Date: 11/23/2013 / Ticker: AIG-WT / Price: $20.50 / Market Cap: N/A / Idea Type: Special Situations
Apologies for not submitting any new write-ups in the last month. I seem to be suffering from a personal version of the “crowding out effect”, as prep work for the December CFA exam has visibly taken its toll on time allocated toward higher ROI activities (like writing this blog). I can’t help it; it’s an expensive exam.
Anyway, the case for AIG’s common equity (NYSE: AIG) has been laid out very well by some high-profile value investors, like this one: http://www.fairholmefunds.com/sites/default/files/CaseStudyII180612.pdf. The stock has been a bit of a hedge fund hotel in the last two years and has seen a nice rally, but the fundamental case remains intact:
1. AIG still trades at a meaningful discount to tangible book, which won’t last forever for a somewhat above-average insurance franchise.
2. Benmosche & company continue to make enormous progress improving operating metrics.
3. Much of the capital return story has yet to play out as the company remains overcapitalized.
You can find plenty of detailed write-ups on the underlying stock, so I’m not going to repeat what everyone else is saying here. This write-up will mainly focus on the AIG TARP warrants, which everybody knows about now but remain poorly understood. I believe the warrants provide a unique levered opportunity for investors to participate in the underlying equity. Under conservative assumptions, if held to maturity, the warrants can be reasonably worth 3 – 7 times their current price by the time they expire.
AIG’s warrants were issued in January of 2011 as part of the US Treasury’s TARP effort. Basically all of the too-big-to-fail institutions were forced to issue these warrants; each came with special terms. AIG was among the last to do the deal which means that its warrants will expire the latest. The special features of these TARP warrants are:
1. They are super long-dated. Most of the TARP warrants (including the bank warrants) expire ten years after issuance. The AIG warrants were issued in 2011 January and expire in 2021 January, so there is still seven years and two months left on them.
2. They are superior to long-dated call options, as the warrants come with special anti-dilution adjustments that protect the warrant holders against dividend payouts and a host of other potentially dilutive transactions (stock dividends, rights offerings, above-market tender offers).
3. I don’t think anyone has written anything about this effect so far, but the anti-dilution adjustments under some circumstances can even be accretive to warrant holders in that they may accrue more value than the notional value of a dividend when that dividend is paid. I will explain the math in detail.
The warrants are somewhat illiquid, because 1) there aren’t that much to trade around which discourages institutions from buying them 2) in the case of AIG a large % of the warrants are held by a few long-term investors whom I suspect will likely hold them to maturity so the effective float is much smaller than it seems. Bruce Berkowitz alone owns 24.3 million of these securities as of his most recent 13F reporting cycle which constitutes just a little under one third of all warrants outstanding.
The warrants have all the features of a long-dated call option except they come with protection (or even accretion) against dividend payouts and a number of other potentially dilutive transactions. The way I think of these instruments is that they are economically equivalent to a 10-year call option with a contingency value right attached. The CVR in this case is the anti-dilution adjustments which have the potential of making the warrants far more valuable than a comparable call options. Here is a summary sheet of what I estimate the CVR to be worth under each of my three scenarios:
The math involved in the calculation of anti-dilution adjustments is quite complicated. Below is the exact language on the subject from the warrants’ prospectus:
There is quite a bit of misinformation on this adjustment clause out there. A lot of people apparently believe that the strike price is adjusted down by the excess dividend over the $0.675 threshold amount (ex. The strike price will go down by $1 when a $1.675 dividend is paid). That is incorrect. In reality, the math is subject to two additional complications:
1. The adjustments will be made most likely on a quarterly basis as the dividends are paid. The dividend threshold per quarter will thus be 0.675/4 = $0.16875. Currently the stock pays a $0.10 quarterly dividend, which is lower than the minimum threshold so no adjustments are being made at the moment with respect to dividend payouts.
2. The adjustment is sensitive to the market price of the underlying stock on the record date. Since the stock will likely trade at different prices on all four quarterly record dates per year, the strike price adjustment downward will vary every quarter.
To illustrate how these complications affect the results, examine the following scenarios:
[Formula: New strike price = old strike price x (market price on record date – excess dividend over $0.16875) / market price on record date]
Assumptions: annual dividend = $1.675, quarterly dividend = $1.675/4 = $0.41875, current strike price = $45. Excess dividend = 0.41875 – 0.16875 = $0.25 per share
Scenario 1: The stock trades at $30/share on the record date, new strike price after dividend = 45 (30 – 0.25) / 30 = $44.625. Effective adjustment in this case = 45 – 44.625 = $0.375 per warrant
Scenario 2: The stock trades at $45/share on the record date, new strike price after dividend = 45 (45 – 0.25) / 45 = $44.750. Effective adjustment in this case = 45 – 44.750 = $0.250 per warrant
Scenario 3: The stock trades at $80/share on the record date, new strike price after dividend = 45 (80 – 0.25) / 80 = $44.859. Effective adjustment in this case = 45 – 44.859 = $0.141 per warrant
Notice the difference? The higher the stock price is, the lower the adjustment will be to the strike price. Notice under scenario 2 that the common (mis)perception that the strike price is adjusted by (excess dividend over 0.675) is only true under the rare circumstance in which the underlying stock trades at $45 per share on every record date before warrant expiration.
If you think that’s confusing, there is more: “Upon any adjustment in the exercise price, each Warrant will evidence the right to purchase the number of shares of Common Stock obtained by multiplying the number of shares of Common Stock purchasable immediately prior to the adjustment by the exercise price in effect immediately prior to the adjustment and dividing that product by the exercise price in effect after the adjustment. All anti-dilution adjustment calculations will be made to the nearest hundredth of a cent or 1/1,000th of a share, as applicable. No adjustment will be required if the calculation results in a change to the exercise price of less than ten cents; however, any such amount will be carried forward and applied in any subsequent adjustment of the exercise price.”
What this means is that whenever a dividend (greater than $0.16875 per quarter) is paid, not only does the strike price get adjusted downward, the investor will also receive rights to purchase additional shares for each warrant held, for free!
Let’s go back to our previous three scenarios to see how this will affect the warrant value:
[Formula: New no. of shares per warrant = Old no. of shares per warrant x old strike price / new strike price]
Assumptions: quarterly dividend = $1.675/4 = $0.41875, current strike price = $45, current no. of shares per warrant = 1.000
Scenario 1: The stock trades at $30/share on the record date, new strike price after dividend = 45 (30 – 0.25) / 30 = $44.625. New no. of shares per warrant = 1.0 x 45 / 44.625 = 1.008
Scenario 2: The stock trades at $45/share on the record date, new strike price after dividend = 45 (45 – 0.25) / 45 = $44.750. New no. of shares per warrant = 1.0 x 45 / 44.750 = 1.006
Scenario 3: The stock trades at $80/share on the record date, new strike price after dividend = 45 (80 – 0.25) / 80 = $44.859. New no. of shares per warrant = 1.0 x 45 / 44.859 = 1.003
The effects here are similar to those highlighted previously; the higher the stock price, the fewer additional shares (after adjustments) you will receive for each warrant you own.
Combining the two effects, you end up with an adjustment schedule that is very sensitive to the market price of the stock, which has a strong inverse relationship with the value of the anti-dilution adjustments – a lower stock price makes the adjustments more accretive.
If this mechanism reminds you of a stock repurchase program, damn right, because that’s what it’s trying to replicate (the end effects will be somewhat different due to other moving parts).
Therefore, if you are dead serious about holding the warrants to maturity, you want the stock price to be as low as possible in the years prior to expiration for the strike price to be adjusted down by a large margin, and then go to full value toward the end of the holding period.
Valuing the warrants is going to be quite difficult given how many moving parts there are. Some of the key variables are:
1. Growth in the company’s tangible book value, which is determined by future ROE. Management’s targets for 2015 are > 10% ROE, assuming achievement of certain operating metrics (combined ratio in the low to mid 90s for the P&C business). AIG is currently doing around 7% – 8% ROE, so 8% by 2015 is a reasonable low case projection in a relatively normal big cat environment. I assume 7% ROE into 2015 and will use 10% ROE (base case) and 12% (high case) for run-rate ROE after 2015.
As a reference, AIG was doing in excess of 14% before the crisis, and most of its P&C and life peers are doing over 10%.
2. Capital allocation – here I assume a 40% dividend payout rate from 2014 and onward, and 40% of the earnings going into stock buybacks. AIG is currently far overcapitalized relative to its peers and management has indicated that significant capital returns to shareholders will be inevitable.
3. Future stock price (for anti-dilution adjustment calculations) – I project 12% annual appreciation in my low case, 14% in base case and 16% in high case. Recall that the higher the stock goes, the less accretive the anti-dilution adjustments will be. I think these numbers are fairly conservative.
4. P/B multiple when the warrants expire – below are my numbers, which are reasonably consistent with the current peer multiples.
Here are my models:
Depending on the future operating performance and capital allocation policy of the company, the range of potential outcomes is very large, which is reasonable to expect given the levered nature of the warrants. Another thing to keep in mind is that I am valuing these securities the old-fashioned way – held to maturity and valued on the basis of the intrinsic value of the underlying equity. If the underlying reaches fair value before the warrants expire, the IRR can be even higher than what I projected. With an option so long-dated, don’t even kid yourself with Black-Scholes; it’s useless.
Another interesting result is that the anti-dilution adjustments in two of the three cases turn out to be not only protective against “dividend dilution”, but also “accretive” to the warrant holders in that the values of their anti-dilution adjustments are higher than the future value of the actual dividends (compounded at 10%).
This is mainly a result of what I mentioned earlier – that the anti-dilution adjustments in fact resemble a share repurchase program. Accretion is realized when the dividends are paid and adjusted while the shares trade below their intrinsic value.
Warrants vs. Common Stock
I think the warrants are currently a better way to gain exposure to AIG than the common stock.
The warrants trade at $20.50 a piece. The “intrinsic value” of the warrant is $49.50 (stock price) – $45 = $4.50. The difference between the warrant price and the intrinsic value is $16.00 a share, so effectively you are paying $16.00 in interest for a non-resource loan to borrow $45.00 for seven years and two months.
That’s an implied borrowing cost of 3.96%! For 1% above the risk-free rate, you get not only the levered upside to the underlying, but also protection against losses exceeding $20.50. If you factor in any potential value accretion from the anti-dilution effects, the effective borrowing cost will be even lower than 3.96%. What a deal!
Warrants vs. Call Options
The anti-dilution adjustments can be very valuable even under conservative assumptions. However, even if we ignore the fundamental case for the stock, the warrants appear quite mispriced relative to traditional call options. Consider:
1. AIG’s 2014 January $45 calls currently sell for an implied borrowing cost of 6% (and you only get to borrow for a year and 2 months).
2. Black-Scholes gets you $23.3 for a 7.2 year call options (I used a 3 year average vol. of 36.5%)
This may sound strange, but I believe the mispricing of the warrants (relative to call options) will be corrected no later than October 14th of 2018, roughly 27 months before the warrants expire. This is because AIG dealers set up 2-year LEAPS on the stock usually 27 months before maturity dates. The most recently available version expires in January 16th, 2016 and became available on October 14th of 2013.
The TARP warrants will expire on January 19th of 2021. Should the warrants stay mispriced, one can simply arbitrage the difference on October 14th of 2018, when the new January 16th 2021 LEAPS gets introduced, by shorting the LEAPS and going long the warrants. Keep in mind that the warrant is simply superior to a call with the same expiration and should always trade at a higher premium.
Hartford Financial Services (HIG-WT) Warrants, $9.79 strike, 2019 June – the fundamental case here is similar to AIG: a tainted but nonetheless solid insurance franchise that trades at 80% of tangible book. The HIG warrants here provide some fairly generous anti-dilution clauses, as the dividend threshold is only $0.05 on a $35 stock. The company already pays more than that so the anti-dilution adjustments have already kicked in. The current strike is around the low $9s.
The stock is at $35.56, so the intrinsic value of the warrant is $26.26. At $26.77 a piece, you are basically paying nothing for the time value of holding the warrants. Accounting for further dividend adjustments, the effective borrowing cost on these warrants will likely be negative for the holding period.
Wells Fargo (WFC-WT) Warrants, $34.01 strike, 2018 October – This one is unique. Wells Fargo is a much better quality business than AIG or the other big banks. Make a table of WFC’s ROA/ROE for the last 15 years and you will see what I mean. The stock trades at a 65% premium over tangible book but given how much faster it will compound its book value than its peers the premium here is probably justifiable (before the crisis WFC regularly traded for 2.5 times book which still didn’t deter Warren Buffett from holding a huge stake).
The intrinsic value of the WFC warrant is $10.35 and the warrants trade at $15.81. You are laying down $5.46 a share in interest to borrow $34 for 5 years, so the effective borrowing cost is 3% before dividend adjustments. The WFC warrants, though, have only five years left before expiration.
General Motors (GM-WTA, GM-WTB) Warrants, strike prices vary, expiration dates vary – GM is the new hedge fund hotel nowadays. I don’t know enough about car sales and labor restructuring to form an informed opinion about the stock itself but the warrants seem interesting from a pure valuation perspective.
The common is at $37.63 a share.
A series (expires July 2016) – trades at $28.01, strike price $10, so you are paying nothing for the time value here.
B series (expires July 2019) – trades at $20.25, strike price $18.33, also close to nothing for the time value. The funny thing here is that GM calls expiring in January 2015 with an $18 strike price is trading at $21.03 a piece. WHAT! An inferior option expiring 4.5 years earlier is more expensive?
There’s a finance professor of mine who has been a staunch proponent of the strong form of market efficiency. I am tempted to show this to him.
The author of this write-up owns shares in the company mentioned (NYSE: AIG-WT) and may purchase or sell shares without notice. This write-up represents only the author’s personal opinions and is not a recommendation to buy or sell a security. No information presented in the write-up is designed to be timely and accurate and should be used only for informational purposes. Readers of the write-up should perform their own due diligence before making investment decisions.