The Credit Crisis: A Look at LIBOR

For several weeks, I have been promising to take a closer look at the financial crisis, the symptoms and causes, and the effect of the EESA/TARP on the crisis. As I began to do my research and create an outline, I realized that this is too much for one article. So, over the coming days and weeks, I will be writing a series of articles which take a look at the crisis and the impact of the EESA/TARP. This is the first article and it's focus is LIBOR. (It is still long...)

Until a few months ago, I don't think I had ever even heard of LIBOR. But, to understand the financial crisis, you need to understand LIBOR. So, what is it? It is an acronym for "London Interbank Offered Rate" and it is set by the British Bankers Association (BBA). It is an important interest rate for a few reasons. First, LIBOR is the benchmark rate for unsecured (no collateral) lending between banks. Second, many other interest rates and financial agreements use LIBOR as a reference rate. For example, some adjustable rate mortgages reset rates based on LIBOR.

As we've discussed here before, fractional-reserve banking requires banks to hold only a small amount of cash in their reserves. The rest of the cash can be lent. The constant movement of money requires banks to monitor their reserve requirements - remember, (most) banks (usually) want to keep their reserves at the smallest possible level so that they can maximize the amount of money they lend. This places banks in a situation at the end of each day where they either need additional money or have excess that they can lend.

Here's a simple example. ABC Bank has $10M of deposits from their customers. Assuming a 10% reserve requirement, they need to have $1M on reserve. On a given day, let's say that their customers withdraw $100k. Now, they will only have $900k in reserves against $9.9M in deposits which would have a requirement of $990k in reserves. So, ABC Bank needs $90k in cash to cover their reserve requirement.

At the same time XYZ Bank also has $10M of deposits and $1M on reserve. But, let's say on this same day, they actually have an increase in deposits of $100k. That leaves them with $10.1M in deposits, a reserve requirement of $1.01M, and $1.1M on hand. So, they have an extra $90k. This $90k can then be lent from XYZ bank to ABC bank. These are typically very short term loans (sometimes just overnight) which all banks to stay liquid (keep lending money) while maintaining reserve requirements.

LIBOR is the basically the prevailing interest rate for the these interbank loans which are originated on a given day. LIBOR is set each day in London. Banks submit details of the interest rates being used in the interbank lending market to the BBA who then calculates the average interest rate. LIBOR is calculated for ten different currencies and various loan maturities ranging from overnight to one year. We are interested in the rates for loans in US Dollars with overnight, one month and three month maturities.

You've probably heard a lot of discussion that the credit crisis is due, in part, to banks not lending to each other. If banks don't lend to each other, then they have to be more careful managing their reserve requirements. This implies that they take less risk, hold larger reserves, and, thus, don't lend as much to consumers or businesses either. The following graph shows the history of LIBOR (US Dollars, 1 Month) since its inception in 1986.

Source for LIBOR data in this article: BBA, Data here (it was a pain to compile!)

The first thing to recognize from the recent crisis is that there was a large jump in the rate which has been followed by a steep decline in the rate. I also graphed and analyzed the difference in the rate from one day to the next since 1986 (not shown). The interesting thing is that rates are usually remarkably stable. The average daily change in the rate is only 0.12 basis points (a basis point is 1/100th of a percentage point). Even more interesting is that the largest changes in the rate from one day to the next have almost been exclusively at the end of November or the end of December in any given year. I'm not exactly sure why this is, but I would guess that it has something to do with the need for lending increasing at the close of the calendar year for reporting or tax purposes. Outside of those jumps, there have only been 26 days in the last 22 years when there was a one day shift of more than 25 basis points.

As you can see from the above chart, LIBOR is relatively stable. Most of the volatility occurs in November and December (as mentioned above). However, September and October of this year saw seven days with 25+ basis point shifts. We haven't had a period of that much volatility since the summer of 1998 as the Asia was reeling from the effects of its own financial crisis and Long Term Capital Management was in the process of imploding. You should also notice that after the post-September 11 volatility in 2001, there was not a single day where LIBOR moved more than 25 basis points from the previous day until September 2007 when the subprime mortgage mess was unfolding.

It is the recent volatility of LIBOR and not the actual rates which indicate the presence of a problem. It is also important to understand that rates are relative. An interest rate of 5% would seem astronomical today, but this is approximately the historical average for 1 month LIBOR. One way to gauge this is to compare on rate to another rate. This is done by simply subtracting one rate from another and expressing that difference (called a spread) in basis points. So, if you compare 3 month LIBOR and 1 month LIBOR you would expect to see little difference in the rates. The graph below shows the history.

There have been a few periods of volatility in this spread. Notice that the spread recently is at high levels (roughly +50 basis points). There was a shock towards the end of 1999 and some very high spreads (averaging almost +100 basis points) in mid- to late-October of 1987 when the stock market crashed. The other times with very high levels generally correspond to times when 1 month LIBOR was volatile from one day to the next as described above (mostly in November and December). Generally, the spread is positive - 80% of the time it is non-negative. It tends to be negative in December. It was negative for most of 2004.

So, what factors drive this spread? First of all, let's think about the purpose of interest rates. If you are going to lend someone money, there are two key factors that are involved: the value of money and the risk of the loan. What do I mean by the value of money? Money represents an opportunity to spend it on something. If money is lent, it obviously cannot be used by the lender. There is an opportunity cost to this which is equal to the value which could be created by other investments with the money. Risk is a bit more obvious. If the lender deems the loan to be risky, they will ask for a higher interest rate to hedge against default.

Generally speaking, it should make sense for interest rates on a 3 month loan to be slightly higher than rates on a 1 month loan. There is more time for an event to occur which increases the risk of a default by the borrower or an increased need for cash by the lender. Also, for most of the last 100 years, the U.S. has experienced inflation. As money loses value (due to inflation), more money is required to be paid back to be cover the original value. The longer the period of the loan and the higher the expectations, the larger the spread will be. If you graph the interest rates against different loan maturities, you get what is called the yield curve.

When the yield curve "flips", it means that interest rates are actually higher for shorter-term loans. This can occur when things are not "normal". If you expect money to gain value (due to deflation), you would accept lower rates on longer-term loans. If you expect rates to decrease, due to either macroeconomic factors or central bank policy, you would also consider lower rates on longer-term loans.

As mentioned earlier, there has been a positive spread (3 month higher than 1 month) 80% of the time. The three periods of time where the spread has been very high were after the market crash in 1987, late 1999, and the recent crisis. All three of these periods correspond to high-risk environments (I'm assuming that the high spread in 1999 was due to Y2K fear). In 2004, the spread was negative for much of the year as interest rates were falling all year, and banks were willing to bet on that.

One more graph... In 2001, an overnight rate for LIBOR was introduced. This graph will look at the spread between 3 month and 1 month as well as the spread between 1 month and overnight.

Notice here that the spread between 1 month and overnight (1m-o/n) is more volatile than the 3m-1m spread that we already looked at. There were shocks after 9/11 and then off-and-on for the last year or so. In the weeks leading up to the week of September 15, 1m-o/n hovered around 30 basis points. Then Lehman Brothers announced they were going bankrupt and Merrill Lynch was in trouble. Overnight LIBOR jumped 96 basis points and 1m-o/n went negative. It got worse the next day when overnight hit 6.4375% (a 333 basis point increase); 1 month remained somewhat stable moving the 1m-o/n spread to -369 basis points. This was the day that AIG failed. By the end of the week, Paulson and Bush announced they had a plan, overnight LIBOR had settled to 3.25% with a -6 basis point spread on 1m-o/n.

The next week, 1 month and 3 month LIBOR grew while overnight fell placing 1m-o/n spreads north of +100 basis points by the end of the week. On September 29, the original EESA plan failed to pass in the House. The next day, overnight again shot up, this time to 6.875%. It settled back down and positive spreads returned until October 8 after Iceland's economy collapsed, stock markets around the world had large losses, and AIG received more funding - all this despite rate cuts by central banks around the world. Overnight again spiked for two days until the beginning of a sharp decline which has continued (with a few bumps) until rates stayed under 0.2% for the entire week of December 8 (data from the BBA is only available through December 12 at this point). This has left 1m-o/n spreads between 100 and 150 basis points for nearly two months.

It is safe to say that the interbank lending market has been a bit crazy lately which is at least a symptom of the market turmoil. There are some reports out there which have questioned the relevance and accuracy of LIBOR in the last year or so. It is argued that the increased role which central banks have played in lending has diminished the need for interbank lending. This leads to a less efficient market and less accurate LIBOR quotes. The BBA has disputed this. Further, a recent article by the Wall Street Journal discusses how the BBA has clarified that loans which are secured by the government cannot be used in defining LIBOR which is intended to represent unsecured lending rates. The BBA has another announcement here discussing the steps they are taking to add more transparency and trust in the process.

Ok. I know that is a lot of information. I hope it is helpful in understanding this important component of the financial system. The next article in this series will take a look at the FED and the famous FED funds rate.