How The "Taper Tantrum" Cost US Banks $25 Billion In Q2 Net Income
Ever since back in 2009 the US financial system effectively suspended Mark-To-Market accounting for the too big to fail (and all other) banks (read our description of FAS 157 and 115 here), the Income Statement impact (i.e., net income above the [Other Comprehensive Income] line) of wild interest rates moves on bank balance sheets has been one thing US banks have not had to worry about. The reason for this is the transformation of large swaths of rate-sensitive holdings (the vast majority of bank assets) on the balance sheet to “Held to Maturity” meaning no matter how much higher or lower interest rates moved, banks would be immune from flowing Mark to Market losses (or gains) through the income statement.
Yet despite best effort to immunize banks from rate swings and debt MTM risk, a substantial amount of duration exposure has remained with the glorified hedge funds known as FDIC-insured bank holdings companies under the designation of “Available For Sale” (AFS) or those which due to their explicit short-term trading fate, would have to be subject to mark to market moves.
It is the bottom line impact of these securities that threatens to crush bank earnings in the just concluded second quarter by an amount that could be as large as $25 (or more) billion.
As an aside, it is technically not true that banks are devoid of GAAP rate and duration risk: over the past four years, banks had found themselves in the paradoxical situation where blowing out spreads impacting bank credit instruments (debt and CDS) due to systemic or industry-specific rate shifts, actually resulted in a boost to the adjusted bottom line since the so-called DVA impact had to be netted out from pro forma net income, leading to a non-GAAP EPS bonanza when things got really rough.
However, a far bigger issue is what happens to banks in a time when rates surge violently and dramatically in a short period of time. Such as in the past two weeks.
As we previously explained, in places like Japan, the rapid blow out in yields (with the 10Y hitting 1% coinciding with the peak of the Nikkei 225 so far in 2013) has a massive adverse impact on banks, and where a 100% parallel shift in the curve may lead to as much as a 35% impairment in bank capital.
But how about the US?
After all, while the Fed is the single largest holder of Treasurys (and whose P&L suffered a massive $200+ billion, or 4 times its “capital”, loss in June alone) courtesy of the ineffable “faith” in fiat, Bernanke has no Mark to Market restrictions (at least until such time as the reserve currency status of the USD is questioned, and so is the faith of the US central bank but that is the topic for another article) US banks do have substantial duration exposure to the tune of hundreds of billions, primarily concentrated in the spring-loaded clip known as Available For Sale securities.
So just how large is said duration exposure? According to JPM’s Nikolaos Panigirtzoglou it’s quite substantial, and amounts to a whopping $30 billion in MTM losses for every 1% increase in yields. Or in other words, the Q2 blow out in the 10 Year should have resulted in a $20-30 billion loss to the US financial system’s bottom line (through the Accumulated Other Comprehensive Income line) in the second quarter alone (all else equal, and for a hyper-levered system in which everything is contingent on smaller and smaller rates all else is never equal implying many more adverse downstream effects will likely be revealed).
How large is the duration risk of banks’ portfolios? The Fed's H.8 release provides some guidance on this for US banks. Each week, the Fed reports the net unrealized gain on banks’ available for sale securities. This is by no means a complete measure of banks’ duration exposure. It does not include held to maturity portfolios (albeit these are much smaller than AFS portfolios), and importantly does not include the impact of any swap hedges. Also, not all bank branches may mark their portfolios to market every day or week.
All that said, unrealized gains had fallen from $37bn at the start of May to $15bn by Wednesday of last week.
We can infer banks’ duration exposure by relating week-to-week changes in unrealized gains to week-to-week changes in the yield of our US aggregate bond index, the GABI. Figure 3 shows this beta, estimated over a rolling 6-month window to smooth through the noise in the estimate. It suggests that banks would make MTM losses on their available for sale portfolios of just under $30bn for each 1% increase in yields.
Estimating the beta over three months would give very similar results.
That beta is around twice as large as in 2010, whereas the size of banks’ bond portfolios has increased by only around 20% over that period, implying that banks in aggregate have been shifting towards longer-maturity securities. The beta of just under $30bn for a 1% rise in yields implies that the 18bp rise in GABI yields since the last H8 report would have reduced unrealized gains by a further $5.4bn.
Oops: this is precisely why when everyone is scrambling to chase yields and in the process increases duration to preserve some NIM in a centrally-planned, manipulated, collapsing rate environment (resulting in a doubling of bank duration exposure beta in under three years!) any rapid inverse move will leave everyone with massive losses. Losses that may be as much as $30 billion or more.
Putting this number in perspective, according to the FDIC, in Q1 banks recorded profits of $40 billion. There goes half of Q1 profits...
But while the US losses may be manageable, if crushing to sellside analysts who have bet the farm – once again incorrectly - on S&P 500 earnings picking up in Q2 due to a surge in financial profits which are now locked far lower at June 28th 10Y levels and resulting in billions in MTM losses that will have to hit the Net Income line, it is things in Europe that are about to get nasty once again.
US banks have been modestly reducing their bond holdings since the start of May, by around $19bn overall. By contrast, Euro area banks bought a hefty €48bn in May, largely because Italian and Spanish banks added to their holdings as spreads edged wider.
And while we don't have a convenient weekly update on banking sector unrealized losses in Europe as per the US H.8., it looks like the European periphery’s Monte Carlo double-down “all in” bluff may have just been called. Or in other words, now that the carry trade tide has gone away, we finally see how many European banks were swimming naked (for a hint from Goldman Sachs, see here). We can’t help but wonder how many of their US brethren will join them with their pants down as bank results are reported over the next several weeks.