Description

I’m recommending a long position in the AIG Warrants as an inexpensive way to buy the growth in AIG’s underlying TBV through the end of 2020.  At expiration (Jan 2021), I believe the warrants will ultimately be worth ~$69 / share, or 3.6x MOIC, offering a 32% IRR. 

The warrants have been written up now 4x on VIC.  The first was by culyer in 2011, and again by jon64 in 2013, and most recently by stanley339 in early 2015.  All of the write-ups were solid and are worth your time to get more background information.

There are three new facts though to add to the story, which I believe justify both a new write up and additional consideration if you’ve spent time on this in the past:

1)     1)  The warrants have MASSIVELY underperformed the stock since 2014; they have never been “cheaper”

2)     2)  For the first time ever, the company has begun buying them back in addition to the common (repurchased 10m shares in the first quarter, prior to that there were only 75m outstanding)

3)     3)  AIG management now has made a ROE improvement plan public (whether you believe it to be credible or not), and there are two activists on the BoD

SUMMARY  

AIG management has laid out a plan to restructure the business over the next three years.  As a result, they expect to improve the ROE to ~9% by 2018 vs ~7% last year.  Many are skeptical of their ability to deliver these results without a tailwind from interest rates.  If they do, however, TBV ex AOCI & DTA should grow from $58.50 / share today to nearly $110 / share by the end of 2020 (the expiration of the warrants).

By that time, the strike price on the warrants should decline to just over $43 / share, and the underlying number of shares exercisable per warrant should increase from 1.006 to 1.045 (more detail on this math below).  Assuming AIGs future multiple doesn’t change (0.98x TBV ex AOCI and DTA), the warrants are worth almost $70 / share at the end of 2020.  

As I mentioned above, many are skeptical that they will be able to realize anything close to what they have represented in their restructuring plan.  To point, consensus is materially below management’s implied guidance, though they do seem to get closer in the out years.

If the business structurally under earns, and the ‘exit multiple’ is meaningfully lower than the current one, I still think you make money owning the warrants. 

I dropped the consolidated ROE to just 6% and the TBV valuation to 0.7x (just about the lowest we’ve seen outside of the crisis), and you still get a 7% IRR.

 WARRANTS OUTPERFORMED IN 2014

I created these great graphs showing AIG vs the Warrants, which you can download here...

WARRANTS HAVE BEEN TERRIBLE SINCE

Again, some really pretty graphs on the download if you care...

AIG common is up slightly since the start of 2015, and the warrants are down ~25%.  The underperformance is even worse when you account for the reduction in the strike (-27c per AIG share) and an increase in the number of underlying AIG shares per warrant (.006 incremental shares of AIG).

I like best to express this as a stub security – the premium you’re paying for the time value in the option:

(AIG/WS + Strike Price – AIG) / Underlying Share Multiplier = Time Premium Stub

This stub is down to $5.70 today after being up as high as $18 just 18 months ago.  The last time AIG/WS was written up on VIC, the stub was ~$10. 

Because of the anti-dilution provisions for the warrants, if the price of AIG common doesn’t change and the dividend never grows, the “time premium” will drop to $3.40 at expiration (18% downside if AIG is flat after 4.5 years).

What this also says is that by buying a warrant today, you essentially get 100% of the upside in AIG above $63 / share ($57.40 at market plus the $5.70 time premium), but only have to put up 32% of the capital ($18.50 / $57.40). 

There is an arbitrage opportunity in here somewhere for someone smarter than me.

I think the reason for the underperformance of the warrants vs the stock are twofold:

1)      1) Fairholme is selling, and is the largest shareholder at 16% (this will end at some point)

2)      2) AIG has been buying back common hand over fist, giving shares a perpetual bid (they are now buying back warrants)

BUYING BACK WARRANTS

In the first quarter the company repurchased 10m warrants at $17.30 / share.  A very small number relative to the $3.5 billion they spent buying back stock. 

Through the end of 2017, management has committed to returning $25 billion to shareholders, primarily though share repurchases.  For those keeping score, that is 1/3 of the current market cap.

The warrants are a tiny piece of the capital structure, 65 million in total at $18.50 = $1.2 billion.  My guess is they will have a bid underneath them as well (and hopefully rocket once Fairholme is done selling).

MANAGEMENT’S PLAN

The presentation of the plan can be found here.

The plan is pretty simple in its presentation, but no doubt will be much more complex to execute.  The goal is to improve the ROE to north of 10% for the operating units, or roughly 9% on a consolidated basis.  The ROE will be based on Tangible Book, excluding Accumulated Other Comprehensive Income (i.e. gains / losses on investment securities that will not be realized) and excluding Deferred Tax Assets (which are substantial, but should be largely worked through by the end of warrant life).

They are approaching it three ways:

1)       1) Shrink to grow; divest assets and buy back $25B of stock

2)      2) Reorganize the business units to improve transparency / show progress

3)       3) Cut costs ($1.6B) and improve the Property & Casualty Accident Year Loss Ratio by 6 points

Of those three buckets, the most controversial by a mile is the AYLR improvement as the others are largely in management’s control (though the cost cuts need to be managed to be sure they don’t affect revenue / profitability adversely).

Here is a history of the adjusted combined ratio and prior year development (i.e. charges) by quarter for the combined P&C business:

The number has been improving, but much work needs to be done.  Notice the $3B charge in the fourth quarter of 2015 – that will hopefully be a nice tailwind for the AYLR going forward.

There have been several research reports going into substantial detail about these targets – whether or not they are achievable, and if so, wondering if AIG is going about it the right way.

I’m not an insurance analyst so I’ll leave the debate to the experts, but from my perspective, I’m not paying for any improvement in the underlying business.  In fact, I will make money on the warrants even if there is material deterioration in results from here.  Anything they can do to improve efficiency is all upside.

I would also note that there are now shareholders in the boardroom (as opposed to Government proxies).  At a minimum I believe that increases the odds of improvement / success.

ANTI DILUTION

This hasn’t been an issue for the prior write-ups, but now that the anti-dilution provisions of the warrants are in effect, I feel like it’s worth discussing.

Per the warrant registration, the warrants have certain rights against dilution from dividends, spin offs, etc.  Specifically, when the cumulative dividends are greater than 67.5c over a single year, the warrant strike and underlying shares get adjusted.  The adjustment math is:

Trailing Twelve Month Dividend ($1.01 at 3/31) minus the Dividend Threshold (67.5c fixed) = TTM Adjustment Factor (33c at 3/31)

TTM Adjustment Factor (33c at 3/31) – Prior Three Quarters TTM Adjustment Factors (13.5c at 3/31) = Quarterly Adjustment Factor (19.5c at 3/31)

New Strike Price ($44.73) = Prior Strike Price ($44.90 at 3/31) X [AIG stock price the day before the dividend ($51.59) – Quarterly Adjustment Factor (20c at 3/31)] / AIG stock price the day before the dividend ($51.59)

AND there is an adjustment to the # of underlying shares:

Prior Underlying Shares (75 million at March dividend) X [Original Strike Price ($45 fixed) / Adjusted Strike Price ($44.73 at 3/31)] = Updated Underlying Shares (75.446 million)

Updated Underlying Shares (75.446 million) / Prior Underlying Shares (75 million) = Per Warrant Underlying Share Adjustment (1.006 at 3/31).

It’s more complex than it probably should be, but whatever.  The important thing to note is that the share price of AIG matters for this provision – the lower the price of the common, the lower your strike price when the anti-dilution sets in, and the more underlying shares you are able to exercise into.  Not a total game changer, but a nice kiss in the event that the stock doesn’t go higher.

With the dividend coming in three days, the updated Strike Price and Underlying shares should be ~$44.58 and ~1.009 respectively. 

RISKS

The risk is that management blows you up.  There is always the risk in insurance (and this is long tailed insurance) that the book is awful and not to be trusted, which has been the Bear Case forever on the name.  It’s an easy Bear view to have because we know Greenberg was extremely aggressive, we know that they have already taken massive charges, and we know the tails are very long.  It’s also easy because you can’t really prove that it’s wrong.  AIG is and will most likely always be a black box.

I don’t consider a future of under-earning to be a material risk here as I think that is already priced into the warrants (as discussed above).  Dead money though is a risk in that you have a time premium in the warrants that could expire worthless.

Catalyst

CATALYSTS

The nearest term material catalyst is continued repurchase of warrants; followed by the completion of Fairholme’s selling program.

Earnings are a catalyst here in that any improvement / signs of progress in the business will increase the odds of the best case.