Tag Archives: housing

Bank funding costs

This post returns to the theme of interest rates on Australian mortgages. The first post showed the extent of the increases in mortgage rates over and above the Reserve Bank cash rate. The rationale banks have been giving for these increases is that their own funding costs have been continuing to blow out in the wake of the global financial crisis. In the spirit of occasional Stubborn Mule contributor @pfh007, it is time for some beer coaster calculations to see how plausible this argument is!

A number of commentators have accused the banks of out and out dishonesty on the subject of their funding costs. A few weeks ago in the Sydney Morning Herald, Ian Verrender focused on banks’ offshore borrowing and argued

if that really is the case, and only half their funding is sourced locally, then logically they should be raising interest rates by only half the rise in the official cash rate

Last week, also in the Herald, Richard Denniss built on this argument and argued that not only are offshore borrowings unaffected by Reserve Bank interest rate movements but so are all of their customer deposits. This led to the following conclusion.

Only about one-third of the banks’ funds come from the Australian money market, which means that when the official rate rises by 1 per cent, the banks’ costs only rise by about a third of 1 per cent.

But these simplistic arguments are incorrect. In saying this I am not defending the actions of the banks. There is no divine right for businesses to be able to preserve their margins at all times. Margin compression is a fact of life for many businesses. But more importantly, the ability Australian banks have to recoup costs from existing borrowers not just new borrowers is inherently anti-competitive.

Nevertheless, given the heat in the issue, it is worth getting a better understanding of exactly what is happening to bank funding costs.

A look at the balance sheet of any of the major Australian banks will show that their liabilities (which effectively represent the “funding” for their assets) are drawn from a range of sources. While the makeup will vary from bank to bank and change over time, roughly 50% of their funding is sourced from customer deposit balances and 50% from the wholesale markets (both domestic and offshore). Within those two categories, further distinctions can be made.

Wholesale Funding

Wholesale funding is a mix of short term “money market” borrowings and longer-term debt. Again, very roughly, about 50% of this wholesale funding is short-term (prior to the financial crisis, quite a bit more would have been short-term) and 50% long-term. Somewhat arbitrarily, “short term” tends to be defined as borrowings with a term of less than one year. Much of this borrowing takes the form of “certificates of deposit” (CDs) which are mostly bought by other banks or financial institutions like fund managers (much of the “cash” component of superannuation funds is invested in these sorts of instruments).

The interest rate paid on these deposits depends on the term and will be closely related to what the Reserve Bank does with its cash rate. For example, since the Reserve Bank just raised rates to 4.75% and it is almost a month until the next rate decision, the 30 day rate on CDs is currently very close to 4.75%. When the Reserve Bank hiked last week, markets were caught by surprise and the CD rate, which had been 0.20% lower jumped up in response to the central bank’s move. The correlation between these short-term borrowing rates and the Reserve Bank’s cash rate is not perfect, but on average over time, they are quite closely linked. So, the cost of this component of the banks funding can be expected to move in line with the cash rate, but should not increase significantly more than the cash rate.

Things are a bit different when it comes to long-term debt. For a start, most bonds are fixed rate: the interest the bank pays investors does not change even if the Reserve Bank cash rate goes up or down. However, while a fixed rate may suit investors, most of the bank’s assets have variable rates. Banks deal with this mismatch by using interest rate swaps (and other derivatives) which effectively convert their fixed rate borrowing into floating rate borrowing. The diagram below gives a simplified version of the mechanics of an interest rate swap. The bank enters into a contract with another party (typically another bank) to receive a fixed stream of interest payments in return for paying a variable or “floating” rate of interest. The floating rate is reset periodically, usually quarterly or semi-annually, with reference to a published rate which tracks short-term bank borrowing costs. The swap is set up to ensure that the fixed rate payments it receives match the payments it has to make on the bond. In this way, the fixed rate the bank pays on the bond is effectively turned into a variable rate from the bank’s perspective.Swap Diagram

Interest Rate Swap

This starts to make the cost of long term borrowing look a lot like the cost of short-term borrowing, but there is another factor: credit risk. If an investor buys a 5 year bond issued by, say, ANZ then it runs the risk that ANZ will collapse some time over the next five years. As compensation for this risk, the investor will demand an extra “premium” on the interest rate. This premium, also known as the “credit spread” or “credit margin” was fairly small before the global financial crisis, but shot up when investors suddenly realised that banks were not so safe after all.

Fortunately for banks (unlike their poor customers), they only had to pay the higher margin on new bonds. Even today, banks would still be paying off bonds issued before the crisis which have very low margins compared to the new bonds they are issuing. The average term of bonds issued by banks is around 3 years and the chart below shows how credit spreads have behaved over the last 12 years* along with a 3 year rolling average which gives a reasonable indication of the overall credit spread Australian banks are paying.

Financials Spreads

Credit Spreads for Financial Institutions (1998-2010)

The first thing to notice is that, although credit spreads have reduced since the peak of the financial crisis, the rolling average effect means that the effective cost of wholesale funds is still going up. Having old, cheap bonds maturing is adding to their cost of funds more than the fall in current spreads is saving them. On this point, at least, it would appear that banks are telling the truth!

But what about all of their borrowing outside Australia? Contrary to Verrender’s argument, Australian banks are not getting huge benefits by borrowing in countries with lower interest rates. Anyone with memories long enough to recall the notorious Swiss franc loans taken out by farmers and other small businesses in Australia in the late 1980s would appreciate that low interest rates do not count for much if the Australian dollar drops, thereby pushing up the amount of money you owe. Banks have no interest in running this sort of currency risk and so, much like their interest rate risk, they use swaps to hedge themselves. A “cross-currency swap” can be understood with a very similar diagram to the one above. Simply replace “Fixed” with, say, “US$ interest” and “Floating” with “A$ interest” and you have the picture for a cross-currency swap. This means that hedging is not a matter of paying some sort of small insurance fee, rather it effectively converts foreign interest rates to Australian interest rates. Even though perhaps half of the term funding raised by Australian banks is sourced offshore, it may as well be raised locally as far as the costs are concerned.

But how much is this increase in spreads costing the banks? As mentioned above, long term wholesale funding provides about half the wholesale funding for Australian banks, which is in turn about half of their total funding. So, a back-of-the-envelope estimate can be made by taking 25% of the 3 year rolling average. While I am at it, I will also project the rolling average forwards, assuming that credit spreads stay where they are today.

Low Funding EstimatesEstimated Impact of Term Spreads on Bank Funding Costs

This suggests that banks will see their funding costs continue to rise for about another year, but the overall impact of elevated costs in wholesale markets is only about a 0.45% increase. Compare this to what has been happening to mortgage rates.

Mortgage Spreads from 1998

Australian Mortgage Spread to the Cash Rate 1998-2010

The increase in mortgage rates over and above the cash rate has been about 1.2%, which is a lot more than 0.45%. So, while it may be true that wholesale funding costs are still increasing, it would appear that banks have already charged home buyers far more than the increase in costs the banks have suffered.

There is another source of costs for the banks that we need to consider: customer deposits. As wholesale funding costs rose during the financial crisis, banks began to compete aggressively for customer deposits as a (somewhat) cheaper alternative to wholesale funds. So, it is only fair to take the cost of customer deposits into account as well.

Customer Deposits

It is certainly true that on some of the customer deposits there is little or no interest paid, but there are also customer deposits which, particularly in recent years, pay very decent rates of interest. These include corporate deposits: imagine if a large mining company were to deposit a lazy $100 million into their account with one of the majors and was offered no interest…how long would it take for that money to move to another bank prepared to pay something very close to wholesale funding rates? Not long.

On this basis, we can reasonably assume that the cost of raising at least a portion of the banks’ customer deposits has risen as much as the increase in wholesale funding costs. To be generous, I will assume that all of their customer deposits have experienced this cost increase (although there are, of course, still plenty of low interest deposit balances out there…have a look at your own savings interest rates). Based on this assumption, I have recalculated the estimates of the increase in bank funding costs (i.e. taking 75% of the rolling average increase in wholesale spreads).

Spread Impact (High)

Estimated Impact of Wholesale and Customer Spreads on Bank Funding Costs

This revised estimate gets to an increased cost for banks of 1.3% which, given that the calculation is definitely too generous on the customer deposits side, is reasonably comparable to the increases passed through to mortgages.

However, the increases passed through to other types of loans (small business, credit cards, corporate loans, etc.) have been even bigger than those passed through to mortgages. So the only conclusion that can be drawn from this beer coaster is that:

  1. The banks are not lying when they say their margins are still increasing, but
  2. They have already gone beyond recouping these increased costs from their customers.

* Data source: Merrill Lynch. This data is the average asset swap spread across the financials sector and includes non-bank financial institutions and thus the spreads for the Australian major banks would, if anything, be slightly lower. I have now also got hold of data on some individual bonds issued by the majors and I will also analyse that to confirm it fits the same pattern.