Citigroup: How do You Value Systemic Risk?

"We have to consider the possibility that the housing price downturn will eventually be as big as that of the last truly big decline, from 1925 to 1933, when prices fell by a total of 30 percent." (Robert J. Shiller, The New York Times, November 25, 2007)

A couple of years ago, when The IRA started to comment on the anomalous, below average loan default experience reported by virtually all US banks, we were accused of iconoclasm, of not understanding the brave new world of structured assets, over-the-counter derivatives and other types of toxic waste which Wall Street mistakes for real investment instruments.

We wrote in May of 2005: "We think that US mortgage and corporate loan default rates are considerably understated. As the Fed raises short-term interest rates to positive real levels for the first time in half a decade, we believe that default rates could swing toward record levels. Only question: Is the coming wave of real estate and corporate defaults going to be a "typical," two-standard deviation event like 1990-1991, or a mega crisis a la 1982, when the banking industry's entire capital base was impaired in a single quarter by the caprice of a few Mexican politicians?"

Well, guess we were right after all. But remember that despite all the fuss and bluster in the financial media of the past two months, and the big valuation hits that large banks have taken to their trading books, bank loan default rates still remain pretty low. News that HSBC (NYSE:HBC) is taking its lumps on a strucutured investment vehicle or "SIV" is more of the same trading book noise.

When an entire industry for a period of years displays abnormal behavior, either up or down; irrational activity that suggests that there is an external factor or group of factors causing members of that industry to skew their financial results, we call that a "red flag." When 2006 turned out to be the 7-year trough in bank loan defaults in the US, the shape of the future was apparent -- at least to those who look at the numbers. Thus our contempt for observers who, even now, express surprise at the scope of the subprime fiasco.

Many of our colleagues in the risk management and supervision communities have known for several years that there is a problem baked into the financial system. We're not talking here about the political appointees and their minions, but rather the examiners, ratings analysts and audit professionals who toil in the financial trenches and do the hard work of fundamental analytics. But what we find really amusing is that there are still "value" investors abroad who think that some of the larger banks have positive net worth available for exploitation.

Last year, when Buy Side observers started to argue with some considerable justification that Citigroup (NYSE:C) was underperforming and should be broken up and sold, we took a contrary position, pointing out in several issues of The IRA (Citigroup: Does a Breakup Make Sense? (Parts I & II)", January 2, 2007) that breaking up C might make financial sense, but would meet with stiff resistance from regulators, who are more concerned with safety and soundness than shareholder returns.

Since we wrote those comments, C has written off billions of dollars in trading book losses. Former C Chairman & CEO Chuck Prince joined the growing list of subprime casualties and the BOD of C showed itself to be completely inept, firing a CEO before a replacement was chosen. Job One of any board, by the way, is supervising the CEO and planning for his or her succession. Guess that latter duty has somehow escaped the notice of C's star-studded BOD lead by former Treasury Secretary and Goldman Sachs (NYSE:GS) honcho Robert Rubin.

The trouble with making the "break-up Citi" argument now, in December of 2007, is that it comes two years too late. The time to take a serious look at unbundling the C empire was when the mortgage craze was still in full swing and both consumer and corporate defaults were at or near all-time lows. Today with C leaderless and facing mounting losses both from the trading and banking books, it is hard to imagine a worse time to attempt a sale of assets or even the whole business. The estimates of $15-17 billion in possible trading book losses for C in Q4 will not be the end of the house cleaning.

The other issue, for those not familiar with the world of bank regulation, is that the various banking business lines which comprise C are now concentrated into three main banking units, with the lead bank, Citibank NA, accounting for the wholesale cash and derivatives trading business, most of the mortgage portfolio and a considerable chunk of credit cards and other consumer loans. The table below shows the summary public data Basel II benchmarks from the IRA Bank Monitor for the three largest C banking units.

click to enlarge

The IRA Bank Monitor (June 30, 2007)

It might be possible for the new CEO of C to sell assets of the parent holding company, but regulators are unlikely to condone any transactions that would weaken the capital position of the bank units, especially lead unit Citibank NA. While you could conceivably sell Citibank South Dakota or Citicorp Trust, the valuations are unlikely to be very compelling and, again, the regulators might resist the sale of units that are more profitable than the lead bank.

At the end of June 2007, Citibank NA had tier one risk-based capital of $67 billion vs assets of $1.1 trillion, for ratio of Tier One RBC to assets of just 5.9%. Even in good times, such a capital cushion would be less than adequate. But in the face of the mountain of bad debt which is now bearing down on C and other large US consumer lenders, single-digit capital to asset ratios represent a grave danger to solvency.

Of interest, the Economic Capital model in the IRA Bank Monitor assigns a ratio of of EC to Tier Once RBC of 4.1:1 to Citibank NA, well above the 3.6:1 assigned to the entire C organization. Through the first nine months of 2007, Citibank NA's charge offs were running at an annualized rate of just 90bp or less than 1% of total loans, hardly a crisis -- yet. As our former colleague at PruSec, Mike Mayo, nicely describes it: a slow motion train wreck.

Of course, the same crowd who are making the "break up the Citi" argument in the past pressed management into implementing a series of extortionate share repurchase programs. When apparent risk was near zero, holding more equity may have seemed costly. Now C and most large banks rightly are perceived to be short of capital. Accordingly, we expect to see C eventually reduce or suspend its dividend in 2008 to rebuild or at least stabilize tangible capital levels until risk from the banking book has passed.

So don't start holding your breath waiting for C to put itself up for sale, in whole or in part. The market turmoil of 2007 was all about a "market disturbance," as the Corrigan group wrote in July 2005, while future losses on consumer and wholesale exposures will be more in the nature of systemic financial shocks.

To review, a market disturbance is caused by uncertainty as to the price or liquidity of an exposure, while a systemic financial shock comes from not getting paid. Or as the Corrigan group put it so well: "...financial markets have a remarkable capacity to cope with financial disturbances so long as widespread credit problems are not seen as an imminent threat. Experience also shows that the fact or the fear of large credit losses is often the key variable through which financial disturbances become financial shocks."

With asset and equity returns at C trending lower, and loan default rates at the major units still near five-year lows, setting a valuation for C would be no more easy than, say, valuing E*Trade (NASDAQ:ETCF) with the problematic mortgage portfolio still attached. Of note, the portfolio of E*Trade Bank is only throwing off 30bp of defaults on an annualized basis, but with a gross loan portfolio yield of just 505bp -- a full 1.5 standard deviations below peer -- you can understand why potential buyers are less than exuberant.

Citibank South Dakota's 350bp plus in run rate defaults in Q3 2007 is less than half the Q1 2004 peak of 831bp, a level we expect to see again by the second half of 2008. Citibank NA, now at 90bp, saw loan and lease defaults reach 230bp in that same period. As 2008 unfolds, we expect to see all of the C bank units test the peak loan default rates of the past five years and then some. The only question is not whether but rather how quickly C will test the loan default rates seens in the early 1990s, which were about 2x 2004.

So if you are a banker or investor trying to get a handle on the valuation of C or ETCF or any other banking asset, the good news is that bank market valuations are well off of the peaks reached during the housing bubble craziness. ETCF changed near $30 in Q1 2006, while C traded at $57 at the beginning of 2007. But as the imbedded losses on the banking book of most institutions become increasingly visible and painful in 2008, bank equity valuations are likely to move even lower. Stay tuned.