What do you buy, when you buy a bank stock?

Valuing (Greek) Banks: A Sum-Of-The-Parts Approach

What do you buy, when you buy a bank stock? You effectively buy the constituent parts of a bank, which are:
  1. its existing book; and
  2. its future business,

the latter of which can further be broken down to:

  1. its deposit-gathering line-of-business;
  2. its loan-servicing business; and
  3. its loan-origination/distribution business

Don’t think we are forgetting something here, so let’s give it a shot at valuing them one-by-one.

Let’s start by looking at Eurobank’s existing book as an example [Greece]. The market value of the bank’s equity and debt is around €72.5bn (liabilities are valued at par, as they mainly comprise deposits and Eurosystem funding). With that, one buys tangible assets of a book value of €71.5bn. In other words, holders of Eurobank equity are valuing the existing book and the future business almost on a par with the book value of the existing assets.

Let’s see then how some of these assets are valued on their own. Take, for example, Themeleion IV, Eurobank’s largest outstanding RMBS issue (placed, not retained). The A tranche currently trades at around 78 cents (if anything, the price of Themeleion IV is supported by purchases form Eurobank, which can book a capital gain by buying Themeleion paper and putting it in its hold-to-maturity portfolio). Also taking into account the B and C tranches, this translates into buying the underlying loans at 76 cents.

Let’s compare that with Eurobank’s residential mortgage book, just north of €18bn in size. According to its Q4 2014 presentation, 22.5% of it was 90dpd, while provision coverage was 36%. This translates into a book value for the mortgage book of Eurobank of 92 cents (1?22.5%*36%). Also applying the very small discount-to- book value at which holders of Eurobank equity value the bank’s assets, we arrive at an implied valuation for Eurobank’s mortgage book of just over 90 cents. The difference of 14 cents between the RMBS pricing and this implied valuation can only be justified by means of assigning substantial value to the future business of the bank.

For the mortgage business, this value is around €2.5bn. By applying the same discount across the bank’s other loan classes, we arrive at a back-of-an-envelope valuation of the future business of Eurobank of around €7.5bn. This is more than four times the current market cap.

Let’s look at the future businesses of Eurobank one-by-one then, starting with the deposit-gathering one. How much is this worth in today’s interest-rate environment? Well, certainly not a whole lot to JPMorgan, which plans to cut as much as $100 billion of some clients’ excess [sic] deposits. Eurobank’s most recent deposit spread is 138 bps. For comparison purposes, ELA costs 155bps, versus 5bps for regular ECB financing.

Then we have the loan servicing business. Well, we could estimate what the value of that is. Total (gross) loans are €51.9bn; let’s further assume Eurobank outsourced the servicing of its assets to someone like Nationstar Mortgage Holdings (NYSE:NSM).

Let’s look at some Nationstar metrics. It generated just north of $1bn in servicing revenue from an UPB of $386bn, had a core pre-tax income margin of 30% and a P/E of 10.50. Running the same metrics on the Eurobank loan book we arrive at a revenue figure of €144m for the servicing business of Eurobank and at a valuation of less than half-a-billion euros. This is around a fifth of the implied value of the future mortgage business of Eurobank, and one-fifteenth of the implied value of its entire future business. Which in turn means, the market relies predominantly on the future origination business of Eurobank to underpin its current valuation.

Right, the origination business, then—shall we just mention that credit to the Greek private sector contracted by 3.1% in 2014? But of course, it will recover, no?

It will. But let’s see what Eurobank’s starting position will be in that race. Let’s not talk about (marginal) cost of funding at this stage—this is a lengthy discourse. Let’s just focus on Eurobank’s cost of origination.

Of Eurobank’s total opex of €1.054bn, we could allocate €144m to its servicing function (approximately how much a third-party servicer would charge). That leaves just over €900m to allocate to the other (profitable) businesses of Eurobank. We just mentioned that the deposit gathering business is not exactly profitable. But let’s allocate enough opex to that function, so that the total cost of deposits (interest expense plus opex) is equal to the (hefty) cost of ELA. To do that, we would need to allocate another €60m to Eurobank’s deposit gathering function. That leaves another €850m to allocate to Eurobank’s loan origination/distribution business.

Now let’s assume things don’t just improve, but they go back to 2008—the best year for loan origination in the history of Eurobank (and the Greek banking system in general). That year Eurobank’s loan book increased by €10.6bn. If one were to allocate the €850m of Eurobank’s opex to a loan origination volume of €10.6bn, this means an additional cost of 8% to Eurobank’s new origination. For Eurobank’s origination business to be profitable, therefore, it needs to be able to charge its customers an additional 8% for its opex.

How much would a third-party loan origination business charge to originate the same loans? Probably around 1% (that’s the standard in Greece—as far as we know, the maximum that has been ever paid is 1.4%), but since Lending Club charges between 1% and 2% for loans with an interest rate below 5.9%, let’s go with 2%. Assuming we go back to the 2008 party (on top of making a few other wildly optimistic assumptions), Eurobank will need to decrease its opex by 75% just to be able to hope that it remains competitive.

Which it probably won’t. Because other competitors will crop up. Aktua, the Spanish servicing arm of Centerbridge, is already in Greece. It’s not terribly difficult to go from servicing to loan origination—funding is not exactly scarce these days (more on that point in particular in a future post). If they do move into origination, these guys will also have zero cost of regulatory capital (being non-bank lenders).

Speaking of regulatory capital: Eurobank’s fully-loaded Basel III CET 1 capital is €4.1bn, of which €3.6bn is DTA. For DTA to remain an asset, of course profitability needs to materialize.

To summarize: traditional banks were built to gather deposits. Their design and infrastructure is geared towards that; they are maladapted to today’s interest rate environment. P2P platforms and other non-bank lenders are eating their lunch. Absent some radical rethinking of their operations model, they are about to go the way of the dodo.

Of course, markets are not efficient; it will take time for competitors to move in. Traditional banks, however, have few weapons to fend them off: their brands (much less valuable for loan-origination than for deposit-gather purposes) and their (hugely expensive) legacy infrastructure.

It took almost a century for the dodo bird to become extinct. That said, equity markets are, theoretically at least, discounting future events. At today’s interest-rates, even events much further into the future affect values a lot.

Oh, and if you think interest rates will rise again: go short some long-dated Bunds. A lot less complicated.