Texas and other Bank Ratios: A continuation of Everydayinvestor's blog
I intended to initially do a response on Everyday's blog....just became too lengthy
Michael.....I was meaning to do a blog on Texas ratio....but you beat me to it. Good - more people read your blog anyway. And I hope you dont mind a little bit of a hijack.
I am sure its great to come up with new measures - but this is just an exposure/leverage measure - really does not tell you much other than - relative exposure. Like generally if a bank is reporting a 9:1 leverage , if you see a number closer to 9 on Reggie then you know the bank did most of its business in this class of loans. Very difficult to see the number not highly correlated with the general leverage number, unless extreme circumstances.
Now the good stuff: My algorithmic addition to Michaels process
Main inputs ( You will need to do most of this - SSBX was a slam-dunk, a pity I did not know about its existence until much later. Most of the time - you'll find fuzzy answers, hence more information, the better in my opinion)
(1) Tangible Capital: Ditto
(2) NPA ( Total non-performing assets , not just loans as Everyday says) This is in essence most 90+days due consumer loans and "distressed/uncollectible" classified assets of commerical variety
(3) Loan Loss Reserves
(4) Tier 1 Capital and Tier II Capital Ratios( You will find them together in the Capital Adequacy section - do a search as Michael says)
If available ( They usually are)
(a) Delinquency (NPA/Total loans) by Category: Commercial ( This usually has the sub-heads of Construction, Land Development etc - also broken out), Residential (1-4 family), Home Equity.
(b) REO and Available for Sale Assets
(c) Growth in NPAs - do this , usually will have Mar 31,2007, Dec 31 2007 and Mar 31 2008 , right now. Create a QonQ growth rate. See where the slope of the curve is going - I dont need to say the interpretation - the Fools world is not so ahem foolish!
(1) Compute Texas Ratio.Actually no , this shouldn't be a 1st step - Texas is indicative only. So if you have a database of Texas ratios - I would simply sort by that and pick the worst offenders - unfortunately I dont know of any such source - so I am assuming you are doing Bank by bank - hence
1a. Compute Total NPA/Loan Loss Reserve ( ie a portion of the denominator used in Texas).
Why, you really can't assume the bank's a gonner - the regulators will start here - basically is the bank providing enough to cover for future potential losses.
Any number above 100% is interesting ...but you are really looking for a high 200%+ number here.
1b. There's a little bit more to NPAs : Here you need to break it down into NPL ( loans) and REO.NPLs are not sure shot loss. Only a % age of them will actually move into Loss ( and then into REO eventually). Post that - it all depends on the underlying value of the collateral - ie home/building or land value. THIS IS A VERY IMPORTANT DYNAMIC. Loans do not go into loss status overnight - due legal process ( foreclosure timelines vary by state - NJ it could take an eternity) takes time - time buys a lot for the bank.
This varies by the loan type/product also: Riskier products have higher propensity than other. Home Equity (HELOCs especially), Option ARMS will have high percentages. So also Commercial - you'll typically find very few loans in this category in NPA as compared to Total loans - but those that go - are in real trouble - because of concentrated exposure - like 10 loans all related to 1 troubled private builder , or 100 loans in 1 project in jeopardy of construction etc.
1c. Then you need to consider what the value of the asset is. As Michael says for Home Equity ( 2nd liens) - its close to zero ie 100% markdown. ( NOT 100% Loss) For Land - Florida says 20 cents to dollar ie 80% Markdown. For general Home - 50% mark possibly.
1d. Thus you get Net expected loss = Total NPL( You will need to extrapolate this based on the rate of growth computation you saw in the QonQ) * % of loans moving into Loss * % age Markdown + REO* % age Markdown ( SEE THE DIFFERENCE - anything held for sale or owned needs to be valued at market , no probabilistic uncertainty here - so anybody with a buldgeoning REO book has some real issues)
1e. This number at a minimum has to be at least 25% less than the Loss Reserves. OR THEY WILL BE ASKED TO BUILD UP TO THE NUMBER. This is what's happening today with RF,KEY,FITB - big regionals
2. From the difference you can see how much of a variance it creates ( raising of the LLR to cover losses) to the Tier 1 and Tier 2 capital ratios - They have to maintain it ( anybody not doing - has given up ). IF THEY HAVEN"T RAISED CAPITAL YET : They have to, no choice. This is a simple immediate short - no bankruptcy required.
3. Chance of raising: Not everybody will be successful. SSBX 40 Million is a question mark. That means they are sunk. If they raise - and bring the ratios in order. Good for now.
4. ALL SAID: If you see (LLR - Net Expected Loss)/Tangible Equity Only ( ie no LLR here) >100% ( Basically a doubling of existing capital!) or really a high $ figure : That's your decision point. If they can't raise that much - well there's only 1 way to go.
Brings you to probably ask - what am I doing with all this theory. Very little - I only come to know of these candidates here at the fool - and most of the time - the opportunity is gone. There have been a few exceptions though - and profitable ones.
Otherwise, of course , if you think shorting banks is easy money - you have no real problems/doubts.