I fell a little down the rabbit hole of this Journal article on “Low Rates Pummel Banks.” This has long been a mystery to me as there are two diametrically opposite narratives of banks and low rates in the world. In one, banks borrow for free from the Fed and reinvest at higher rates and print free money and spend it on parties where they beat up retirees. In the other, banks are, um, pummeled by low rates. Neither of those theories is absurd on its face, as you can tell from the Journal article:
Superlow U.S. interest rates are squeezing bank profits, complicating the industry’s nascent recovery from the financial crisis. … U.S. banks earned $114.39 billion last year, their best showing since 2006.1
Profits are squeezed, but to their best level since before the financial crisis, so.
The Journal’s thesis is basically that banks are getting screwed because net interest margin, specifically, is squeezed, as mortgage rates keep dropping, deposit rates are kind of floored and can’t drop any further, and other sources of revenue are drying up:
As higher-yielding loans and securities acquired before the crisis mature, the banks are forced to replace them with assets that carry much lower rates. With some sources of lucrative fee income such as debit card charges capped in 2010′s Dodd-Frank law, the margin squeeze has an outsize impact on the bottom line.
On the other hand, other other sources of revenue, like originating mortgages at record spreads to MBS rates, or just trading securities whose prices have been bolstered by declining rates, are still robust – but that business seems to be concentrated in big banks that also benefit from economies of scale. A Goldman equity research note yesterday came to the same place, noting that regional banks get ~58% of revenue from net interest while the big universal banks get only ~48% there:2
Given acute NIM pressure and only modest asset growth opportunities, we favor large-cap banks given 1) less reliance on spread income, and 2) greater offsets from expense / capital leverage.
The Journal gives us an example of one of the more pummeled regional banks:
Hudson City Bancorp, a Paramus, N.J., lender whose profits have been hammered by falling income on its large mortgage portfolio and which agreed in August to sell itself, said in a recent filing that the Fed’s moves “had an adverse effect” on the company, which has $43.6 billion in assets.
Here is the bulk of Hudson City’s balance sheet:
Other stylized facts about Hudson City:3
- those loans are essentially entirely mortgages – $28bn of first mortgage loans;
- those $13bn of mortgage-backed securities are essentially all agency pass-through certificates;
- it’s funded by $39bn of liabilities of which $24bn are interest-bearing deposits and the rest are other short-term things (repo, FHLB advances); and
- its 2Q2012 NIM was 2.12%, with average yield on assets of 4.09%, including 2.70% on MBS and 4.81% on mortgage loans.
So Hudson City is sort of a perfect stylized little home-loan bank; it is a machine for turning short-term deposits into mortgages, and it turns them roughly into 2/3 mortgages that it holds and 1/3 MBSes that it buys from Freddie and Fannie.
Here’s a chart from that Goldman note – this is average yields of various assets held by banks:
Note the “Resi” line – 4.32% in the second quarter, 4.26% in the third – and the “Securities” line, 2.97% and 2.85% respectively. This compares to Hudson City’s 4.81% on mortgage loans and 2.70% on MBS in the second quarter; Hudson City did a little worse than the big banks in its securities investing and quite a bit better in its mortgage investing. But of course it did much worse over all because it’s all NIM; the big banks are more than half other stuff that’s more profitable in a low rate environment.
Another thing to note, perhaps, is that at ~2/3 mortgages and ~1/3 MBS, Hudson City’s average yield was 4.09%. At all lend-and-hold mortgages, it would have been around 4.81%, for a 2.84% NIM – below Goldman’s regional-bank 3Q average (3.49%), though slightly above the large-cap bank average (2.81%). One thing worth pondering is why the Fed’s buying of all the agency MBSes in the world hasn’t pushed banks to do more (non-conforming) lending on their books instead of feeding conforming loans into the GSE machine and/or buying back GSE products at tiny tiny yields. Some of that is mentioned in the Journal article – “In its search for more yield, Wells Fargo said it kept nearly $10 billion of residential mortgages it would normally sell to investors.” Hudson, which was 2/3 its own mortgages and 1/3 GS machine, might have benefited from being even more lopsided, though of course capital regulations, liquidity needs, etc. militate toward holding the agency bonds instead of actual mortgages.
Yesterday we discussed S&P’s claim that the Volcker Rule might force it to lower big banks’ credit ratings as trading revenue dried up. One reaction you could have is “heh, that’s kind of weird, why would prop trading make you safer?”; I sort of had that reaction and so did Sallie Krawcheck, who tweeted: “S&P Warns of Volcker Rule Downgrades: implicitly assumes prop trading additive thru a cycle. Hhhmmm.”
One possible explanation is that too-big-to-fail-ness is marbled throughout financial decisionmaking; having a big trading operation both diversifies your revenues and makes them bigger and more tied up with everyone else, making you more indispensable to the financial system, which is good for your creditors even if for no one else. This profit-squeeze perplexity might add to that explanation: if you’ve got a big trading operation that is helped by low rates propping up asset prices, or for that matter a big securitization operation that is helped by high spreads between new mortgages and new mortgage-backed securities, then you stand a decent chance of doing well even in a low-rate environment.
On the other hand if you’re a wee little bank that borrows at 1%, lends at 4%, buys GSE pass-throughs at 2%, and generally sulks about how it’s getting pummeled, then the spreads that the big securitizing banks are making – which after all come out of your GSE yields – are not really a positive, and their trading revenues don’t do much for you. Until you give up and one of those large interconnected banks buys you.
Low Rates Pummel Banks [WSJ]
1. To complete the syllogism: rates were low in 2011! As low as they are now (though, QE3, whatever)! And not in 2006!
2. To wit:
3. From the 10-Q – loan breakdown page 15, MBS breakdown pages 12-13, liabilities breakdown page 5, NIM page 37, yields page 43.