On LIBOR, part 2

At the end of my last post, I posed the question of whether or not any possible attempts to manipulate the daily LIBOR quote had been successful.  To look at this question, I analyzed 3 years worth of  daily 3-month LIBOR data, from January 2005 thru December 2007, a time during which routine manipulation was allegedly being attempted.

Now, just to be clear, my analysis cannot be said to be definitive.  To determine if, on any given day, an attempted manipulation was successful, and to what degree, one would need to look at all of the individual submissions that comprised the average resulting in the published LIBOR rate that day,   determine which if any of those submissions should have been a different rate, establish what the submission should have been, and then recalculate the average using that new rate.  That is beyond the bounds of what I can do here.  Nor have I attempted to determine the effect of any possible collusion between two or more panel members, which would be even more complicated.  All I will attempt to do here is to look at daily LIBOR settings in the context of the rates before and after, and the longer term trends, both actual and expected given market conditions at the time, to determine if the published rate makes sense and seems reasonable or if it seems odd and counterintuitive.

In looking at the data, one interesting thing appears immediately.   For almost an entire year, from mid-September 2006 until mid-August 2007, the daily LIBOR quote barely moved at all.

From September 20, 2006 until August 7, 2007 the high rate was 5.3875% (on September 20, 2006) and the low rate was 5.33% (on March 5, 2007).  Why were rates so steady during this time period?  Primarily because the Fed was maintaining a steady rates policy, and in the absence of any credit crisis, LIBOR will tend to mirror Fed policy.  So, with rate policy on hold for nearly a year, this would seem like an ideal time period to look for manipulation of LIBOR.  If it was going on during this time period, one would expect to see seemingly arbitrary daily moves both up and down in the otherwise steady published rate.

In fact what we see is remarkable consistency.  Indeed, for two extended periods, from Dec 26, 2006 until February 27, 2007, and then again from May 10, 2007 until July 26, 2007, we saw quite literally not a single change in the daily LIBOR rate of 5.36%.  For these two periods the graph is a perfectly flat line.

It seems highly unlikely that any manipulation could have been taking place during these two flat-line periods.   Overall, from Sep 20, ’06 until August 7, ’07, on days in which LIBOR did change from the previous day, the average change was .0029%, or less than one-third of a basis point.  And there were only 5 days over the course of this entire  222 day period in which the change in LIBOR from the previous day was even greater than 1 bp, with the biggest single change being 1.8 bps.

Now, perhaps it might be interesting to investigate further into why the rate changed by more than 1 bp on the 5 days in which it did, with an eye towards whether the change was the result of the same bank or banks altering their LIBOR submission on those days, and whether those days coincided with e-mails from traders requesting a higher or lower submission.  But regardless of the outcome of such an investigation, we can say, I think, that whatever attempted manipulation may have gone on during this period, it was neither all that regularly successful nor, if it was successful, did it alter the rate all that significantly, averaging barely a quarter of a basis point.

As an aside, some of you may be wondering why I chose September 20 and August 7 as a starting/stopping point for this particular analysis.  I chose September 20 because that was the date on which the Fed first announced that it was holding rates steady after over a year of rate hikes, and I chose August 7 because it was in early August that Bear Stearns’ impending meltdown began to really effect the market, with the collapse of two of its sub-prime funds and the requested resignation of its president Warren Specter.  This graph shows how rates began to react first to the news of Bear Stearns in early August by spiking from what had been essentially a flat-line, and then to the 50 bp rate cut by the Fed on Sep 18, the first of many cuts that would result in the zero rate policy that we still have today.

What about periods prior to September 2006?  Can evidence of successful manipulation be found there?  Well, it is somewhat more difficult to analyze this period because the Fed had an active rate-hiking policy prior to September 2006, and so the LIBOR rate was changing pretty much on a daily basis in response to this rate hiking policy.  For example, see this graph for the period between August 9, 2005 and December 13, 2005, a period in which the Fed hiked rates by 25 bps on September 20, November 1, and again on December 13.

With the exception of two consecutive days, September 1 and 2 on which the rate oddly dropped first by 1.5 bps and then by 9.4 bps, the graph looks pretty much exactly as you would expect it to, slowly increasing incrementally every day to keep up with the anticipated (and ultimately realized) Fed actions of raising rates.  For example, between the November 1 and December 13 Fed announcement dates, on which the FED announced a 25 bp rate hike, the average daily change in LIBOR was .9 bp, with a max of 2 bps (both times over weekends, so accounting for 3 days worth of change) and the rate steadily moved from 4.2606% to 4.4913%, almost exactly the expected 25 bps.   This steady and consistent change is not indicative of any manipulation, or, if there was any, its effect was just fractions of a basis point.

That 9 bp drop on September 2 is certainly an interesting outlier.  What could have caused that?  Frankly I don’t know.   It is possible that it is the result of manipulation, although I doubt it because in order to get that much of a change, any manipulation would have to have been both coordinated with many other panel members and fairly egregious.  The absence of other equally odd outliers at other points suggests to me that something peculiar happened that particular day in the market causing the drop rather than being an example of regular manipulation.  But, again, I don’t really know the answer, so I could be wrong.

In any event, my look at the 3 years prior to 2008 does not suggest to me any obvious evidence of significant, successful alteration of LIBOR from what one would expect.  Which brings us to what I originally said needed to be the subject of a separate discussion:  possible manipulation of LIBOR during the crisis of 2008 in order to hide funding difficulties from the market.  That will be the subject of my next post.

On LIBOR, part 1

So I figured it was time to weigh in with a post about the LIBOR scandal.  I will say upfront that, while it is a scandal, I think that, because of the general bank-bashing atmosphere that prevails in certain media and political circles right now, it has been blown far out of all proportion.  Our own jnc has tended to focus, not unreasonably, on the implications that the scandal presents for the kind of culture that prevails in the finance industry.  While I don’t agree that it is nearly as indicative of a widespread culture of dishonesty and corruption as jnc seems to, I think it is at least a fair question.  I have no time, however, for the Taibbis and Simpsons of the world who are doing their utmost to portray this as some kind of huge and concerted rip-off of Main Street, municipalities, widows and orphans, or whoever else is their latest chosen victim. That is just a bunch of bunk.

There are actually two separate issues here, and they need to be discussed separately. The first is the claim that individual traders were routinely manipulating the daily LIBOR set, both higher and lower, to benefit their trading positions on any given day.  The second is that, during the crisis of ’08, some banks were reporting falsely low borrowing costs in order to keep their funding difficulties from becoming public information.  Let’s address the former first, and perhaps a little history would be useful.  

Prior to the advent of LIBOR swap contracts used many different floating rate indices such as Prime or T-Bills (which is of course manipulated by the Fed as a matter of course), and even at times an individual bank’s own lending rate, determined solely at it’s own discretion.   Then in the mid ’80’s the demon derivatives market invented LIBOR in order to standardize swap contracts by basing them all off of a single, reasonable and standard measure of a “premium” bank’s borrowing cost.   Banks liked this because, even if the published rate was not exactly their own borrowing cost, it was a fair enough approximation to make their contracts sensible. Spreads could be added to this standard for individual contracts to account for the varying credit quality of the contracting parties, but LIBOR would represent a baseline index. In the beginning there were actually competing publishers of daily LIBOR rates, each with their own methodology of calculating the rates.  There were different contributors as well as different averaging and rounding conventions, and so the “standard” wasn’t quite that.  But over time the British Bankers Association’s published BBA Libor became the dominant rate and eventually the market standard.

From the first, LIBOR was defined broadly and was never meant to be some exact rate that mirrored an already established market rate.    In the first place, it’s nature as an average of many different bank rates, with both high and low submissions being deliberately excluded, makes it nothing more than an approximation anyway.  Second, the rate that the BBA requested as submissions was always very subjective.  Originally the BBA simply asked each contributing bank’s opinion of  where an unnamed, generic “prime” bank might be able to borrow:  ““At what rate do you think interbank term deposits will be offered by one prime bank to another prime bank for a reasonable market size today at 11am?”. 

Later, in 1998, it changed it’s question to “At what rate could you borrow funds, were you to do so by asking for and then accepting inter-bank offers in a reasonable market size just prior to 11 am?” So now, rather than asking about a hypothetical bank’s borrowing costs, it was asking about the bank’s own borrowing costs.  But note the qualifying and indeterminate language.  What is “reasonable”?  What if you weren’t actually in the market “just prior” to 11 am?  What if you borrowed from three different banks at three slightly different rates? What if you borrowed at 10am, and then saw that the market sold off between then and 11 am? Do you submit your actual rate or what you guess it would have been an hour later? What if you weren’t a borrower at all, but were in fact lending to other banks on a given day?

The main point here is that, contrary to the perception being fostered by outraged pundits and opportunistic politicians, there is no single, objectively knowable “correct” rate to be submitted by any given bank.  Obviously there is not only room for a subjective interpretation and judgment on the part of the  individual submitters, such a judgement is often actually required.  There is some range of rates within which it is perfectly reasonable to have submitted, even of it is not an actual rate that you transacted.  

Now, if one is tasked with submitting the daily rate to the BBA, how does one make this judgement, and is it ethical to be influenced by the entreaties of traders who stand to benefit or lose from your judgement?  That is a fair question and perhaps a point worth debating.  But I actually think it is debatable.  If the nonexistent “correct” rate falls within a range of, say, 3 or 4 basis points, is it really wrong or unethical to take into consideration the financial well being of the company you work for when deciding whether to report the high or low end of the range?  I don’t think so, at least not necessarily.  And the fact that one rate within the band is chosen and not another does not mean that the rate has been criminally “manipulated” or that the rate has been “fraudulently” submitted. As long as the submitted rate is defensible on the merits, regardless of how the judgement was arrived at, I don’t see a problem.  Certainly, in any event, this is not some “criminal conspiracy” akin to the mob shaking people down.  

And in fact the whole process of calculating LIBOR takes all of this into account.  That is precisely one of the reasons that such a wide number of banks (18) are polled and averaged, and why the high and low submissions are excluded from the average. As the BBA itself says, “The decision to trim the bottom and top quartiles in the calculation was taken to exclude outliers from the final calculation. By doing this, it is out of the control of any individual panel contributor to influence the calculation and affect the bbalibor quote.” Which brings us to another, and I think more important question:  Regardless of whether routine attempts to influence the submissions are ethical, were they successful in changing the rate?

That will be the subject of a future post.

%d bloggers like this: