At December 31, 1998, when bank stock valuations were at or near their all-time highs, an investor could have bought shares in Wells Fargo (WFC) at a 4.95x price/tangible book value per share ("P/TBV-PS") multiple, US Bancorp (USB) at a 10.06x multiple, or JPMorgan Chase (JPM) at a 2.87x multiple.
Which stock looked like the safest bet?
Fast forward to last Friday. Each bank's P/TBV-PS multiple has contracted materially, WFC's from 4.95x to 1.72x, USB's from 10.06x to 2.73x and JPM's from 2.87x to 1.06x. Had you bought shares in each at the end of 1998, you would have earned the following cumulative total returns: WFC 144%, USB 66% and JPM 20%, or annually compounded returns of 6.8%, 3.8% and 1.3%, respectively.
Even though all three banks experienced severe multiple compression, especially USB, JPM looked "cheapest" at the end of 1998, and yet ended up delivering the worst return So WFC and USB both did something operationally to overcome their multiple compression that JPM could not or did not do.
What are the reasons for the operational gap?
Some reasons are independent of managerial influence, and others are a direct result of it. First, JPM's net interest margin ("NIM") is considerably lower than that of WFC or USB. Average NIM for the three institutions from 1999 to 2011 were 2.48%, 4.89% and 3.99%, respectively. JPM has a large securities portfolio, which it funds with wholesale borrowings, and this makes the NIM gap impossible for JPM to close. Sure, changes in the yield curve that began late in 2007 helped the NIM of JPM while hurting that of WFC and USB, but these changes are not permanent. They will revert, partially if not completely, when the economy recovers, likely to JPM's detriment. JPM's trading income and other forms of income aren't high enough to compensate for its low NIM, and as a result, JPM's return on average assets ("RoAA") is well below that of WFC and USB; average RoAA from 1999 to 2007 (I exclude subsequent years because of abnormally high loan loss provisions) were 0.78%, 1.63% and 1.82%, respectively.
All other things being equal, I'm willing to pay a higher P/TBV-PS multiple for a high RoAA bank than for a low one. That does not mean JPM should exit businesses that depress its RoAA. It just means that JPM might have to accept a more modest valuation multiple. Or can it do something?
In reality, all other things aren't equal. Shareholders ultimately care about return on average tangible common equity ("RoATCE"). RoAA is one input to this, and leverage, or tangible common equity/tangible assets ("TCE/TA"), is the other. Within limits, leverage is under management control. JPM is more aggressively leveraged (has a lower TCE/TA ratio) than either WFC or USB. At Q2 2012, JPM's, WFC's and USB's TCE/TA ratios were 5.95%, 7.99% and 6.70%, respectively. By way of comparison, a 1.00% RoAA for a bank with a TCE/TA of 7.99% translates into an RoATCE of 12.5%, but with a TCE/TA of 5.95%, this 1.00% RoAA translates into an RoATCE of 16.8%. Leverage is hugely important to how RoAA translates into RoATCE.
JPM's more aggressive leverage helps JPM's RoATCE. Even so, from 1999 to 2007, average RoATCEs for JPM, WFC and USB were 17.0%, 28.5% and 33.0%, respectively. JPM's RoATCE lags considerably. Using median sell-side EPS estimates, and assuming no share buybacks or material changes to dividend payouts, I estimate Q4 2013 RoATCEs for JPM, WFC and USB of 13%, 17% and 21%, respectively. JPM's projected RoATCE gap is narrower, but only because of WFC's and USB's reduced RoATCEs.
JPM has argued that it isn't riskier despite its relatively low TCE/TA, and that its large quantity of low-yield, low-risk assets makes regulatory capital ratios a better measure of its capital adequacy than stated capital ratios. I agree. JPM's, WFC's and USB's Tier 1 common ratios at Q2 2012 were 9.87%, 10.08% and 8.78%, respectively. JPM is slightly below WFC on this basis, and meaningfully above USB. But that's not the point. The point is that JPM's RoATCE lags even with the addition of all the low-risk assets.
When JPM CEO Jamie Dimon expresses frustration over increasing regulatory capital requirements, low RoATCE is why. These requirements are likely to force JPM's RoATCE even lower.
I suspect that low RoATCE at least partially motivated the speculative bets that JPM's Chief Investment Office (CIO) has made in recent years, bets that according to a recent article in the New York Times supplied $4 billion of profits over the last three years, equating to 10% of JPM's total profits. Then in Q2 2012, the CIO lost $4.4 billion pre-tax, or $2.6 billion after tax, assuming a 40% tax rate.
I'm perplexed by what edge JPM's CIO thinks it has in deploying the $350 billion it is charged with investing. Who is on the other side of JPM's trades: individuals, companies outside the financial sector, or equally (perhaps more) sophisticated investment banks and asset managers? To use a recently popularized term, who is the "muppet"?
Can asset growth save the day? Generally, the more rapidly a bank grows assets, the lower the RoATCE is on the "new" assets. When RoATCE dips below the estimated cost of equity capital, additional growth actually destroys value. Bank equity discount rates are currently in the neighborhood of 10%. This is low relative to where they've been historically (14-15%), because of how low current long-term Treasury yields are. If JPM can grow assets while maintaining the projected RoATCE of about 13%, growth will be valuable, as long as discount rates do not rise above 13%. WFC's and USB's RoATCE's are higher, so their growth is that much more valuable.
What is the historical asset growth record for these three banks? Since 1998, JPM grew its assets per share ("A-PS") at a compound annual growth rate ("CAGR") of 5.6%, versus 11.0% for WFC and 9.0% for USB. The period from 1999 through 2005 was especially bad for JPM on this front, partially but not completely because of the pricey acquisition of Bank One. Acquisitions don't necessarily supplement organic asset growth. If we recalculate the CAGRs from Q4 2005 to present, we get 9.0%, 9.1% and 7.7%, respectively. Of course, JPM's and WFC's A-PS were both helped by large acquisitions of troubled institutions announced late in 2008; such future growth opportunities are unlikely.
I wouldn't expect JPM's A-PS growth to materially lag WFC's and USB's in the future. But I would expect long-term future A-PS growth for each bank to be modest, on the order of 5-6% per year.
On Friday, JPM closed at $37.17, with a P/TBV-PS multiple of 1.06x. This seems like too low a multiple, today. I think a more appropriate one is 1.3x, which would give a "fair value" estimate of $48. While this implies that there's easy money to be made, I wouldn't be so sure. The headwinds are strong. When the economy recovers and interest rates rise, JPM's NIM will suffer. Therefore, its ROATCE will suffer, even if JPM doesn't need to reduce its leverage, which it might have to do. Lastly, the applicable discount rate will be higher. The tailwinds? Investors may assume that the robust asset growth accompanying the economic recovery will extend well into the future, and accordingly award JPM a high valuation multiple.
According to a recent article in Bloomberg, in Q2 2012, Warren Buffett increased his ownership of WFC by 9.7 million shares, to 269 million shares, a stake worth over $9.2 billion as of last Friday. WFC traded on average at about 1.70x TBV-PS during the quarter. You have to ask yourself why an investor as astute as Buffett hasn't chosen to load up on JPM shares, if they're so clearly "cheap".
The praise and financial rewards that Dimon and the JPM management team received for avoiding the disasters that befell Citigroup (C) and Bank of America (BAC) were well-deserved. But if JPM wants to tout itself as "best in breed", C and BAC, both of which are still working through their problems (and both of which arguably look cheap today, at P/TBV-PS multiples of 0.57x and 0.62x, respectively), are not the right comparables. WFC and USB are.
For the reasons I've articulated, JPM won't be able to match the operating performance of either WFC or USB. But can JPM keep improving, or has it reached a wall? Because if it can't improve, if a 13-14% ROATCE is all JPM can do, and this 13-14% hinges on speculative activities and requires large share buybacks, then maybe pursuing growth isn't what's best for shareholders. Maybe low growth and high dividends is best.
JPM's all-time high closing price of $65.67 occurred on March 23, 2000, over twelve years ago, a point at which JPM was 16% of its current asset size. Let's see how long it takes to get back there.