Funds Transfer Pricing

  

Published: 17th January 2012 by William Webster

What is FTP?

FTP is part of full transfer pricing. That is identifying how much things cost to produce and then charging for those costs. In this way the final price to the customer should contain the costs of production.

When banks and societies deal with customers there is a transfer of risk. This may or may not involve the transfer of money. (Transactions like derivatives and contingent commitments don’t involve money but do involve risk, this risk may also include future liquidity risk).

An example helps. If a bank or society makes a mortgage loan the following risks are incurred:

1. Credit risk
2. Interest rate risk
3. Liquidity risk
4. Basis risk
5. Prepayment risk

There are also costs:

1. Staff
2. Premises
3. Capital
4. Regulatory expenses

If fees and interest don't meet the costs losses arise. But what about risk? We know ignoring risk can temporarily lift profit. But all risks have prices.

With a fixed rate mortgage swaps are used to hedge interest rate risk and the swap cost forms part of the mortgage price. But what about liquidity risk?

Until recently this was overlooked and it is this cost of liquidity that FTP seeks to measure and then charge back to the product line.

Why do FTP?

First if you can't quantify liquidity risk how do you know that you are being paid sufficiently for it? Second profits can be inflated by under pricing liquidity risk and not hedging it. This has proven to be a disastrous long run strategy. Maturity transformation and contingent commitments contributed to the crisis. Evidence shows FTP particularly where senior management got involved improved survival rates.

For banks and societies FTP is not an option it's mandatory. 

How do you do FTP?

FTP isn't new but for many implementation is. The process is subjective, it's not rocket science and some common sense needs to be applied - done badly it is dangerous.

There are several FTP methods, which you use is up to you. It needs to fit what you do and it should be proportionate. Let's look at three approaches: 

 

1. Pooled funding

This is an old method. The funding pool (treasury) supplies liquidity to business units at the pool rate. This rate is normally a market rate or market derived rate. Typically this is Libor or Libor plus a margin that reflects the bank's cost of borrowing in the wholesale market. It's simple and transparent but there are disadvantages:

• Libor is not a true cost - wholesale unsecured markets don’t trade.
• Failure to accurately transfer liquidity transformation costs - a short term rate doesn't reflect the liquidity premium for long term money. Some business units are overcharged others are undercharged.
• Failure to transfer interest rate risk - term money should be priced using term interest rates, if it isn't interest rate risk can remain outside treasury.
• Attributes of assets are ignored - a loan with the borrower's right to extend imposes additional liquidity risk. Using a short term rate fails to take this into consideration.
• Attributes of liabilities are ignored - fund raising parts of the bank deposit money with treasury at Libor. Money from a three year retail deposit receives the same Libor rate as money from an internet instant access account. The retail deposit taking function is not encouraged to seek stable term funding.
• Basis risk is ignored - loans made by the bank not linked to Libor can lead to basis risks (for example Base rate - Libor) pool funding fails to capture the price of this risk.
• Buffer costs are ignored - banks now hold significant liquidity buffers, pool funding can ignore the associated cost or include an average cost that is unrepresentative of the buffer required for individual products.
• A single pool rate can therefore be misleading - firms invariably use a mixture of retail and wholesale money, wholesale costs can vary whether they are secured or unsecured.

Pooled funding is insufficiently granular for the risks that are now apparent.

2. Multiple pooled funding

This tries to overcome some of the problems associated with a single rate by matching the characteristics of assets and liabilities with a particular type of funding profile.

The simplest approach uses a short term rate (Libor) and long term rate (swap) which are applied to assets and liabilities by their approximate maturity. A spread can be added to Libor and the swap rate in order to more accurately reflect the bank's true cost of borrowing.

This approach is still relatively blunt and fails to assess the true cost of liquidity.

3. Matched maturity method

This method prices each transaction by way of reference to the funding cost for the specific term involved. The starting point is a yield curve related to the bank's funding costs.

For some banks that are active borrowers in capital markets this curve can be constructed from yields on the bank's traded senior unsecured debt. If there are insufficient data points a derived curve can be used. This uses a swap rate plus a credit spread. For example a bank with a single A credit rating can derive a single A credit curve related to financial institutions. The data for these curves can be obtained from information vendors like Bloomberg. A deposit or loan now has a base price determined by reference to the appropriate maturity on this curve. For example a five year loan has a base price from the bank's five year cost of funds. Let's look at the advantages and disadvantages:

• The base price is obtained using the bank's real funding cost - if the bank's debt is actively traded this is an observable yield and should give a good indication of the cost of borrowing additional money.
• If a derived curve and credit spread are used this is more subjective - the method is an approximation based on firms of a similar rating, we know that the same rating does not guarantee the same funding cost so the bank's real cost of borrowing may differ.
• Some firms do not access wholesale markets. They are entirely funded by retail money. In this case it is appropriate to build a funding cost curve that is based on the source of funds.
• Assets and liabilities have different characteristics and these affect risk - a five year amortising loan should be priced off its duration. For mortgages pre-payment risk and extension risks are much more complex.

There are also some questions you need to consider:

• Some assets are better than others - if you hold liquid government debt it can be used to collateralise borrowing, this is beneficial. Should the cost of funding pledgeable securities be the same as that for illiquid loans?
• Some liabilities are better than others - long term fixed rate retail deposits are less subject to withdrawal than short term internet accounts. How can you influence the area that raises liabilities to take this into consideration?
• If you lend money and link the rate to the Bank of England's Base rate (repo rate) but borrow at rates linked to Libor how do you factor the basis risk cost?

For most firms some type of matched maturity method either using market yields, derived yields or retail funding costs provides the base cost of funding.

Refinement - adding to the base cost

The "on the road" price of a car is always higher, likewise funding costs. What extras do you need to add and are they optional?

1. The buffer cost (contingent liquidity cost)

Firms hold significant buffer assets. In stressed conditions these can be converted into cash. Holding such assets is expensive and the cost needs to be charged to the product that creates the need to hold the buffer.

The details of the buffer requirement are in your liquidity adequacy assessment (ILSA/ILAA). It will cover:

• Retail funding outflows
• Wholesale funding outflows
• Derivative collateral calls
• Pipeline risks
• Liquidity guarantees and commitments

Each product creates its own requirement. Instant access retail money attracts a higher buffer than term retail deposits. Short dated swaps create a lower buffer than long dated swaps. Mortgages create pipeline risk and a commensurate buffer. The buffer requirement should therefore be product specific and you should be able to calculate the associated cost.

2. The basis risk cost

Business units have a choice of pricing products off different indices and need to be charged for cost of hedging the basis risk. If a mortgage is linked to base rate and the cost of swapping this to Libor is 75 basis points the business line should pay for the hedging costs. Failure to do so will mean that basis risk can be treated as a free lunch.

3. The spread for credit risk

This is outside FTP but within full transfer pricing. The credit spread should reflect at least the probability adjusted loss in default. Various market and historic models can be applied these include stress scenarios. It is the writer's opinion that this spread is governed by two main factors - the competitive market place and a subjective opinion about what is "fair value". Firms should be able to demonstrate that the credit spread charged on a product is appropriate, this goes beyond mirroring the price at which competitors are prepared to write business.

4. The capital charge

This is outside FTP but within full transfer pricing. Capital supports the business you write and the risks it incurs. The amount of capital used to support a product and the cost of that capital needs to be factored into the cost attributed to that product.

In practice

This means the transfer price of an administered rate loan is as follows:

• Base funding cost + Buffer asset cost + the cost of capital

The transfer price of a tracker rate loan is as follows:

• Base funding cost + Buffer asset cost + Basis cost + the cost of capital

The transfer price of a fixed rate loan is as follows:

• Base funding cost + Buffer asset cost + the cost of capital

Note that the buffer charge should, where appropriate, include pipeline risk and collateral support for derivatives. It is worth considering breaking down a product's buffer costs into the individual risk drivers.

Each product you have will have a different set of costs based on the risks that it creates and the costs associated with hedging those risks.

What does the FSA want?

The regulator wants firms to identify and appropriately charge for the risks incurred in doing business. This is not unreasonable and as a minimum this means a product should have a pricing or costing "template".

Furthermore FTP should be embedded in the business process. That means it must influence the way you conduct your business. This goes further than just acknowledging it is there, it includes:

• Governance - does the board understand and implement the principles of  FTP, what's the evidence?
• Who owns the policy
• What's in place? - A well documented explanation of how it works
• How is new product design and approval influenced by FTP?
• How transparent is the breakdown of what constitutes a product's price?
• Do you include all costs?
• Is there sufficient differentiation between products?
• How has FTP affected product pricing?

It's an area that's open to regulatory challenge and you need the evidence to back you up.

What's the role of treasury?

Treasury manages risk. This involves warehousing, offsetting and hedging. This is not free and treasury needs to be paid the economic price of managing that risk. Failure to do this will lead to transferring costs to treasury and profits to business units - a recipe for suboptimal decisions. This can get political and what's fair may require an independent assessment from risk management.

Are there pitfalls?

Two you should be aware of:

1. FTP can create mistrust. It's easy to see how FTP can be seen as a transfer of profit from one department to another. Those who are affected should understand why and how you do it. Transparency is a must.
2. FTP can become obsolete. Unidentified risks aren't factored in. Market prices change. Customer behaviour alters. Small errors are inevitable. Large ones can be costly. Has your funding cost altered? Do you have new products? Can you link your costs to the buffer held? When did you last review the assumptions? Do you use FTP in budgeting? Do the actual results get compared with the budgeted ones? What does that tell you?

 Summary

• For banks and societies FTP is not an option it's mandatory.
  
• Pooled funding is insufficiently granular for the risks that are now apparent.
• For most firms some type of matched maturity method either using market yields, derived yields or retail funding costs provides the base cost of funding.
• This needs to be adjusted to include the additional costs you have identified, for example the buffer cost.
• Each product will have a different set of costs based on the risks that it creates and the costs associated with hedging those risks.
• Embedding FTP is as important as calculating it. It's an area that's open to regulatory challenge and you need the evidence to back you up.

Q & A

Q. I understand FTP is mandatory but do I have to charge the full cost of all the risks in the product price after all the market place is very competitive?

A. I don't see why you can't offer a product at a price that is lower than the FTP. But it raises questions. Who authorises this and do they fully understand the issues? How much of this will you do and can you afford it? Why is your FTP rate higher than the price you charge? Are your FTP assumptions accurate? Are you in the wrong market? Should you hold back and reconsider your action or redesign your product?

Q. Most firms don’t match fund loans. They borrow short and lend long term. Suppose a five year loan is funded by a one year deposit. (It's all variable rate and there is no basis risk). Surely this is a cheaper way of funding despite a potentially higher buffer charge. Shouldn't the FTP be based on what we do rather than the five year cost of money?

A. This is exactly what banks did. They relied on being able to roll retail or wholesale deposits. It creates a risk that not only can your funding cost increase but you can't borrow. To hedge this risk you need to borrow term money, this is more expensive. Who is to say whether the premium the market charges you is too high?

The FTP cost of the loan should be based on the rate you must pay to borrow cash for the full term of the loan. If you then decide to borrow short the success or failure of your strategy will feed through in your net interest income. But you have at least priced and understood the risk from the onset. Furthermore if your long term funding cost is high or you can't access long term funding doesn't it tell you that you should rein in long term lending?

Q. We use a weighted average cost (WAC) of funds. Are you saying this isn't now appropriate?

A. Yes. Weighted average costs are convenient but create the wrong incentive. They overprice the cost of short term money and under price long term money. It makes business units want to increase asset duration. WAC is also insensitive to changes in your cost of funding. If your funding cost suddenly jumps the WAC will initially show little change and you will continue to price money as if it was cheap and plentiful.

Q. We don't have a treasury so isn't FTP a theoretical exercise.

A. On the contrary FTP is relevant to you. You borrow money and there is a fully costed price for that money. If you know what it is it takes the guess work out of margin calculations. It should also give you a better grip on what you expect product profitability to be and therefore your budgeting process should improve.

Q. You mention budgeting; can you explain how FTP fits in?

A. Yes. The budget tells you what you expect your income and expenses to be and based on the planned volume you can calculate your expected income. I believe the regulator calls this forward looking. Deviations tell you whether what you expected actually happened. Investigation means learning about your business and the assumptions. For example if you cost using the long term funding rate but in reality fund short your actual cost may be lower than your planned cost. This is a reward for risk.

Q. Is this back testing?

A. Back testing is about seeing whether your model (in this case FTP) works. Comparing the budgeted FTP to the actual FTP will show you how effective your estimates for the cost of funds have been. It gives you insight into how you can improve your model and the budgeting process. It also tells you more about the risks you have. I believe this is what the FSA means by back testing the FTP.

Q. Should we stress test FTP?

A. Yes. Stress testing is good practice.

Q. How do we stress test FTP?

A. I thought you might ask that. Stress testing would ask what happens to the FTP if the individual costs alter. For example if the funding spread increases or the buffer charge increases how does this affect the cost and therefore the price of a product? What is the effect on the budget? Are you in a position that you can do something to mitigate the affect? Can you pass the cost on? Can you reduce the business volume? Can you change your product mix? If not you have identified weaknesses in what you do.

Q. FTP sounds straight forward but when we try to do it it's difficult.

A. You are not alone. The FSA has sidestepped how you do FTP. I think this is sensible it's not a one size fits all approach and is unique to each firm. Putting FTP in place requires step-by-step improvements. The more you do the better you become. In order to allay regulatory concerns I believe your first stab must be a good one - refinements come later.

Q. I thought FTP was about the cost of money and not credit risks and equity costs. Am I missing something?

A. I agree with you. FTP is about the cost of money. The credit and equity charges come in with full transfer pricing. There isn't much point in arguing with the regulator. If they want full transfer pricing and happen to call it FTP so be it.

Q. Does FTP mean writing out more dealing tickets?

A. That depends on how your business works. If you have a treasury that does arms length transactions with business units you will already write out tickets for those deals. Only now the rates on those tickets will change. If your business doesn't do this it's about working out product costs in a formal manner and then using that information in your accounting system.

Q. Should the rates we pay to access the central bank and covered bond markets apply in FTP?

A. This is an issue for larger firms. I think that using rates you pay the central bank are inappropriate. Repo rates have been kept very low for a reason and this offers banks a funding arbitrage. If you use the repo rate are you not implying continual ongoing central bank support? You could argue funding rates from collateralised or secured programmes are applicable to FTP provided you fully cost the provision of collateral and any affect on the cost of other funding programmes.

Q. Should we add the FSCS levy?

A. In the sense that the levy is a cost on retail funds, yes.

Q. A lot of discussion focuses on pricing asset funding costs. Does FTP apply to liabilities?

A. Yes. Liabilities are products. They contain risks and those risks need to be priced. You need to be careful you don’t double count. For example if you hold a 20% buffer for retail funding how is that cost apportioned? Is the cost priced such that you pay a lower rate on the deposit or charge a higher rate on the loan it supports? This goes to the heart of risk - which side of the balance sheet creates it? This is why FTP is subjective.

Displaying 1 to 4 of 4 results in total.