an 8 hour day and 80% efficiency. This figure would probably double at month end and quarter end.

Charges Calculations

The personal customer base enjoys free “in credit” banking. Let us assume that of the 15m accounts, 2m correspond to non personal customers and that all these are eligible for periodic fees, which it is assumed, are calculated monthly. In our manual model, this would involve.

  • For each non-personal account, count from the customer ledger, quantities of entries by transaction type (e.g. cash deposit, cheque, standing order etc). Write the totals onto a charges and interest sheet. (5 minutes).
  • Multiply each total of entries by the unit cost per transaction type and then total to arrive at the overall cost to the customer. (3 minutes).
  • Independent check of calculations. (3 minutes).
  • Type out customer charges and interest statement (1 page of A4, 40 or 50 lines including customer name, address, bank account number and charge breakdown). (10 minutes).
  • Put in envelope and write name and address on envelope. (20 seconds).
  • Create a pair of charges posting vouchers and put in the brought forward file. (1 minute).

It would be in the bank’s interest to spread these calculations over the month but in practice they are clustered around month end for many banks/customers nonetheless. We will assume that there are 22 working days per month and that the above work can be spread evenly over 22 days.

Using the above, an 8 hour day, and 80% efficiency, it would require 5,300 extra FTE’s to calculate and pre-notify the charges each month.

Interest Calculations

Let us assume that 9m of the 15m customer accounts earn interest (deposit or loan/overdraft). Most banks do a daily accrual of interest with a monthly posting. In our manual bank this would need the following daily activities:

  • Copy today’s closing balance onto interest calculation sheet for account. (30 seconds).
  • Multiply the balance by the (previously calculated) daily interest rate for the account (credit or debit) and enter accrued interest into the daily columns. (30 seconds).
  • Pass the sheet to a checker and check the calculation. (20 seconds).
  • The next task happens once a month. Sum the daily-accrued interest and create a voucher to pass the interest onto the customer ledger. (1 minute).
  • Pass to a checker and check the sum of the daily interest calculations. (45 seconds).
  • Create a pair of interest posting vouchers and put the vouchers into a brought forward file. 1 day for deposit interest, 30 days for debit. (1 minute).

For borrowing accounts (let us assume 2m accounts of the 9m interest accounts), in the UK, the customer needs a pre-advice of any debit interest charges. So we have the additional monthly tasks of:

  • Type a letter detailing customer address, name, account number, interest incurred. (1 minute).
  • Get copies of the customer statements. (1 minute).
  • Check the letter against the customer statements. (30 seconds).
  • Return copies of customer statements. (1 minute).
  • Put letter in envelope and write address on envelope. (20 seconds).

We shall assume that the monthly processes are spread evenly across the month, which has 22 working days.

Using the above factors and an 8 hour day and 80% efficiency we see that we need 50,200 additional FTE’s.

Payment Processing

In our manual banking world, all payments are effected using vouchers; both inter branch within the bank and inter bank. This would cover those payments currently made using vouchers (cheques and giro credits) but also those payments that currently are electronic such as direct debits, direct credits and standing orders. In addition all other entries (such as charges and interest) would also be made via vouchers.

Let us assume that 10% of all payments are “house”, that is they are from one account in a branch to another account in the same branch. Earlier (see the section called “Entries and Daily Balance Calculations”) it was assumed that there were 7m non-cash entries per day, each of which is assumed to be one half of a credit and debit pair). Thus this number of entries corresponds to about 3.5m customer payments daily.

The primary aim of voucher processing is to get the payment voucher to the account owning branch. In our manual world this involves several levels of sorting of the voucher.

In the payment originating branch the vouchers need to be sorted into piles corresponding to;

  • Payment drawn on accounts in the branch (“House Items”).
  • Payments drawn on other branches of our bank.
  • Payments drawn on other banks; (one pile per bank).

Let us assume 3 seconds per voucher.

  • Each pile needs to be broken into batches (of 100 say). And the total value of the batch calculated and entered as a control total on a batch voucher. (5 minutes).
  • Check the control total calculation. (5 minutes).

In the central clearing area.

  • The batches from each payment originating branch (2,000 branches) need to be sorted into a big pile for each bank. (20 seconds per batch).
  • The total value of the batch totals needs to be calculated. (63k batches into 13 totals – say 1 hour for 10 people.)
  • The vouchers coming in from other banks are assumed to represent half the total payments made off our bank accounts. Consequently half the payments made in our bank are drawn on another bank’s accounts. We thus have half of the batches of vouchers in the centre originating from our branches and half from other banks. Let us assume we trust our own staff to have totalled the batches correctly but not other banks. Thus for the other banks’ batches we have to check the batch totals with the contents of the batches.
  • Recalculate the batch total. (5 minutes for each batch).
  • Recalculate the total of the batches per bank. (10 people 1 hour).

Each batch of vouchers need to be broken up and sorted into 2,000 piles corresponding to each branch in our bank’s network. This would probably be effected by a two level manual sorting with each voucher (6.3m vouchers) being handled twice for 5 seconds.

At this stage we now have 2,000 piles of vouchers in Central Clearing. These need to be broken down into batches of 100 with control totals for each batch calculated twice (Timings as before).

For each branch a control total for all of the batches is calculated. (Assume 10 people for 1 hour).

The batches are then sent to the branches and each branch has an average 81 incoming batches. The branch randomly takes 4 batches and recalculates the control totals. (5 minutes per batch).

Each branch then has to sort the contents of batches into account number order. It is assumed there are 4m accounts receiving entries spread over 2,000 branches so the average number of accounts receiving entries per branch is 2,000. Hence the incoming vouchers have to be sorted into 2,000 piles. This would be achieved by a two level sort with each voucher being handled 2 times for 5 seconds a time.

The same sorting process has to occur for the “House items”, i.e. those vouchers that never left the branch because they applied to accounts held at the branch.

Finally each pile of vouchers per account has to be physically moved to the customer’s ledger so that the entries and balance calculation” can start. (30 seconds per account, 4m accounts).

A calculation based on all the above, an 8 hour day and 80% efficiency means that a total of 20,600 FTE’s would be needed to process vouchers in our bank with this figure doubling at month end and quarter ends. Not all of these FTE’s would be additional as banks currently employ staff to deal with voucher processing, but we will assume this is negligible in comparison with the work in the manual scheme. What is not considered an additional cost is the transportation of vouchers to and from branches as van and bike runs today exist and it is assumed these could handle the extra volumes.

Standing Order and Brought Forward File Processing

We will assume our bank has 2m standing orders. The mechanism for manual standing order processing is assumed to be to maintain a set of manual standing order records and a record of brought forward files. For each standing order payment the next payment voucher is created and put into a brought forward file. (1 month’s time, 1 week’s time, tomorrow etc). Let us assume that, on average, each standing order generates a payment every two months. Let us assume that each standing order has to go through the following brought forward files.

  • 2 months or more away
  • 1 month away
  • 3 weeks away
  • 2 weeks away
  • 1 week away
  • 4 days away
  • 3 days away
  • 2 days away
  • tomorrow

When a standing order payment voucher is moved into the tomorrow brought forward file (or when a new standing order is set-up), the next voucher has to be created by the following steps.

  • Get the Standing order file for the account (1 minute)
  • Create the next payment voucher and put in the appropriate brought forward file (1 minute)
  • Put the Standing order file back (1 minute)

Additionally each standing order voucher is handled 9 times at an average of 3 seconds per handling as it moves through the brought forward files. There would be 2m standing orders generating 6 payments each per annum.

Using these assumptions the bank would require an extra 400 FTE’s assuming an 8 hour day and 80% efficiency.

Statement Production

To create customer statement a person has to:

  • Get the account ledger. (1 minute).
  • Type a page with the name, address, account number and for each transaction the date of the transaction, the transaction narrative, the value and the resulting balance with the customer ledger as the source. (10 minutes).
  • Independently check the statement against the ledger. (30 seconds).
  • Put the statement in an envelope and write the address on the envelope. (20 seconds).
  • Return the ledger to its home. (1 minute).

It is assumed that there are 500k statements per day using these factors it is calculated that an additional 33,400 FTE’s are needed for manual statement production (assuming an 8 hour day and 80% efficiency). This probably doubles at month end.