LIBOR class-action lawsuitThe lawsuit- Annie Bell Adams, et al. v. Bank of America, et al, 12 Civ. 7461--alleges that members of the rate-setting panel of the British Bankers' Association manipulated the LIBOR rate higher on or before January 2000 through at least February 2009 (referred to as the “Class Period” in the lawsuit) and "unjustly enriched themselves" as a result. The lawsuit alleges three separate LIBOR manipulations:
- “Throughout the Class Period, the LIBOR 6 month rates on the first business day of each month are, on average, more than two basis points higher than the average LIBOR 6 month rates throughout the Class Period.”
- “Additionally, from August, 2007 through February, 2009, the LIBOR 6 month rates on the first business day of each month are, on average, more than seven and one-half basis points higher than the average LIBOR 6 month rates.”
- “Finally, the LIBOR 6 month rates on the first business day of each month are, the great majority of the time, higher than the five-day running average of the LIBOR 6 month rate surrounding the first business day submissions throughout the Class Period."
What is LIBOR, anyway?LIBOR is essentially an opinion of a price of money produced from a trade group of banks. The definition provided by the British Bankers' Association says LIBOR is “The rate at which an individual Contributor Panel bank could borrow funds, were it to do so by asking for and then accepting inter-bank offers in reasonable market size, just prior to 11am London time.”
The methodology of how the LIBOR rate is collected makes it difficult for a single institution or even group of institutions to move the needle on the LIBOR value by much (if at all), either up or down. The final value of LIBOR used for ARMs is constructed by taking submissions from the contributor panel of banks, where the rates are ranked into four groups. The lowest and highest groups are discarded, and the remainder is averaged together to produce a value.For example, if the contributor panel was 20 banks, the five lowest and five highest quotes would be excluded and the rest of the values would be averaged together. As LIBOR pertains to the lawsuit, it's important to know that many of these kinds of ARM contracts generally specify the use of the LIBOR value available on the first business day of the month as published in The Wall Street Journal. As such, the value is presented on a one-day-delay basis; that is, a value created on the 31 st of the month is released on the first of the month, and it is the first of the month value used as the basis of adjusting interest rates. If you pick up a today's copy of the Wall Street Journal and look, you will see the value from yesterday, and it is noted by date as such.
We launched our own review of the LIBOR dataWhile there have been allegations of LIBOR manipulation during the worst of the financial crisis (2007, 2008) banks were arguably quoting LIBOR on the low end, since being able to lend and borrow at low rates is an indicator of financial strength. At the time, and with major banks and financial institutions failing, there was great suspicion about the health of many market players, and if an individual firm was looking to borrow and being presented with high-rate offers, it might appear as if they were less than solvent. As such, the bias of such quotes, if they were manipulated, would have tended to be generally lower than reality would dictate.
The allegations in the class-action lawsuit were rather different, since they included a period well before the financial crisis, and also alleged a higher bias to LIBOR throughout. These allegations all seemed like a bit of a stretch to us, so we thought we'd conduct some rudimentary analysis to determine the credibility, if any, of the claims in this lawsuit. Although we conducted a couple of reviews, we paid particular attention to the "five-day running average" in the third allegation listed above.
Lawsuit's analysis is unclear to usThe lawsuit is rather vague and thin in terms of providing clues as to how the plaintiff's analysis was put together. The allegations of a two-basis-point differential are all well and good, but the statement does not indicate how this figure was arrived upon. On a straight up, as-defined-in-the-lawsuit basis, it's not possible to compare a single value available on the first business day of any given month against the entirety of the Class Period (110 months) and find such a figure, so there must be some other means by which they did their evaluation. The allegation of a period where LIBOR on the first of the month was “seven and one-half basis points higher …” suffers from the same issue: higher than when, exactly? The middle of the month? Some rolling or moving average? Again, it is not clear what the allegation intends or the comparison used to arrive at such a figure. That said, we thought we could analyze at least a contrived version of the five-day average, although the definition of that period isn't completely clear, either. It's unknown if the “five-day running average of the LIBOR 6-month rate surrounding the first business day” uses a “two days before, first of month, and two days after” arrangement, “three before and two after,” “two before and three after” or other method. Since the allegation was that LIBOR was higher on the first day of the month, we chose to use an average of the five values leading up to the value available on the first business day of the month. If LIBOR was specifically manipulated on the first of the month, it would be expected that the days leading up to it would tend to be lower.
Since the analyses that produced the first two allegations weren't clear, we chose what we think is a fair method to try to determine if LIBOR on the first business day of the month was higher or lower than the value that came immediately before or immediately after it (value available on last business day of the prior month and second business day of the new month, respectively).
Our analysisWe conducted three main pieces of analysis for the Class Period (January 2000 through at least February 2009) to determine if the 6-month LIBOR was in fact higher on the first business day of each month:
- Analysis 1: We averaged together the five prior daily values of the 6-month LIBOR and compared them against the value that would have been available on the first business day of the month to determine if it was higher or lower than the five-day average. Although it doesn't strike us as entirely fair to compare a five-day average value against a single one, we wanted to closely investigate the third allegation mentioned above.
- Analysis 2: We also compared the individual value for the last business day of the month against the first business day of the month.
- Analysis 3: Finally, we examined the first business day of the month against the individual value on the second business day of the month.
Analysis 1Of those 110 months, and relative to the five-day average of the days before it, here's what we found:
- Higher: LIBOR was higher on the first of the month 57 percent of the time (63 months). In the months when the LIBOR was higher, it was higher by an average of 0.03339 percent.
- Lower: LIBOR was lower on the first of the month 43 percent of the time (47 months). In the months when the LIBOR was lower, it was lower by an average of 0.05193 percent.
Analysis 2When we compared the LIBOR value on the first business day of the month against the value immediately preceding it, the single value on the last business day of the month, we found:
- Higher: LIBOR was higher on the first business day 36 percent of the time (40 months). When the value was higher, it was higher by an average 0.25530 percent.
- Lower: LIBOR was lower on the first business day 46 percent of the time (51 months). When the value was lower, it was lower by an average 0.27020 percent.
- Same: LIBOR was the same on the first business day of the month about 17 percent of the time (19 months).
Analysis 3Finally, when we looked at the LIBOR value available on the first business day of the month compared to the value available on the second business day of the month, we found:
- Higher: LIBOR was higher on the first business day 45 percent of the time (50 months). When the value was higher, it was higher by an average 0.22650 percent.
- Lower: LIBOR was lower on the first business day 44 percent of the time (48 months).When the value was lower, it was lower by an average 0.22020 percent.
- Same: LIBOR was the same on the first business day of the month about 11 percent of the time (12 months).
And another thing…If LIBOR really was manipulated to produce a higher value for the first business day of the month, wouldn't it be higher than both the value before and the value after it on a regular basis? Of the 110 months, the LIBOR value available on the first business day of the month was:
- Higher than both the previous and next values: 12 percent of the time (13 months)
- Not higher than both the previous and next values: 88 percent of the time (97 months). This includes periods when the value on the first of the month was the same as a previous or next value.
The effect on borrowers is hit or missBefore it becomes lost in the conversation, it's also important to remember that most ARMs tied to the six-month LIBOR adjust only once every six months. In this way, and even in the cases where a borrower held a mortgage throughout the Class Period, they would have been exposed to only two rate changes per year, for a total of 20 possible rate changes--and that if the loan was a fully floating six-month ARM from the beginning. Even then, a given borrower might have hit a "lower/lower" pattern, "lower/higher" or "higher/higher" one, so all borrowers with these products might not have had the same experience.
For more on how ARMs work, including adjustments, rounding and more, read: “ARMs: Hows, Whos and Whys”
At the same time, if the analysis which is the basis of the complaint is proven to be true, what then? The suit alleges a two-basis-point differential during the Class Period (and more during a sub-period). However, our analysis found both higher and lower readings on average throughout.Even if we consistently found a two-basis-point increase on the first of the month, which we didn't, it wouldn't make much of a fiscal difference. Why? A two-basis-point differential in the interest rate (from 5.00 percent to 5.02 percent) on a $100,000 loan with a 30-year term would amount to $1.22 more per month in a given six-month period. According to the complaint, the “plaintiffs aver that the class exceeds more than 10,000 borrowers nationwide," and that's fine, but what might the actual damages to an individual borrower amount to, if they even exist? This is the question which needs to be answered.