Manage your bank, not your model

Need-to-know basis

As a kid I enjoyed many TV shows, books and films.  One of my favorite themes was espionage - James Bond in particular.  Add a splash of science fiction and I would be hooked (e.g. The X-files) Often buried in the plot lines of intrigue, investigation, and government cover-up was a situation where some characters were only provided information on a “need-to-know” basis.  I knew when I heard a line like “need-to-know” there was some mystery or jaw-dropping plot twist that I was likely to find out about.  “Need-to-know” in this context implies secrecy or high security.

A different way of looking at “need-to-know” is to think about our “need-to-know” certain information in order to perform our job or even go about our daily lives.  For instance I don’t need to know how an internal combustion engine works in order to drive my car.  In our professional lives, albeit much less glamorous than 007, it seems to me that we all deal with some information that is “need-to-know”.  It’s not likely to be the kind of information that’s particularly intriguing or sexy.  It’s more likely to be data, terminology, or phrases that amount to specialized “shop-talk”.  The details or specifics about certain policies and procedures are really only of interest to those whose job it is to “need-to-know” such things. 

For me no where is this more evident than when I accompany my wife to a parent-teacher conference at my daughter’s school.  My wife is also a teacher.  It’s not uncommon for these “conferences” to dissolve into a discussion between just my wife and my daughter’s teacher that cover a dizzying array of teaching strategies, student-progress evaluation methods, and performance standards.  It’s not necessarily complicated stuff, I just don’t “need-to-know” it.  I just want to know if my daughter is doing ok in class or not.

Generally for most of us this type of thing is not a problem.  It’s a sign that certain jobs and functions are compartmentalized and it often means that highly specific niches have developed.  The downside is that things get much more complicated if specialized information like this must be communicated outside the originally intended group.  This new group has it’s own set of biases and perspectives and may interpret the specialized information in a different way (or worse, incorrectly).

I can’t help but think about the stress-testing we do for our community bank clients and who “needs-to-know” the results.  It used to be that it was only senior management and ALCO.  Then it was also the examiners…then the Board.  Sarbanes/Oxley (SOX) expanded the circle to include accountants and auditors.  The FASB (and other groups) have added investors and shareholders to the mix.  Finally the circle has expanded to include attorneys.  I was recently asked by one of our publically traded clients to explain our IRR and Fair Value modeling processes to their attorneys.  Imagine an attorney that looks like a deer in headlights as he’s desperately trying to grasp unfamiliar modeling terms like duration, discount rate construction, economic value of equity, rate shocks, rate ramps, and twists.  (I have to admit it was childishly satisfying to render a lawyer speechless, if only for a minute.)

In October 2010 CSBS published a white paper entitled, “The Case for Stress Testing at Community Banks”.  Towards the end of the document they address several issues related to widespread use of stress-testing.  Here’s one in particular:

Should stress testing results be publically available?
Public disclosure was one of the elements of the SCAP program which proved successful. As Chairman Bernanke noted in his May 6, 2010 speech, public disclosure enhanced the credibility of the stress test. The disclosure provided transparency of the participating companies and the amount of capital required, if any. While the SCAP was very different in purpose and was conducted in an abnormal economic environment than any routine, general purpose stress test, an assessment must be made if and what type of data will be publicly available. While there is a legitimate case to be made that the results should be protected as supervisory information, there will be ongoing pressure to disclose results by analysts and market participants. There was a time when the bank’s call report and Uniform Bank Performance Report (UBPR) were considered confidential information.

Over the past year or so I’ve spent a considerable amount of time pondering this question.  I’ve decided that my answer here is NO, stress-testing results should not be publically available, at least as far as community banks are concerned. 

Seldom are there clear right or wrong “answers” produced through stress-testing.  It’s not like an engineering stress-test where you find the breaking point of some material or structure.  It’s more like a medical stress-test that tells you about the likelihood of a stroke or heart-attack, but there are no guarantees.  However, the medical stress-test analogy can only take us so far.  There is widespread agreement among medical professionals about what certain risk factors mean.  This is only because there is a wealth of medical data available.  Countless test results can be analyzed and correlated.  Conclusions can be drawn and often times even verified.

We don’t have this wealth of stress-testing data for banks.  And a large part of what we do have is based on subjective assumptions (take core deposit decay assumptions for example).  The bank stress-testing that was performed as part of SCAP and the stress-testing we run for our community bank clients uses far fewer “scientific” facts than most people realize.  There are many industry professionals who have a deep understanding of modeling, accounting, and analysis that believe the tools used for stress-testing are not ready for prime-time (see one example here, “SFAS 157, Fair Value and Other Fairy Tales” | Calculated Risk – 8/30/2008).

I generally agree that community banks can benefit from stress-testing.  The process is not necessarily complicated but it can be extremely detailed and it involves making many assumptions.  There is a real danger that the information will be “dumbed-down” and ultimately misinterpreted if made widely available.  The general public still can’t make a meaningful distinction between a Wall Street bank and a community bank.  Why should we believe that they’ll correctly (or meaningfully) interpret stress-testing results?

The circle of who “needs-to-know” the details and results of the stress-tests has expanded quite a bit.  At this point though I don’t think the results are ready for public consumption and probably won’t be for some time to come.

Debate about the discount rate for loans – a liquidity adjustment?

We make no secret about how we calculate the discount rates used for fair value calculations in our model.  For loans it’s a pretty straightforward approach that combines a risk-free rate, a credit quality adjustment, and a cost (or overhead) adjustment.  Using this approach I think we’ve been doing a reasonable job of calculating fair values for the past 20 years.  There have been occasions when we’ve been asked to explain or defend our estimates, and on the whole, I think our approach has held up pretty well.

However, lately there’s been a trend of accountants and auditors telling us our fair value estimates are “wrong”.  The reason?  Apparently we’re not making an explicit “liquidity” adjustment.  Their basic question is always, “how much should the value be adjusted if these instruments needed to be sold or liquidated?”

This question has always puzzled me for a number of reasons.  First, hasn’t FASB made it clear that fair value is meant to represent value between a willing seller and buyer?  It is not supposed to represent fire sale values. 

Fair value of an asset (or liability): The amount at which that asset (or liability) could be bought (or incurred) or sold (or settled) in a current transaction between willing parties, that is, other than in a forced or liquidation sale.
FASB Concepts Statement No. 7 – Glossary of Terms. Issued 2/2000; Last Amended 2008.

Second, banks will typically make a loan and expect to hold the loan until it matures or is paid off.  I understand that many banks originate and sell mortgage loans - I’m not talking about those loans.  I’m talking about the huge volume of commercial loans (C&I, CRE, ADC, etc.) made by community banks all across the country.  Given this unique business model it’s hard to see how any “liquidity” adjustment would be anything other than a “fire sale” type of mark-down.

Third, loan pricing hasn’t changed.  The bank’s loan officers are still pricing new loans the same way they always have (hopefully they’ve increased their credit spreads lately).  I assume the bank is booking these loans at Par value.  If not, how is that working?  New loans are being booked at a loss?  It seems like the accounting theory for valuing the loans and the business practice for pricing the loans are not in alignment.

Finally there’s the question of how much should the adjustment be if you made one?  I don’t have an answer for this because, obviously, I don’t think making a liquidity adjustment is appropriate in the first place.  If forced to come up with one I’d start by researching what kind of liquidity “haircut” is charged by other organizations.

I asked our Senior Financial Analyst, Kurt Schneckenburger, to research this one.  He didn’t have to look far to dig something up.  Here is a list of haircuts for different types of collateral for the Chicago Mercantile Exchange (CME).  The purpose of the haircuts is to insure that the collateral backing an account is enough if it was forced to be liquated to cover any losses in the account.

For example, someone has an account at the CME and they deposit $1,000,000 dollars.  The first item shows there is no haircut on US Dollars (but, obviously, the account holder would earn no interest on their $1,000,0000 either.)  As an alternative the account holder could place on deposit a $1,000,000 T-bill.   If this was the account’s collateral, it would earn some interest (whatever the T-bill earned).  However, in order to protect the CME in the event of an account liquidation at the end of the day, given any volatility in the T-bill market, the exchange would demand a 0.5% haircut on the T-bill, meaning that it would take a T-bill deposit of $1,005,000 to equal the protection of $1,000,000 of cash.   If the account holder deposited a US T-bond with a maturity date 6 years in the future the CME would demand $1,045,000 to protect the same size of an account that $1,000,000 in cash would.

The size of the haircut is an attempt to insure that there would be enough money left in the collateral after accounting for any potential change in the collateral value that could occur during a single day.  In other words the size of the haircut is the maximum expected volatility for a day.  This seems to answer the basic question posed above, “how much should the value be adjusted if these instruments needed to be sold or liquidated?”

Some of the haircuts listed are 0.5%, 4.5%…11.0%, 30.0%.  These are forms of collateral that have active (and supposedly efficient) trading markets.

I’d hate to see the haircut we’d come-up with for “Main Street” Community Bank’s C&I loan portfolio.  I think many folks would find it painful.

Former FDIC Chief says FASB proposal is "irresponsible"

"FASB's proposal is incomprehensible and irresponsible," Isaac said. "There is a high risk that just proposing that loans be reported at market value can cause banks to tighten up even further on lending than where they are now. They will become extremely reluctant to lend to small and medium-sized businesses in an already weak economy that is dependent on these businesses for growth, especially in employment.

 

"The three members of the board who voted in favor of this proposal are like religious zealots worshipping at the altar of fair value. Not only is this proposal bad for the economy, it is bad accounting because it doesn't recognize the banking business model," Isaac said.

via www.accountingweb.com

This article and the related articles below provide some very good reading. It's incredible that FASB is still trying to push this through given the number of people with years of experience in the industry (private & regulatory) that are against it.

If it goes through, it will be really bad news for banking.

• Mark-to-market plan could have seismic impact.

• PWC observations on FASB fair value proposal.

• Fair value exposure draft alarms CFOs.

• FASB plans will increase risk, stunt growth.

• Former FDIC Chief: FASB proposal irresponsible.

FASB just doesn’t get it when it comes to banks

I was hoping to avoid any more rants about mark-to-market accounting, but it seems FASB is intent on ramming it down our throat.  Recently there have been several great articles and posts arguing against MTM for banks.  They all specifically mention the loan portfolio:

FASB Mark-to-Market Plan Could Have a Seismic Impact
American Banker  |  Thursday, May 27, 2010

'Fair Value' Proposal Could Cost Banks
WSJ | May 27, 2010

FASB's Big Bad Idea
Bank Lawyer's Blog | May 31, 2010

FASB Proposes Mark-to-Market For All 'Financial Instruments'
CSBS Examiner | June 4, 2010

But, it’s the first article listed above that caught my attention.  Right near the end of the article they finally mention core deposits:

Core deposit values would be based primarily on present value, and would not reflect intangible values to the extent typically considered by banks when they decide how much of a premium to pay for deposits in an acquisition.

The valuation of deposits has long been a sticking point in the debate over mark to market. Those who oppose applying the method to all loans have argued that it is unfair to use mark-to-market on only one side of the balance sheet. But arriving at fair value for deposits is no simple task; in the early 1990s the FASB determined that fair-valuing core deposits was unworkable, and left them out of a rule addressing accounting and reporting requirements for investments.

The solution the FASB has come up with for deposits is "more of a present value of future cash flow," Fisher said. "It's not full-fledged fair value. It's not the same thing as if the bank went out and got a market price for its core deposits. It's partway between the two. That's illogical to a lot of people, but it becomes more logical as you move to a fair value model" for other parts of the balance sheet.

FASB Mark-to-Market Plan Could Have a Seismic Impact  |  American Banker  |  Thursday, May 27, 2010

Thank goodness, I thought, someone is finally shedding some light on the problem with valuing core deposits.  I’ve already commented on this in a previous post, “Valuing non-maturity deposits is a huge road-block to implementing fair value accounting”  But then I saw this: “The solution the FASB has come up with for deposits is ‘more of a present value of future cash flow’…” - Huh?

I know what “present value of future cash-flow” means.  And of course you should, at least partially, base any fair calculation on the present value of future cash-flows.  But there’s a problem - what is the future cash flow of a checking or money market account?  There’s no contractual maturity date; the customer can withdraw money any time they please.  How do you determine the present value if you don’t know the term structure of the future cash-flows?

Because we model interest rate risk for banks we know there are various ways to solve this problem.  But there is no wide-spread industry-accepted way to model this (according to the article though, FASB has come up with a “solution”).  Believe me, bankers and regulators would love to have a workable standard here, but there’s still much debate.

Without agreeing on some standard way to model the term structure of the these deposits (a way that truly represents how valuable these deposits can be), there’s no hope of making bank financial statements “more transparent”.  Fair valuing core deposits without an acceptable standard will just leave us further lost in a fog.

Valuing non-maturity deposits is a huge road-block to implementing fair value accounting

Whether you are for or against full mark-to-market accounting for community banks, I think you’ll agree that it will have a big impact.  In the bank lobby’s tug-of-war with FASB you might say it’s a good thing when a regulator chimes-in in-favor of the banks’ position.  One Fed Governor has done just that:

Fed Seconds Fair-Value Plan Critique
American Banker, Tuesday, September 15, 2009 (subscription required)

WASHINGTON — Federal Reserve Board Gov. Elizabeth Duke sharply criticized key aspects of a pending accounting proposal that would expose a bank's assets and liabilities — including loans held to maturity and deposits — to fair-value treatment.

In a rebuke Monday to the Financial Accounting Standards Board, Duke echoed industry complaints that the approach fails to recognize how banks conduct operations and could cut off credit to small businesses.

Here is one of the main points made by Duke (as quoted in the article):

Duke appeared more concerned about financial institutions' ability to comply with the rule — "We have very little actual experience in fair-valuing liabilities," she said — and its impact on small banks…"Smaller banking companies likely will incur substantial costs and experience great difficulty in applying the new standards," Duke said. "But will financial statement users see any real benefit?"

While I applaud her observations, I think the point about “very little actual experience in fair-valuing liabilities” can’t emphasized enough. 

I don’t think the financial world-at-large truly understands the magnitude of the problem – especially when it comes to valuing non-maturity deposits.  A fundamental piece of any fair value analysis is the timing of the future cash flows.  What timing are you going to use for checking, savings, and money market accounts?  Those of us who make a living modeling interest rate risk know that there isn’t an easy straight-forward answer to this question and, so far, there surely is no standard we can all agree on.

To give you an idea of how big this issue is – a typical bank funds 30% to 50% of their assets with non-maturity deposits.  That’s a pretty big chuck of the balance sheet to be marking-to-market using level 3 inputs.

Discount Rate construction

Our senior analysts are often asked how we construct the discount rate we use to calculate loan (and deposit) fair values.  The Fair Value Disclosure section of our A/L BENCHMARKS Executive Report sums it up pretty nicely.

For a given balance sheet category, the model will employ one of two techniques to construct the discount rate. They are the Build-up approach or the Risk Premium approach.

Build-up Approach
The Build-up approach views the discount rate as consisting of four possible components:

First the Risk-Free Rate forms the foundation of the discount rate and is derived from the Treasury yield curve. It is the point on the Treasury yield curve which corresponds to the maturity structure for future principal and interest cash flows. A geometric interpolation is used between the discrete points of the yield curve to approximate the curvilinear shape of the yield curve. Thus the values of this component vary across time to reflect the term structure of interest, while the remaining components are constant across all time periods. This component also includes an adjustment to reflect the value of imbedded prepayment options.

Next the Credit Risk component is the annualized yield needed to cover the loss of value expected over the entire life of a portfolio. For loans, the derivation of this component is based on an analysis of underperforming loans and the gross charge-off experience of the past four quarters. For funding sources, the derivation of this component is based on financial ratios that reveal the “credit-worthiness” of the financial institution for which present values are being computed. In both cases, the higher the credit quality component, the higher the credit risk.

Next the Operating Expense component represents an annualized cost rate derived from operating expense allocations. This component is used to adjust the risk-free rate in order to compensate for operating expenses. When dealing with core deposits, the operating expense component is reduced by a service charge component (see below).

Finally the Service Charge component represents an annualized income yield derived from operating income allocations. It is used to reduce the operating expense component.

Risk Premium Approach
The risk premium approach views the discount rate as the sum of two components: the risk-free rate, and a risk premium. The risk-free rate is the same as defined above. The Risk Premium is the annualized yield needed to cover the risk reflected in the portfolio. It incorporates all forms of risk in a single spread to the treasury yield curve. Consistent with an entry rate concept of selecting a discount rate, the marginal pricing rate for each account serves as the basis for determining an appropriate risk premium to the Treasury yield curve. This risk premium is calculated by subtracting the value on the curve which corresponds to the average maturity of the account from the account’s marginal pricing rate.

I notice that we published this nearly ten years ago in our old “Communicator” newsletter.  The approach is still the same, although we have updated the weightings and allocations in the Credit Quality and Operating Expense components with newer statistical data.

Mark-to-Market: Fair in Theory, But of What Value?

I’m glad to see the financial press is focusing more on the downside of mark-to-market accounting for banks.

Mark-to-Market: Fair in Theory, But of What Value?
American Banker, September 8, 2009 (subscription required)

I was a little disappointed in the author’s answer to the direct question: “Which aspects of fair value have been most damaging to banks?”:

The real damage to banks occurred when they had certain investments or commitments that were risky and that ended up working out very poorly. Disclosing it is just really telling the world exactly what happened.

But fair value is a costly compliance issue. Where it also could be costly is if you misinterpreted the standard to mean that you really have to use the most recent market price, even if the market price was less than fair value because the market was inactive or because there were distressed sales…

While I agree with these points, it misses a huge problem: marking assets to market - but ignoring liabilities.  Writing down capital and earnings due to changes in asset values while ignoring the tremendous stability (and therefore value) of community bank core deposits just makes no sense.

Overall I don’t agree with using mark-to-market accounting for the entire balance sheet anyway, but if you’re going to use it, apply it equally to both sides of the balance sheet.  Isn’t that why they called it a “balance” sheet in the first place?

Fair values are dropping…and just wait until market interest rates start to rise

Lately every discussion I hear about fair value talks about the impact of credit quality.  It’s certainly understandable, the current credit climate stinks.  And most experts agree that the crisis will be fueled again by problems in the commercial real-estate market.

“High on my worry list is the possibility of another shock to the still-fragile financial system.  Commercial real estate is a particular danger zone…”
San Francisco Fed President Janet Yellen: Outlook for the U.S. Economy and Community Banks

Federal Reserve and Treasury officials are scrambling to prevent the commercial-real-estate sector from delivering a roundhouse [kick] to the U.S. economy just as it struggles to get up off the mat.  Their efforts could be undermined by a surge in foreclosures of commercial property carrying mortgages… Commercial Real Estate Lurks as Next Potential Mortgage Crisis, WSJ August 31, 2009

Commercial Mortgage Delinquencies Increase Sharply, CalculatedRisk on 5/07/2009

I quite agree that the commercial real-estate sector could deliver that “roundhouse kick”.  But I find it curious that there is little discussion about the potential impact of a rising interest rate environment.  Fair values are dropping as we speak, just look at the recent data reported by the largest banks in their SEC filings:

Loan-Value Data Shows Ills Persist, American Banker, Monday, August 31, 2009 (subscription required)

But if fair values are dropping at alarming rates now (presumably because of default rates and declining collateral values), just wait until market interest rates start to rise.  It’s basic finance 101, when rates rise values fall.  We’re already seeing sharp write-downs in bank capital (“In a Blink, Bank's Assets Revalued, Capital Is Gone”, American Banker, 8/31/2009 – subscription req.)  Marked-to-Market accounting on commercial loan portfolios in a rising rate environment will only pour gas on the fire.  Variable rate pricing will help soften the impact, but it won’t block the punch.

How much of an impact?
Although I can’t say precisely, I can look back at recent history.  Between 1st quarter 2004 and 3rd quarter 2006 the average fed funds rates rose +425 basis points.  Over that same time frame the average loan premium for banks $300M-$500M dropped by approximately –2.70%.  If you are using fair value accounting, that’s a pretty big bite out of capital. 

It’s also worth mentioning that this rate rise took place over nearly 3 years.  My sense is that when rates start rising again, they are likely to rise much faster.  Talk about a roundhouse kick and punch combination.

Viewpoint: No Accounting for Fair-Value Benefit

(Shameless self-promotion warning!) I guess I stood on my soapbox long enough.  My opinion piece made this morning’s American Banker (subscription required).  It’s been hacked and shortened by the editors, but it’s their paper – it’s their privilege.

Deposits marked-to-market – Idiocy

It turns out that yours truly is not the only one that’s quite cynical about the FASB’s “master plan” for marked-to-market accounting.  According to an article in this morning’s American Banker, “In FASB Fight, New Front Opens on Deposit Values” (Friday, August 14, 2009, subscription required), FASB is considering how/when to implement marked-to-market accounting for deposits too and more than a few industry experts think it’s a bad idea.

I won’t parrot the entire article, but here are two outstanding (and I think insightful) quotes from the text:

…to Christopher Whalen, managing director of Institutional Risk Analytics, the whole exercise would be a waste.  "The mark-to-market of a deposit is par, and to say otherwise is absurd. Deposits don't trade, there is no observable market, and they're mostly a function of how aggressive a bank's management wants to be in terms of growing or maintaining liabilities in that particular class instead of the alternatives" such as bonds or short-term funding, Whalen said. "The FASB has to hold back from this idiocy and just admit there are certain things we can't know."

and this one:

…Ethan Heisler, a managing director with Hexagon Securities LLC in New York.  "It's such a complicated process," he said. "Even a small bank offers so many different kinds of deposit products with different kinds of assumptions and different kinds of incentives. It's hard for me to see how boiling this down to one line item on a fair-value basis means much."

So if the fact that deposits don’t trade, and there’s no “observable market” is a good argument for not using marked-to-market accounting for deposits (which I think it is), then why are we talking about using it for loans that don’t have an “observable market” either? - Idiocy.