A number of people are beginning to see a light at the end of the Citigroup (C) tunnel. I still have significant doubts. In an article on December 30, 2007, Todd Sullivan pointed out that if Citi is short the $11 billion needed to pay the dividend for one year, that shortfall could be made up by selling an equivalent amount of assets. If we assume that the value before this one time loss made sense, and that $11 billion is the full extent of the loss, then that sounds very reassuring. Collecting 7% from a reliable financial firm, with a 15% tax rate sounds really good. Citi might be very good for my retirement.
Is it correct that $11 billion is all that needs to be made up, and that it's a one shot deal? Does anyone know the magnitude of the losses Citigroup has incurred relating to the incredibly unwise risk in taking on poor quality A.R.M. residential mortgages labeled as AAA creditworthy? I have seen a number of sources saying no one knows the extent of the losses. That surprises me. Financial people can't be making decisions without any analyses or estimates at all. I wonder how would one go about making such an estimate. Can anyone help me with that? I am just a retiree who knows almost nothing about financial evaluations. Then too, off-the-books entities would seem, to my untrained mind, to make it more difficult. But perhaps it's not totally impossible. In the interim, I will take a naive crack at it. Here goes:
We have to start somewhere, so let's try this approach: (Please point out my errors.) Let us assume for the moment that the subprime losses for a large bank are $50 to $100 Billion. Then we will see if that makes practical sense.
Here is how I arrived at that guestimated (engineering jargon) range. A typical number for the loss already admitted to by a number of large banks is in the vicinity of $10 billion plus or minus a few billion. What kind of activity could have generated a loss of $10 billion? It could have been rung up by 40,000 non performing mortgages of $500 thousand each. I am assuming that each property lost 50% in resale value. That means they wrote only twenty thousand non performing mortgages a year in about two full years during 2005, 2006, and 2007. That would translate to about 1700 non-performing subprime mortgages being written per month. What fraction of the actual number of subprime loans written or purchased by Citibank in twenty four full months between 2005 and mid 2007 does that represent? That is the real question. To start, I will assume the fraction is one tenth. Otherwise, why bother with moving risky loans to off-the-books entities? That fraction would mean that 17 thousand loans likely to be in default were written or purchased by Citibank, per month.
Is that number of loans realistic? Let's see. There are roughly ten large banks, and I will assume there were about ten large subprime writing factories and many smaller ones. The key is I am assuming one large subprime loan generating factory per large bank.
How many loan writers were there? Lets leave the many smaller ones aside for now. I seem to recall First Magnus in Arizona had a layoff of something like 200 loan writers when they couldn't sell any more of the clearly worthless loans, even with slicing and dicing tricks. If each bank had about 200 people writing about 4 loans a day, 20 days a month, that allows a $100 billion loss in 2 years. If we assume most mortgages were $250 thousand instead of half a million dollars, that means a loss of "only" $50 billion per bank, everything else remaining the same. What is the effect of all the other low initial rate, ARM, no documentation loan writers? Is a factor between 3 and 10 reasonable? Are these numbers available somewhere?
Since all parts of this first cut estimate are subject to significant variation in terms of the unknown by me) facts, I would be cautious and say the number is likely to lie somewhere between $50 billion and $500 billion per large bank like Citi. Getting to $250 billion is easy with a few factors of two or three, so I will write $500 billion to be cautious. Even $50 billion is not small change, and that assumes only 200 bad loan writers. If there were only ten First Magnus size offices devoted to defrauding Citigroup's investors, we are in the neighborhood of $500 billion dollars of undeclared losses. Five hundred billion dollars represents a significant fraction of Citigroup's total asset value of $2.3 trillion (taken from Todd's article), or $2300 billion. That's roughly 20% of total assets. It doesn't include any other liabilities.
Would that mean the dividend would be cut? That would really hurt the bondholders. The stock would take a significan long term hit I would assume. I am not competent to evaluate the effect of only a fifty billion dollar real loss on the value of bonds or on the stock valuations. What would happen to the book value seems important. I assume it would be difficult to sell financial assets if they are suspected of being tainted by fiscal irresponsibility of the type of Citi itself. Even if they are pure as the driven snow, there may not be enough buyers for the sheer volume of the assets to be liquidated by so many banks at about the same time frame. There may not be enough foreign buyers when we need them.
Again, this is not my area of expertise, but if this really simple estimation process is reasonable, we have trouble in Citi City.
If the actual loss lies anywhere in the uncertainty range ($50 billion - $500 billion), it could be risky to assume that a 40 percent cut in the dividend is sufficient to fix Citigroup's problem. Cutting the dividend would be a major statement. Is it a correct statement or not? Also, what does this imply for reassessing Citi's actual value? Wouldn't that mean that the current 40% drop in Citi's NAV over the last year is not enough? Would there likely be an additional 20% reduction, or would it be larger or smaller?
If you see one or more flaws in this initial approach, please let me know. Also, please suggest a better one. I am not a financial guru, just a retiree trying to see what are reasonably safe investments, and what are not. I would like to find a way of estimating the real risk. Inflation is a risk I understand somewhat. However, the value and risks of financial paper issued by Citi or of stock in Citi are still a mystery to me. I feel I have just begun to peel away the layers of the onion. I am startled, amazed even, to discover that supposedly mature, knowledgeable, experienced managers, who had a fiduciary responsibility to their institution, purchased highly correlated investments whose outcomes were obviously not statistically independent. Those investments by their very nature had a high probability of failing at around the same time, when "tickler" rates ended. This was an easily reachable and predictable circumstance. What were they thinking? What were they doing in Finance 101 classes? Since their financial actions have not engendered a sense of trust, would their institution be a good vehicle for a DRIP investment or only for a drip down an unstoppered drain? I saw Ben Stein on Kudlow ascribe it to the temptation of having fabulously rich grandchildren overriding fiduciary responsible and I would add honor, honesty, and decency. They were too easily seduced to becoming thieves.
If anyone understands this subject and can suggest other ways of estimating, either deterministically or stochastically, Citi's likely stock and bond values and the associated risks, I would really appreciate your help!
I came across a Forbes article quoting an analyst as saying the immediate loss is $30 billion, with possibly another $100 billion. No analysis was shown. This puts her in my range.