Selecting the Right Treasury System - Part 3: DIY Internal Invoice Management
Out of all the treasury processes, FX multi-lateral netting of internal invoices is typically the biggest and most easily achieved return on investment. It is somewhat more complicated to manage than a cash pooling structure, but offers more in terms of hard savings and soft benefits. The critical outcome of this process is to get an affirmative answer to the following question: “To the extent that our tax structures allow, are we executing all of our inter-company invoices and payments internally?”
The need to implement an FX netting programme requires a few basic elements to be in place. The first, and most obvious, is that there has to be a significant amount of multi-currency inter-company invoicing and payments occurring. How much is significant? It depends on the size of the invoices, but 100 a month with an average size of around $40,000 would certainly qualify. The second, and most common barrier is the culture or organizational emphasis of the company. Some companies have such an emphasis on local operational independence that they operate more as a holding company than an integrated corporation. One company that is very often put forward as an example of operational efficiency settles literally billions of dollars a year in inter-company invoicing without any sort of FX netting. This company places such a high emphasis on its business heads being 100 per cent responsible for delivering their results that the company simply chooses to forgo most of the benefits of a central FX function.
Many corporations today have grown well beyond the founding company’s core market or geography. This successful growth, whether organic or done through acquisition, usually creates a complex corporate organization. It is little wonder that some treasury processes (at companies with many different products and locations) simply don’t get a chance to be implemented.
There are many costs or forgone benefits of not netting inter-company invoices and payments.
Wire charges vary, but a conservative estimate of the cost of sending a wire is about $10. If you are sending only 100 wires a month, then the $12,000 annual cost is relatively insignificant. There are companies, however, with annual sales in the $2bn range that are doing around 3,500 inter-company wires per month. At $10 per wire this is $35,000 per month or $420,000 per year.
The cost of this is real and theoretically visible. However, since the cost of wires is frequently bundled into the bank fees bucket during budgeting it gets rolled into the ongoing cost of the business bucket and essentially escapes scrutiny (except, perhaps, to try to negotiate the cost of each wire down).
The spread between bid and ask (buy or sell) for a market sized deal, of $5m is usually four pips (hundredths of a cent) minimum. This means that asking a bank to pay a GBP inter-company invoice of the USD equivalent of $5m would amount to about $1,000 in FX conversion costs. (This makes the $10 cost of a wire seem quite small.)
Most companies, however, have very few $5m inter-company invoices to settle. More often, they have hundreds or even thousands of much smaller invoices to settle. The bid / ask spread from a bank on these smaller payments (known in the industry as nuisance payments) ranges from 50 to 200 pips. If a company were to settle 100 wires at the equivalent of $50,000 per wire, it would cost between $13,000 and $52,000 in currency conversion costs and $1,000 in wire costs. Generally speaking the smaller the size of the average wire, the more it costs to convert it to the desired currency (ie 1,000 payments of $5,000 each would almost certainly cost $50,000 plus $10,000 for the wires)
Avoiding the FX spread can be a big cost saving. In the above example, it equates to a one per cent drag on the profitability of the company. There is, however, seldom a slot in the budgetor executive compensation plan that puts a value or measure on FX spread costs. This cost therefore is a hidden cost.
There are many ancillary benefits to an FX netting programme. When companies run inter-company netting programmes they can set common business terms for the netting cycle. This has the very practical benefit of shortening payment cycles while at the same time improving payment predictability. Both of these effects help in terms of forecast reliability and reducing working capital. The third benefit is the reduction in timing risks of currency hedging. With a netting programme companies are forced to net and settle on a designated day of the month. This means that the currency risk manager can hedge to and settle on a specific date like the 15th of the month. This specific date match is much less risky than hedging to the middle of the month for a set of payments that happen throughout the month.
All of the big global banks such as Citibank, JPMorgan Chase, Bank of America, ABN Amro etc. offer netting services. These banks provide corporations with software to consolidate and net internal invoices. Operationally, the subsidiaries enter their invoices over the Internet into the bank’s system. The bank’s system then nets the invoices and presents a series of deals to be executed.
There are three primary costs in asking your bank to act as your netting provider:
This is a common way for the bank to recoup its costs for running the programme. This charge varies from customer to customer. The cost ranges from zero (it’s part of the relationship!) to $100,000 a year.
As this is a classic system-driven process, software costs are inevitable.
Most, if not all, netting arrangements with banks require that the subsequent residual currency transactions be done with said bank. In theory, this leads to larger FX spreads. In practice, even if the netting cycle is managed using your own systems, the set of currency transactions that result from the netting programme are not always competitively bid.
One of the most common ways for a bank to charge for the netting service is to apply an invoice fee. Once again, this fee varies widely, but it is often in the $10-20 per invoice submitted range. It is a nice, and you might say, painless, way of charging for the service. $10-20 is, after all, about what a company would have had to pay for a wire had they chosen to forgo the netting process altogether.
Part one in this series of articles started with the fact that 40 per cent of the Fortune 1000 does not own a treasury system. Obviously, that means that 60 per cent do own one.
FX multi-lateral netting is a treasury process that depends heavily on software. The basic technology elements that are needed are inter / intra net access for invoice submission and a central system to manage and consolidate all of the requests.
In the more advanced systems, there can be an extra layer of sophistication added that allows for online invoice disputing etc.
Once again this tends to be the most economical solution, but it is also the most difficult to get budget for. It requires, therefore, a strong focus on managing processes centrally and doing / controlling the process yourself.
Cost for FX netting solutions from Software vendors varies somewhat, but expect to pay a license cost of between $30,000 and $70,000 when purchasing it as a module of a larger system. Stand-alone FX systems are often sold on an annual rental basis for approximately the same amount.
People are needed to operate systems. In the case of FX netting the main operator is typically the currency risk manager. The currency risk manager’s job almost always becomes less onerous when an FX Netting system is installed.
It is not uncommon for a central treasury to fund its operations through charge backs. FX netting provides a very clear and simple method for doing this. Invoice charging banks do it and shared service centres often adopt this model. Costs vary, but it is an internal cost.