Description

I am recommending a short of Astoria Financial (AF), a highly levered Long Island bank with deteriorating fundamentals that will likely see its earnings collapse in 2007 from the current consensus of $1.94 to somewhere near $1.35 (putting it at a PE of 23). Earnings have already come under pressure because of the inverted yield curve, and unless the Fed cuts rates in the next six months, the sell-side will be forced to substantially lower their 2007 numbers again.

Company Description—Continued Deterioration in Revenue

Astoria is very easy to understand, as it is a classic bank thrift. It holds jumbo mortgages (mainly 3-1 and 5-1 hybrid mortgages), a small amount of commercial real estate loans (mostly apartments) and investment securities (i.e bonds and AAA mortgage backed securities) and funds these with a combination of retail deposits, brokered cd’s and borrowings.

This model worked great with a steep yield curve, but has encountered problems as the curve has flattened and inverted. To its credit, AF has done an admirable job trying to stave off the inevitable earnings collapse by cutting operating expenses (down 4% in the last year). It is my belief that they have cut costs as much as they can (without unduly hurting their operations) and levered up as much as they can (20 to1) and that the net interest margin will continue to compress from 1.92% to near 1.50% by Q2/07. Much of the sell side is predicting a mild margin decline in Q3/06 and then a rebound either in Q4/06 or Q1/07. A brief analysis of the balance sheet and its funding structure indicates that this will be nearly impossible unless the Fed cuts rates soon.

The most telling numbers in the table above are yields earned on assets and paid on liabilities. Over the last year, the yield on AF’s residential mortgages rose 23 basis points from 4.83% to 5.06%. Even more interesting is the yield on its commercial real estate portfolio which has actually fallen by 2 basis points over the last year to 6.07%. Overall, the yield on Astoria’s total earning assets has increased only 24 basis points in the last year from 4.91% to 5.15% despite 200 basis points of Fed increases. On the liability/funding side, AF has proven to be more sensitive to the Fed’s rate increases with total funding costs rising 54 basis points from 2.79% to 3.33%. Upon first hearing that funding costs have increased by only 54 basis points despite 200 basis points of Fed rate increases, one might conclude that AF has actually been doing pretty well. And that is exactly what makes them such a good short—because the funding costs will continue to rise even though the Fed’s pause has created some degree of optimism and hope. AF’s funding base is not that good. Of a $22B balance sheet, low-interest core deposits, which are largely insensitive to changes in interest rates, constitute only $4.5B. (Despite management’s attempts to characterize some cd’s as core, they are not, as the 4.30% cost for their “liquid cd’s” attests—these are the single most expensive funding source on their entire balance sheet!) That means that about 75% of the earning assets are funded with rate sensitive products such as borrowings and broker-originated cd’s that are driven by the Fed funds rate, currently at 5.25%. With brokered cd’s costing Astoria only 4.07% in Q2 and borrowings costing only 4.27%, it would seem that both of these, totaling $15B, have substantial room to continue repricing closer to Fed funds over the next several quarters. This seems especially likely given that most banks in the NYC metropolitan region continue to report intense deposit pricing competition. There has been an explosion of branch building over the last five years and banks are desperate to fill them with deposits, even if those deposits are in the form of high-rate cd’s that don’t allow for any meaningful interest rate spread. Citi and JPM, in particular, continue to aggressively promote 5%+ cd’s, determined to defend their majority market share after years of slippage to companies like CBH.

To state it simply: Astoria’s yield on its $21B in earning asset is increasing about 5 basis per quarter (and with the recent pull-back in the 10-yr. this could be even less in Q3/05 and Q4/05). The cost on its $15B in non-core funding has been increasing about 15 basis points per quarter and remains roughly 100 basis points below Fed funds. The interest rate margin could easily fall another 40 basis points over the next few quarters, which would compress quarterly net interest revenue to something near $80M from the current $100M level. This translates into a quarterly EPS decline of $0.14, or almost 30% on the $0.49 that it just earned.

Operating Expenses—An Added Bonus

Astoria has recognized its unsolvable revenue problems and has done the only thing it can do—try to cut expenses. From Q2/05 to Q2/06 FTE headcount fell 12% from 1,864 to 1,635 and operating expenses fell 4% from $57.6M to $55.2M. Astoria deserves applause for this, but it is my belief that in an intensely competitive marketplace such as New York, expenses can only be cut so far before there is either an employee backlash or customers stop walking in the door. (Astoria was also known to be a lean institution anyway). We can already see the effect in AF’s sharp drop in core deposits—over the last year, savings accounts are down 15%, demand deposits are down 4%, and money market accounts are down 33% (see table above). I think expenses will have to start rising soon, certainly for the 2007 budget at least, which could take off another $0.10 of EPS in 2007.

Risks

Acquisition risk is a legitimate concern. CEO George Engelke is 67 and would probably sell, although any buyer would have to be wary of the potential earnings collapse. Hudson City (HCBK) is often mentioned to be the most likely buyer. While it seems to make sense at first (similar business models and HCBK has a market cap twice as big and excess equity), the deal would be quite big for HCBK in an operational sense, as AF has over 1,600 employees and HCBK has under 1,200. HCBK’s CEO, Ron Hermance, also recently noted at the Lehman Brothers conference that they were very uninterested in doing large acquisitions at this time, particularly given the uncertainties with interest rates and the yield curve. TD Bank North, another acquisitive bank in the Northeast, has recently admitted problems with its Hudson United acquisition and noted they will be more disciplined with deals going forward. I think a marriage between HCBK and AF ultimately makes sense, but I don’t think it will happen for a couple of years at least. (Additionally, because AF has already cut expenses so much, there isn’t nearly as much left for an acquiror to take out.)

Astoria has also been involved in a court case involving the IRS and some taxes involving poorly performing S&Ls they were forced to acquire during the crisis of the late 80s and early 90s. If Astoria wins, the end impact could be anywhere from $2.80 to $4.40 in book value per share. I don’t have any real insights into the case, but this has been going on for several years and a resolution does not appear imminent.

Valuation

AF trades at 2.8 times tangible book value,16.2 times 2007 consensus earnings, and 24 times my 2007 estimate, leaving room for substantial compression. The stock currently trades 28% above its November 2005 low of $24.43, even though EPS was $2.26 in 2005 vs. current consensus of $1.93 for 2006 and $1.94 for 2007,both numbers which should come down. I think AF could easily test those lows again within the next 4 months, translating into a decline of roughly 25%.

Incidentally, LSV, a quant driven shop, is one of the largest owners of AF, at roughly 6%. If the fundamentals continue to deteriorate as I expect, with falling ROE and ROA, missed estimates, and negative momentum, we could have a major holder liquidating as well.

The Fed and Interest Rates

While divining the Fed and interest rates is nearly impossible, I do think it is unlikely that Bernanke will cut rates within the next six months. If he cuts rates before some of the speculative excesses of low interest rates is purged (most obviously housing), he will lose credibility as an inflation fighter. He probably won’t cut rates until he sees some pain developing in the economy—I would say at least six months of it. Currently, we have almost zero pain— the country is at full employment, corporate profits remain strong, there has been no big spike in housing foreclosures, and most banks continue to report good credit quality statistics (with a few small blips) and solid demand for loans and credit. Anecdotally, if our recent conversations with several bank executives from across the country mean anything, credit quality will remain benign for at least the next two quarters (a reasonable timeframe with which lenders have good visibility into the operational health of their borrowers). In order for any real steepness to occur, the Fed will need to cut by at least 100 bps, and in order for that to happen, the U.S. will likely be in major recession caused by a bursting housing bubble, in which case almost every stock will go down, especially stocks of banks involved in residential lending—like Astoria.

Thornburg Mortgage as a Hedge

We are long Thornburg Mortgage (TMA), a mortgage REIT with minimal credit risk that already trades at 1.1 times book. There is risk of a small dividend cut next year (quarterly dividend could fall from $0.68 to $0.55), but even then, the dividend yield would equate to 9%. TMA will be a huge beneficiary if the Fed cuts rates. Both the founder and the president recently bought stock. Several Mortgage REITs have been written up on VIC, including LUM and SFO. Reading through them may help in understanding how to think about mortgage REITs, but basically they buy low risk mortgages and MBS and fund them with a variety of short and medium term borrowings, employ between 10 and 20 times leverage, and pay out 100% of earnings as a dividend.

Why AF vs. Some Other Thrift?

While shorting most thrifts probably won’t lose you a ton of money in the near term, AF is a particularly good short candidate because of its high P/B ratio and the high leverage. Hudson City (HCBK—a New Jersey thrift), New Alliance (a Connecticut thrift) and BRKL (a Boston thrift) have very similar looking operations, but both have tons of excess capital which they can continue to lever, and trade at P/B’s below 2.0 (although with much lower ROEs in the 4-7% range). Theoretically, they should trade at about book value, but some investors seem to be hung up on the fact that anything below 2.0 times book is “cheap” even if the ROEs are extremely low.

Catalyst