Description

Short FirstBancorp (FBP)

We recommend selling FBP above $3/share ($277m market cap).  We think recovery value for the common stock is about $150m, and suggest covering at $2 ($185m).

Overview

FBP is a holding company with banking subsidiaries in Puerto Rico, Florida, and the Virgin Islands.  They have $19.7b in assets against stated common equity of $0.9b.  80% of their loan book is in Puerto Rico, where they operate the island's second largest bank.  11% is in Florida, mostly around Miami. Over the past few years, FBP has grown its construction & commercial book without properly accounting for potential principal losses embedded in their loans.  The numbers indicate that, after accounting for credit losses, their normalized earnings over the past few years are negligible or potentially negative.  Currently, they have $1.3b in non-performing assets (6.5% of assets) against which they have $407m in loan-loss reserves.  FBP is reliant on brokered deposits for funding and faces high core deposit costs in Puerto Rico, at about a 100bp differential to BPOP, their larger competitor.  They received a $400m TARP injection; although they've suspended dividends, the 5% to the taxpayer will accrue - a funding source that isn't included in interest expense, but reduces income to equity holders.  Their lending environment is unlikely to improve anytime soon, with 14% unemployment in Puerto Rico, a debt-strapped government, and an overbanked economy where FBP lacks a clear edge (BPOP has their franchise, OFG has a clean balance sheet).  Given this negligible franchise value, we think recovery is limited to tangible book value, which we estimate to be about $150m, which is below the current trading value of ~$300m.  We think the mispricing is due to their policy of under reserving, and expect revaluation as the realize losses on their Puerto Rican construction and Florida portfolios.

Valuation

Catalysts

a) Regulatory action leading to a capital raise & reserve build to highlight their equity issues, b) recognition of other-than-temporary impairment on LEH receivable and DRL loan, c) further charge-offs on their Puerto Rican construction portfolio, d) downwards adjustment of the deferred tax asset in their annual report, e) 2Q09 30-90 day delinquency data being released through FFIEC reports (FBP cover this in their Q).

Mispricing

We think buyers at these prices are attracted by the one or more of the following: an optically low ~0.4x P/TBV, annual net interest income of $480m, continuation of the Scotia buyout story, a Puerto Rican recovery, or the chance to participate in a low-dollar short squeeze.  Aside from the last, we think investors will change their minds as problems reveal themselves through charge-offs, reduced returns, and capital raises.

Management

FBP's management is generally bullish in their approach, and they tend to underestimate downside risk.  They avoid quarterly calls, which reduces the opportunity to get clarity on their filings, but we've had the chance to meet them in person.  In 2008, they were excited about the shopping / retail commercial lending opportunity in Florida and in Puerto Rico - loans that are currently causing them concern.  Their expected recoveries on a newly acquired auto loan portfolio were far in excess of what more experienced island auto lenders were reporting.  They've consistently been more excited about lending opportunities than BPOP, whom we've found to be more pragmatic about Puerto Rican prospects.  Their optimism is evident in their reserve assumptions, which are highly subjective and are drawn from near-term historical data and are loosely tied to realistic outcomes.  2Q09 was the first time they comment extensively on the problems they face in their operating environment.

Implementation

On 7/31, the short interest was 11.7m shares, 14.5% of the 92.5m float.  Borrow is easy, at top rate, and the issue has been trading ~1m shares a day.

Risks

Complicit regulatory inaction is the main risk, as in the case of BPOP.  The US Treasury has agreed to exchange $935m TARP preferred stock for $935m TruPs, which they are allowing BPOP to recognize as a $500m gain to the common equity.  The accounting process essentially indicates that BPOP is issuing debt below market rates, and allows them to record the new debt at "fair value" on their liability sheet, with the ~$500m differential flowing through to common equity.  The cash flows are not affected, but there is an implicit transfer of value from the Treasury the stockholders, as a liability is reduced without penalty.

Additionally, a risk for shorts at this point is path volatility due to the low dollar price.  We see this position revaluing over the year as credit information comes out, but there might be rallies in the interim.  Some of the Puerto Rican banks have had 50% rallies (even during calmer times) before selling off 90% lower over time.

Holders

Scotiabank (10%), First Trust Portfolios (8.4%; income-focused fund), Fidelity (7.8%; mainly in their $2.4b "Low Price Fund" FLPSX), Barclays (6.3%), Vanguard (4.7%).  Anecdotally, the common stock and preferreds are also held by local retail investors for income purposes.  Scotia paid $12.67/share for a 10% position in May 2007, instigating buyout rumors.  They have since begun their own operations on the island with a clean balance sheet.

---

Shorting low-dollar-price small-cap stocks is not our usual style, but we know this company and its issues fairly well, and see limited opportunity for fundamental improvement over the next two years.  We have been involved in Puerto Rican names for four years, during which their public market valuation has shrunk from a combined $22b valuation in 2005 to $1.7b today, with RGFC & WHI on the brink of bankruptcy.

Our hypothesis depends on three key issues: a) low future earnings, indicating a negligible franchise premium, b) poor asset quality, c) a refutation of management's claim of being adequately reserved against potential losses. We examine each in turn.

Earnings Analysis

To establish FBP's core earnings potential, we look at four drivers: historical operating earnings, unusual contributors to net income, funding sources, and their loan environment.  After taking into consideration the principal losses embedded in their portfolio, we believe FBP has been running negative value operations for some time, and see little room for improvement in the future.

Operating earnings

Since 2006, FBP has averaged quarterly net interest income of $120m versus non-interest expenses of $80m. Their net interest margin was on par with their competitors, which is implies that they might have stepped along the WHI/DRL Taking into consideration the potential cumulative losses of the loans they held in their portfolio during that period, and annualizing over the various loan terms, it appears that the franchise lost money steadily through their loan decisions.  We take a simple look at their normalized earnings, accounting primarily for their construction losses and holding consumer loan losses at their current NCO rate. Our estimates for constructions writeoffs are rosier than those Puerto Rican lenders are already experiencing on their portfolio, but even so, they indicate that credit losses overshadow the yield provided on these loans.

Unusual gains

Over the past few quarters, FBP posted opportunistic one-time gains which are either unlikely to be repeated, or do not have an impact on book value.  They have had the effect of making net income look less unattractive than reality indicates, which would lead some modelers to build in high income expectations.  These gains, which should be viewed in light of FBP's quarterly normalized earnings of $40m, include: 

• - $9.3m gain on sale through the 1Q08 VISA IPO
• - $1.5m increase in mortgage banking activities, primariy from GNMA issuances
• - $4m interest income recognized on impaired loans
• - Income tax benefits of $14m and $95m over the past two quarters
• - Gains on sale on their MBS portfolio, $18m in 1Q09, $10m in 2Q09

Securities gains are already accounted for in OCI when considering book value - the company opts to realize the gains through the income statement while leaving losses in OCI.  The real effect of these sales is to lower the yield on earning assets moving forward.

---

Past performance is a judgment on previous operating realities.  Looking ahead at FBP's funding and loan origination issues, we see little room for improvement in their future earnings.

---

Funding

FBP has a relatively weak deposit presence in Puerto Rico.  Their core deposits have consistently cost more than BPOP (at 3.6% vs 2.5% in 2Q09).  Consequently, FBP has been heavily dependent on brokered deposits as a source of funding, and these account for 36% of their liabilities, versus 7% for BPOP.  They've mentioned that this source is getting more expensive - which might throttle their net interest income further this year as most of these deposits expire within the year.  Over the past few quarters, they have indicated that repo counterparties, who provide $4.1b (20%) of FBP's financing, are reluctant to renew financing.  If questions arise regarding FBP's capitalization, the bank might find it hard to replace their repo book, and only have a combined $1.1b of availability from their FHLB and FED lines.  Additionally, their interbank credit line shrank from $200m to $100m over the past quarter - not a large source of liquidity initially, but potentially indicative of a trend.

Lending opportunities

Puerto Rico is a bleak environment for most industries.  They've been in a recession since 2006, and their government finds itself with few levers to pull in the absence of monetary authority.  Unemployment is at 14%, and the largest employer - the government - has been cutting jobs.  A major local paper regularly reports on the glut of condos in an already depressed housing market, with increasing inventory at the high end ($150+ for PR) of the market.  On the commercial front, several of the resort projects of '06-'07 are currently on hold as builders try to work out financing and exit sale issues.  As the second largest lender on the island, FBP's asset quality and prospects for growth are tied to local economics. 

From 2004-07, the Puerto Rican banks expanded their loan portfolios rapidly in mainland Puerto Rico and in Florida (the latter due to an IBE tax incentive). Much of the growth came from increasing LTVs chasing rising pro forma housing valuations. The loans were priced competitively (FBP's current yield on their construction portfolio is 3.6%). At the moment of writing, the combined market cap of the Puerto Rican banks is $1.7b, half of which is accounted for by BPOP & FBP, the two largest. The other major commercial players - WHI, DRL, & RGFC - are already on the brink of bankruptcy.

In this environment, even a well-capitalized bank (of which OFG is the only local example) finds it hard to make good loans.  FBP has been originating ~$1b/quarter recently, we question the deployment of their funds. In 3Q08, they purchased a $218m auto loan portfolio from Citibank. From our conversation with them, their severity assumptions were much rosier than those of BPOP & EUBK - the more experienced Puerto Rican auto loan managers. Almost half of the $1.1b in 1Q09 came from a $500m loan to COFINA (a government agency), that was paid in the next quarter. The major contributors to the $900m generated in 2Q09 were drawdowns on previously existing lines of credit. $570m went to construction projects, which is a cash drain as about a third of the island's construction loans are in non-accrual.

Given their operating situation, it is highly likely that they will have a much smaller balance sheet in the future - FBP might be better off downsizing operations ahead of time.

Asset Valuation

We arrive at a ~$150m recovery value for the common stock by taking their stated tangible equity and adjusting for earnings and expected loan & asset writedowns over the next year.  Most of the credit losses are expected in their Puerto Rican construction portfolio, where they have yet to reserve against their increasing non-accruals.

FBP operates in three geographies (indicated credit exposure): Puerto Rico (81%), Florida (11%), and the Virgin Islands (9%). We look into the different loan issues they face in the first two portfolios.

FBP entered Florida along with the rest of their peers to take advantage of a tax break offered by the Puerto Rican government.  The group as a whole cited a strong housing market and an untapped local Puerto Rican population for the reasoning behind their asset growth in these areas form 2004-07.  At the moment, they're left with under collateralized construction and mortgage portfolios.

• $153m condo-conversion loans in Miami, net of a $48m charge they've already taken. Some of the borrowers have curtailed their conversion plans, and are using rental income to cover interest payments - principal repayment is another issue. Most of these were underwritten to high LTVs against boom-time pro forma projections. Estimated recovery value from the collateral indicates cumulative losses of about 50%, which means there are about $50m of unrecognized losses in this portfolio.
  
• In 1Q09, they listed a $6.7m construction loan for an 18 unit development ($416k/unit loan) classified as non-accruing. Current Florida data points indicate a 50-70% writeoff on such a loan, and they have $437m in construction loans left on their balance sheet, with $67m in reserves against them. They've taken writeoffs against this portfolio already (down from a ~$500m balance last quarter), but given the absence of a new boom in Florida, we expect larger charge-offs in the future.

In Puerto Rico, FBP faces significant near-term issues in their construction and commercial loans and long-term capital lock-up problems with their residential portfolio.

• $965m local construction book consisting of $422m in residential construction, $300m commercial, and $200m land loans. Against these prime assets, they have a total of $60m in reserves. Construction loans were originated with six-months of interest reserves over a two to three year building period. Loans are underwritten against the assumption of security from cash-generating fully-developed properties. A developer who has at least two loans outstanding with FBP once boasted to us that he had been extracting egregious terms from the local lenders as they'd been eager to allocate capital over the past few years. In Puerto Rico, most of the developers have either slowed down or stopped building as unit inventory rises on the island. They fund their interest payments by drawing down their existing credit lines, but there's no value created in the process, with a big question mark over how principal payments will be made at maturity. Tellingly, in their latest filing, they report "Construction loans originated in 2009 consisted mainly of disbursements on existing commitments". Basically, cash is being forwarded to an industry where NPAs are ~30% and writedowns range between ~30% when realized (In 1Q09 they took a $5.2m charge off one $15m residential construction loan in PR). We think this is where FBP will be forced to recognize the majority of their losses over the next few years, as interest reserves dry up, and the loans are marked to (non pro-forma) asset values. Loan loss reserves of 6.2% seem inadequate against this picture, with the $60m barely enough to cover expected losses on the ~$200m they have tied up high-rise projects.
  
• $336.3m commercial loan to Doral, secured by mortgages on residential and commercial properties. Doral was one of the more aggressive mortgage lenders, and the quality of the underlying loans are dubious at best. FBP previously had a similar loan out to RGFC for $205m. In 2Q09, they brought the loan bank onto their balance sheet, and took a $33.8m (15%) reserved against the underlying portfolio of residential mortgage loans. We expect the DRL writeoff to be higher than 15%, but kept the charge-off at $50m for our projections.
  
• $3.7b commercial loan portfolio, primarily to retail borrowers. This portfolio is just starting to show deterioration, with NCOs moving from 30 bps to 80 bps from 1H08 to 1H09. Puerto Rico has been in a recession for at least four years now, so the poor quality of commercial loans does not come as a surprise. However, there had been some expectation of fundamental changes following the election, which led to lenders such FBP being bullish on shopping & retail loans throughout 2008 and 2009. The recovery has failed to emerge, with the new government offering caution on the precarious economic situation. We expect higher losses as a new wave of defaults emerge in these sectors. 
   
• They have a $2.8b residential mortgage book with 11% non-accruals. Given the overhang of housing inventory in the Puerto Rican economy, we think recovery for NPAs will be low. However, the banks benefit (from an accounting perspective) from the long foreclosure process on the island - it takes about two years for properties to pass from NPA to being charged off. However, from an investor's point of view, that is $280m (or, more realistically, ~$150m) locked up, earning negative returns (after baking in property maintenance costs).
  
• In 1Q09, they restructured $90m of their loans, of which 40% were past due within the quarter. In 2Q09, those numbers changed to $230m and 78% respectively. We think restructuring merely delays NPA recognition, indicating greater losses to come.

Other assets with potential impairments include:

• $64.5m Lehman receivable. FBP had earlier posted collateral to Lehman Brothers against an interest swap agreement. In the last quarter, they classified this as a non-performing asset. They have yet to provide clear guidance on recovery. Although the Master Agreements will be vastly different, we have the following datapoints on the collectability of Lehman receivables from the FHLBs, which indicates recovery values between 7-15%. We assume 50% recovery on this asset.

FHLB Atlanta: "In accordance with the Master Agreement, the net amount due to the Bank as a result of such excess collateral held by LBSF is approximately $189.4 million.  Management evaluated this receivable in accordance with the guidance provided by SFAS No. 5, Accounting for Contingencies ("SFAS 5"), and related pronouncements.  The Bank recorded a $170.5 million reserve based on management's estimate of the probable amount that will be realized."

FHLB Pittsburgh:  They have a book of derivatives with LEH holding an excess collateral position as well.  They claim that the net receivable is $41.5MM and reserved $35.3MM since they determined that it was probable that a loss has been incurred with respect to this receivable.

• $218m deferred tax asset. Over the past few quarters, FBP has been running income tax benefits through their income statement ($14m in the first quarter, $95m in the second) to add on to their deferred tax asset. This asset is only recognizable if FirstBank is a) able to persuade the regulators that their claims are legitimate, and b) they are profitable enough to realize the tax benefits. We think that this asset will prove to be weak under the year-end audit scrutiny and assume a $100m write-down. DRL & RGFC have already recognized similar DTA losses.
  
• $110m private label pass-through MBS purchased from RGFC. Last quarter, RGFC paid FBP $4.2m to eliminate the 10% recourse provision in the documents, and FBP has allocated $30m in losses to OCI against this portfolio. We are not sure about the structure of this security, and have not factored in losses from this MBS, but this could be a source of future losses.

Reserve Policies

In 1Q09, FBP's LLR/Loans was 1.5% vs 4.3% for BPOP and 2.7% for Puerto Rico as a whole.  They have consistently underreserved versus an increasing NPA balance, as indicated below.  One of their explanations for an ever declining LLR to NPA ratio is that they take higher charge-offs up front, pointing to the 3.7% in the last quarter. We look at the same data and temper it with LLR information (below) to come to a different conclusion - that they've been delaying write-offs, which indicates worse to come.  The FFIEC numbers show that their charge-off rates have been in line with BPOPs over the past few quarters (they chose to take the majority of their write-downs in 2Q09 as opposed to steadily realizing impairments; this makes them seem more draconian within the moment). In fact, FBP has consistently taken fewer write-downs that BPOP on similar portfolios.

We compare FBP's portfolio to BPOP's.  As the largest two lenders on the island, they have very similar footprints, with similar expected loss profiles.  Matching FBP's book to BPOP's credit profile and approach to reserving, the numbers indicate that FBP should have about $800m in LLR, roughly double what they've reserved.  To be fair, BPOP's portfolio includes tenuous local credit card loans and as well as a larger concentration of US assets - which might cloud the numbers. However, a 2x differential in reserves for similar portfolios indicates something's amiss. Unlike most other banks, FBP does not release their 30-90 day delinquency data on their reports. We will have to wait for the 2Q09 FFIEC reports to come out for more data on impending NPAs.

Catalyst

We think most of the current mispricing is due to misconceptions about the value of their loan portfolio, the true earnings power of the company, and a lack of awareness of some of their impending non-loan writedowns.  These catalysts might provide datapoints to clarify these issues.  Anecdotally, we might see the buy/sell discrepancy to correct itself once the income-focused holders realize that the dividend check is no longer in the mail.

Regulatory action: Executives at another PR bank recently admitted that regulators had been lax about shutting down Puerto Rican operations despite the lack of tangible equity in the face of under collateralized mortgage and construction portfolios.  Alfredo Padilla, the commissioner, is skeptical about the banks' prospects, but regulatory authority lies with the NY Fed.  At an audit, we would expect regulators to look closely at the discrepancies we've outlined versus BPOP's reserve policies.  While the timing of these events are a function of Fed human resources and regulatory intent, it's likely that regulators will take a closer look at FBP's capital adequacy in the months to come, and that the current 4.73% TCE/risk-weighted assets (soon to be lower) will be a talking point.  Conversations with regional bank CEOs lead us to believe 6% is the targeted number in this category, which would have dilution implications.

FFIEC data on delinquencies (banks): Unlike other banks, FBP does not disclose their 30-90 day delinquencies on their quarterly reports.  Consequently, investors do not have the opportunity to examine impending NPAs until their call reports are released, about 45 days after the end of the quarter.  FBP recognized about $600m of impaired loans in the quarter (indicating potential non-recovery of principal), but it is unclear whether these loans are still current on their payments, especially taking into account the construction loan issue.

FFIEC data on capitalization (BHC): BHC has $82m in cash and had increased their investment in their Ponce $42m (from $123m to $165m).  Depending on how they are accounting for their reserves at the different subsidiaries, the banks might require further capital support, which would necessitate an equity injection.  In 1Q09, the BHC's only income was a $20m dividend from their Puerto Rican bank, as their other subs had stopped paying up.  The $400m TARP injection seems to have been injected directly into FirstBank Florida (where they've recognized most of their losses this quarter).  This might leave the FirstBank Puerto Rico undercapitalized once they're required to recognize construction losses. We expect more disclosure in the 2Q09 FR Y9C statements.

Loss recognition: We expect continued writedowns on impaired commercial and construction loans, with data points coming out in their quarterly filings.  Their 10K should offer more clarity on their accounting for the Lehman receivable and their deferred tax assets.

Regulators shutting down DRL: This will cause secured loans to come back on balance sheet, requiring impairment tests and consequent reserves.  FBP seems to have done this preemptively with RGFC, closing out a $205m commercial loan to the bank by taking in the pool of residential mortgages onto their own balance sheet, accompanied by a $33m reserve.

Funding issues: FBP might be unable to roll all of their $4.1b of repurchases due in 2009. They replaced $1.1b of the $4.5b that have expired so far this year with FED financing (at sub 1% cost) and FHLB, but only have $711m unpledged assets and $400m available from their FHLB line.  Additionally, they have $1.5b in outstanding commitments to extend credit, mostly to existing construction borrowers.  Brokered deposits are an expensive option for them, and these funding issues will lead to either capital injections, loss recognitions, or lower future income as margins narrow.