Description

I am in no way an expert on banks so I’d be grateful for thoughts on the framework as well as the idea.

Citigroup is a fairly standard bulge bracket bank which means that, in large part, it’s a black box.  Given that, I think it’s helpful to ask “how bad might things get?”.

Citi today has $644bln of net loans on its balance sheet and another $89bln of receivables from broker dealers.  So, call it $640-730bln of loans today.

In Q3 2007 (just before the banks acknowledged that things were going to shit), Citi had $761bln of net loans and $69bln of receivables from brokerages from ~$760-830bln of loans.

Then, the financial crisis happened and Citi, who had been taking around $8bln of provisions a year, took $80-85bln of excess provisions in the 4 fiscal years that followed.  Notably, Citi’s business was generating some $15-20bln of operating income so these higher-than-normal provisions led to a cumulative GAAP loss of ~$14bln over 4 years with losses in the first two years reaching $40bln (and the absolute largest cumulative loss over the period being about $49bln).

Coming back to today, the Fed stress test estimate for 2022 estimated loan loss rates of 6.4% under the severely adverse scenario.  GFC loan losses were 6.8%.  The stress test analysis estimates that Citi’s CET1 ratio would bottom out at 8.6% (more or less in the middle of its peer group: BofA 7.6%; JPMorgan 9.8%; Goldman 8.4%; Morgan Stanley 11.4%).  The Fed also estimates that Citi would have just $27bln of Net Income losses through Q1 ‘24 under the severely adverse scenario putting it right at the median for banks as a % of assets.

Stepping outside the fed framework, on the basis of TCE/Loans, Citi is in a much stronger position today than in Q3 2007:

Going back to Citi today, the company has $154bln of Tangible Common Equity and 1,942m shares outstanding putting the stock, at $44, at about 55% of TCE.

Over the last 10 years, the company has traded at an average of .88x TBVPS.  This includes periods of substantial market stress - memories of the GFC were still fresh in 2012, late 2015 there was a mini panic, obviously the market was dislocated in 2020, and the stock is basically at its 10-year P/TBVPS low in recent months which, I think, is where the opportunity lies.  That is to say the stock routinely traded above TBVPS when the market wasn’t in the dumps.

Over this period, ROTCE averaged 7% (or more like 8% excluding 2018 when the company had a large tax charge):

So let’s say the worst happens and the company takes, all at once, $80bln of impairments.  The stock would be trading at 1.15x TBVPS and would have 23% downside to its 10-year average P/TBVPS multiple.

But even this seems too punitive as Citi has fewer loans today than pre-GFC and the Fed estimates that a severely adverse scenario would result in lower loan losses today than in the GFC. Even beyond that, this isn’t what happened in the GFC when these losses were recognized over 4 years and the largest peak to trough drawdown was $49bln.  That is, we can either say these losses will be recognized over a few years and offset them by on-going normalized-post-provision profits or, as was also the case after the GFC, we can assume the company goes back to generating something close to its prevailing level of operating profits.

If, using the latter scenario, we assume Citi takes $80bln of impairments tomorrow and then generates ~$20bln of operating profits or ~$13bln of net income (against $74bln of TCE) going forward, we have a bank with a ~18% RotCE which would seem to deserve to trade above .88x.  If 1.5x is appropriate (it seems kinda low) that’s approximately 30% upside:

If, alternatively, we roll the losses out over 4 years we end up with some $20-30bln of losses and, at .88x this forward TCE of ~$129bln (current TCE less $25bln) the IRR is 8% but, again, I’d expect the forward RoTCE to be higher and thus deserving of a higher P/TBV multiple.

Now, of course, if Citi indicates it has $80bln of impairments to take the stock is going to trade down but the IRR should still be fine and, most importantly, I see no indication we’re in for anything like that.  The point is that even if we get a scenario that is materially worse than the Fed’s severely adverse scenario, Citi isn’t expensive by historic (or, really, absolute) measures.

This is all sort of a long way of saying: (i) I don’t think Citi has a ton of knockout risk; (ii) I don’t think Citi is materially over-earning on 2022 estimates; and (iii) that 4.7x trailing, 4.2x 2021 and 6.4x consensus 2022 earnings is far too cheap.

What’s it worth? I don't know but .88x book (10x earnings) seems reasonable enough and would result in 60% upside.  In the meantime we should get a 4.7% dividend and, while the buyback may be on hold for a few quarters to build regulatory capital, the company should again start repurchasing substantial amounts of stock in the next year or so.

Catalyst

Buyback?  Time?  Non-occurence of major recession?