Banking: The finance view and the macro view
A recent post by Tyler Cowen discusses one possible reason why the Fed has refrained from setting a higher inflation target, despite the fact that many economists believe that doing so could boost AD. The argument is that higher near-term inflation expectations would raise short term nominal rates, and cut into bank profits that are now earned by borrowing short at very low rates and lending long at higher rates.
I can’t say whether Tyler Cowen is correct that worry about bank profits may be a factor discouraging the major central banks from doing additional monetary stimulus. But I haven’t seen any better explanations for their seemingly perverse behavior. And I should add that Tyler has mixed feelings about the desirability of such a policy:
I also regard this as a somewhat gruesome hypothesis. It means that “Main Street” is paying for “Wall Street” (forgive me the use of those awful terms) in at least two ways: high unemployment and inability to earn much on one’s savings. Risk on the Fed balance sheet is also paying some big part of the bill, since presumably that is helping to maintain the interest rate spread.
I’d say it’s even worse—it isn’t even clear that this strategy would help the banking industry. I believe this “finance view” misses the most important factor influencing bank profits–the state of the macroeconomy. Earlier this year the IMF lowered its estimate of the worldwide losses to banks from $4 trillion to $3.4 trillion. The explanation was a slightly better than expected recovery in the world economy after March 2009. The fact that a modestly better than expected macroeconomic outlook could shave $600 billion off expected banking losses is just one indication of how devastating the worldwide drop in AD was during late 2008 and early 2009. It sharply reduced all sorts of asset values and severely damaged the financial system. Indeed the damage was much worse than that from the earlier sub-prime fiasco.
If the sharp drop in AD had not occurred, i.e. if inflation had continued at its normal 2% to 3%, then the banking system would have survived the sub-prime crisis is much better shape. Yes, banks may gain, ceteris paribus, from lower short-term nominal rates. But in this case ceteris isn’t paribus. The very thing that drove nominal rates to near-zero levels is the same thing that caused the bulk of the financial crisis. And reversing that fall in AD would be a huge boon to the banking industry, even if nominal rates were higher as a result. Put simply, in most cases when X is bad for an industry, then “opposite of X” is good for that industry.
[BTW, interest rates are often more closely related to the level of NGDP relative to trend, than the rate of change in NGDP. So interest rates might remain fairly low even with a vigorous recovery in spending, at least until the economy got closer to full employment. This pattern occurred in the 1930s.]
Tags:
30. December 2009 at 18:43
Nifty new blog.
Nice that Cowen has finally caught up with reality – it only took 6 months.
http://baselinescenario.com/2009/05/06/is-everyone-confused-yet-bank-stress-tests/#comment-13203
“It seems to have become quite clear that the administration’s goal is to let big banks “earn their way out” by providing a favorable yield curve through low interest rates. The insidious thing about this is that it will work only so long as inflation does not return. If inflation returns, and this forces the fed to raise rates (commensurate to what the bond market is charging), the effect is synergistic – and we have the S&L version of this crisis (they are paying higher rates on short borrowing than they are earning on long lending).
Presumably, if that happens then assets should reflate in value (isn’t that what inflation means?). This would reduce loan defaults by increasing home/commercial land resale values. But the question is how fast bank loan losses are reduced by asset value reflation versus how fast bank earnings are reduced by unfavorable moves in the yield curve.
Real estate value reflation is a slower-moving phenomenon than bond prices; they are stickier. Bond prices project expected future rates instantaneously into the present. If inflation returns, therefore, it seems likely that we will have a “death valley” in which bank earning power will decline dramatically and asset prices will see a delay in catching up.
Thus, the Fed needs to reverse asset price declines _before_ inflation peeks through. Or massively recapitalize banks prior to losses (while it still seems that they are massively profitable on an operating basis).
As I’ve said before, the Fed is trying to thread a needle – I’ve wondered why, but it now seems like the commitment to allow banks to earn their way out is a plausible explanation.”
30. December 2009 at 21:02
Statsguy, That is possible, but I am not at all sure that interest rates would rise before real estate prices. In the Depression interest rates stayed low as long as the economy was depressed, even in 1933-34 when prices rose rapidly.
Both interest rates and asset prices fell in tandem in 2006-08, so I don’t see why they counldn’t rise in tandem. And of course (I think we both agree) there is much more than bank profits to worry about right now. Indeed I think we all three agree on that point.
Having said all that, it is a reasonable hypothesis; congratulations for getting there 6 months before Tyler.
30. December 2009 at 22:18
>Earlier this year the IMF lowered its estimate of the
>worldwide losses to banks from $4 trillion to $3.4
>trillion.
…
>If the sharp drop in AD had not occurred, i.e. if inflation >had continued at its normal 2% to 3%, then the banking >system would have survived the sub-prime crisis is much >better shape.
A few remarks to reinforce some of the finance/monetary links here…
As a slightly informed guess, the $600bn lower loss estimate is mostly comprised of a “write-up” of previously written down mark-to-market values of securities held. That is, these are paper gains on previous paper losses. The striking thing is that the write-downs have been massive multiples of any reasonable estimates of losses.[1]
If you look closely at the older IMF reports, their own projections of losses going forward on the loans held in portfolio (which aren’t marked-to-market) are smaller than the losses already embedded in the mark-to-market write-downs of comparable securities. I think the only plausible way to obtain the enormous loss estimates predicted by the decline in the value of securities is the macro deflationary view (i.e. the market is pricing in huge deflationary risk).
As this particular crisis manifest itself as a run on the shadow financial sector, which is outside regulatory oversight, you do have to wonder about the optimal Fed policy given the constraints on its mandate… I doubt the Fed could have done nearly enough with available short-term rate cuts. I also doubt that feasible levels of QE along the Treasury curve alone would have been enough. My reasoning is there was too much asymmetry in the exposure of market participants to the stigmatized sectors, giving enough incentive for those with liquidity to withhold from those in need for any plausible level of Treasury yield cuts. (Batten down the hatches, wait for competitors to go belly up and welcome a brave new world without them)
Thus, I don’t think the Fed had a choice but to heavily target its liquidity interventions onto the shadow sector, which it didn’t do nearly enough of (but try defending that on the Hill). Given those constraints, it should be no surprise we ended up with TARP. Perhaps gearing up its “extraordinary” programs earlier (i.e. those programs that got it into business with the shadow sector) would have helped a great deal. Anyway, I think this might be the real beauty of an NGDP futures targeting regime. Any and all activity that contributes to NGDP becomes a legitimate target of Fed policy. Ideally, the concept of a shadow financial sector would disappear (other than truly black markets) leaving just that pesky too big to fail problem to deal with…
[1] My favorite example of this anomaly from the crisis was the spike in prices of securities backed by federally guaranteed student loans, where despite almost zero risk of loss, these floating rate instruments were yielding a 100 or more basis points above LIBOR, which itself was a couple of hundred basis points above Treasuries.
31. December 2009 at 05:18
Just copying over my comment from the old blog:
Two problems with Tyler’s argument:
It assumes that negative real rates are not possible at the short end. In terms of just the funding cost of the banks, the Fed can clearly keep lending at whatever low rate they feel like.
As regards “market” interest rates, the low lending rates from the Fed to the banks can feed through partially to the market. Yes, I know the usual argument against negative real rates – hold a basket of commodities with no storage costs etc, but this just doesn’t hold given that there are storage costs and transaction costs and there is also uncertainty. There is no arbitrage one can enter into after taking into account costs that precludes negative real rates of upto say 2-3%.
In fact, the usual result of negative real rates is an increase in real asset prices – equities, real estate, gold etc but this is a risky strategy.
The other problem is that no one has shown me any proof that banks are generating all these profits by borrowing short and lending long. This was the strategy that the S&Ls got into so much trouble with initially in the early 80s and I doubt any bank does this now on a large scale. Most banks are reasonably maturity matched atleast on an interest rate risk basis i.e. they hedge out the interest rate risk due to maturity mismatch via interest rate swaps.
The real source of their profits is a combination of the negative real rates at the short end and the elevated credit spreads they’re earning on “safe” assets relative to pre-crisis levels. They may be earning extra credit spread by investing in longer maturity assets but the interest rate risk is usually hedged out.
31. December 2009 at 06:25
Can the banks really hold on to profits generated by Tyler’s mechanism? Competition between greedy banks should lead to most of their profits getting passed on to their consumers.
31. December 2009 at 06:56
“They may be earning extra credit spread by investing in longer maturity assets but the interest rate risk is usually hedged out.”
Do you have any data on this? Hedging the interest rate risk on the entire stockpile of 30 year loans sounds rather expensive. I suppose one mechanism for doing this is Fannie/Freddie – which have been buying up large lots of long term mortgage debt with Fed money. In that sense, I can’t say I disagree with the Fed program – if that means the Fed’s low rates get transferred to consumers, this is a reasonably equitable way to inject money into the economy (compared to other mechanisms).
It’s possible that the banks are primarily holding onto 5/1 ARMs, and much of the refinancing activity in the past year has focused on 5/1s because the rate spread for consumers is rather large. But I have no idea about the composition of bank balance sheets. So perhaps banks have shifted interest rate risk to consumers…
ssumner:
It’s not just that _I_ beat Cowen by 6 months… so did most of the rest of the world. The debate about the Fed’s excess concern for bank profits vs. concern for the macroeconomy has been vicious. So, welcome to the debate Tyler…
31. December 2009 at 07:09
These comments will appear under my name, but they are written by others. The author appears above each comment. I have moved them from the old blog, and will answer them later.
Jon
30. December 2009 at 15:05 (#)
Where is JKH. Seriously I’ve heard this theory before-its the closure that seeks to explain how modern policy is conducted in the absence of (meaningful) reserve requirements.
The basic idea is that FOMC can set the profitability of the banking system generally by manipulating the yield curve. When the FOMC wants a neutral policy, they flatten the yield curve (minimizing the rate of retained earnings). When the FOMC wants a very tight policy they invert the yield curve (causing banks to lose capital, and discourages investing new share issues of banks).
In this way, the creation of geared money is regulated.
bill woolsey
30. December 2009 at 15:27 (#)
Scott:
This is as bad as the theory that paying interest on reserves is a good idea because it enhances the profitability of banks.
Money and credit confused.
Anyway, for the theory to work, all the loss in real output must be due to an adverse productivity shock. This could be due to the loss of financial intermediation, shifts the the composition of demand, or even something unrelated to the crisis. What is most important is that increasing nomininal expnediture has no impact on output-only inflation.
rob
30. December 2009 at 22:47 (#)
I call shenanigans. No f*cking way. If this is what the Fed is really thinking then I am in favor of flying planes into tall building in Manhattan. For the record.
rob
31. December 2009 at 02:52 (#)
Scott, um, maybe you should delete my previous post considering the questionable state of the 1st amendment. thanks.
31. December 2009 at 07:54
Thruth, Thanks for the info on the IMF. There is one comment you made that I don’t agree with:
“I also doubt that feasible levels of QE along the Treasury curve alone would have been enough.”
They already did far more QE than necessary–they essentially doubled the monetary base. That is enough to double the price level. At that point all they needed to do was to set a price level or NGDP target, level targeting, and AD would have risen fast. Indeed they would have been able to get by with a much smaller QE if they had a more aggressive target.
Ash, Good comment. I agree with you that real rates can be negative, even ex ante. I will defer to your views on current banking practices. What you say sounds very plausible, but I am not knowledgeable enough to comment.
Tom P. Good point. I suppose the response would be based on the short run/long run distinction. But even in the short run banking is pretty competitive.
Statsguy, I probably should have paid more attention to the debate. I have been scratching my head all year trying to explain why the Fed was so perverse in its behavior. Perhaps it was merely special interest politics. Maybe I am naive, but when you see things like the Bernanke interview on 60 minutes, he really does seem to care about main street.
Jon, I had heard this theory mentioned in some of my comments, but admit to not paying close attention until I saw Tyler’s post.
Bill, I strongly agree. I really hope the Fed is not basing its policy on this idea.
rob, I censored it to make it more “child-friendly” but didn’t alter your political views. Just remember not to make this sort of wisecrack while waiting in a security line at the airport. 🙂
31. December 2009 at 08:21
“I can’t say whether Tyler Cowen is correct that worry about bank profits may be a factor discouraging the major central banks from doing additional monetary stimulus.”
Considering that our central bank is quite literally owned by the banks in the US, this theory doesn’t really strike me as unlikely.
31. December 2009 at 10:10
On the subject of bank asset-liability management, you could try reading this http://www.bos.frb.org/economic/neer/neer1995/neer195b.pdf or you could try this book if you have too much time to kill http://books.google.co.uk/books?id=5yDC2uyGCdMC&lpg=PA1&pg=PT9#v=onepage&q=&f=false .
My point is not that banks are always perfectly hedged out in terms of interest rate risk, just that they don’t just naively fund their trillion dollar balance sheets at the 3-month rate and invest their proceeds in 30 year mortgages. Again, this was a problem before the development of the interest rate derivatives market but not now. And even then, banks tried their best to minimise maturity mismatches.
To put it differently, if banks really did this, you’d expect them to lose huge amounts of money every time the Fed raised rates faster than market expectations.
Another paper by the BIS on the implications of corporate interest rate hedging on the usefulness of interest rate changes in monetary policy http://www.bis.org/publ/work95.pdf?noframes=1
The point that I’m trying to make is that interest rate changes just don’t have that much of an effect on either bank profitability or corporate profitability as they used to. We’re far away from the the 1950s when the entire banking sector simply held large amounts of long-dated investments outright, mostly govt bonds and the Fed could technically bankrupt the entire banking sector just by raising rates by 1% or so.
Now it has to resort to much more direct methods to impact banking sector profitability, primarily repoing assets at generous terms etc.
31. December 2009 at 10:25
But many think that policymakers were “heroes”!
http://www.ft.com/cms/s/0/a2e35910-f30a-11de-a888-00144feab49a.html
31. December 2009 at 11:56
“To put it differently, if banks really did this, you’d expect them to lose huge amounts of money every time the Fed raised rates faster than market expectations.”
Last I checked, they did.
31. December 2009 at 12:35
Count me as willing to ‘confess’ that I don’t entirely get Bill’s point.
31. December 2009 at 12:50
Doc Merlin – banks didn’t lose money because of the unhedged interest rate risk on their balance sheet. They lost money because the interest rate hikes tipped their consumer/mortgage lending customers into insolvency i.e. interest rate hikes led to a deterioration of the credit quality on their balance sheets. This is not what Tyler’s referring to.
31. December 2009 at 13:58
Ash, that’s a helpful paper but it lacks info on the magnitude (proportion) of assets that are insulated from rate risk, or the structure of rate risk insulation.
Banks may be insulated from extremes (for instance, they may have bought a floor on the rate spread between cost of funds and earnings on lending), but we may not be close to that floor – in which case an increase in the spread (which is very very high right now) would still be a huge gain to banks even as they’re suffering high rates of insolvency. And, in the current range where the spread is located, a decrease in spread may sharply decrease income.
In other words, even if all of the assets were shielded from rate risk (who has that data? the Fed? is it public?), it’s likely that banks are partly shielded from taking a _loss_; the contracts don’t guarantee a high level of profits. But the Fed needs the banks to actually be making a lot of money on spread to cover losses on their other risks (default, legal, etc.).
And that assumes all assets are insured… which is doubtful given the 10 trillion in mortgage debt outstanding in the US. I think you make a very good point, but it doesn’t completely undercut the argument.
1. January 2010 at 04:14
Statsguy – I’m not claiming that all interest rate risk is hedged out, just that it’s actively managed at the highest levels of most banks.
Regarding data, The Fed certainly has it but I doubt if its publicly available to anyone else.
If you want to read something else that illustrates bank asset liability management, try this HBS case study on Banc One in the early 90s which explains it quite well. http://sy.zlgc.org/Upload/20070918161219718.pdf
1. January 2010 at 19:17
Everyone, Due to lack of knowledge, I am going to refrain from taking a position on the question of Fe dpolicy and bank profits. I lean slightly in Ash’s direction, but mostly because after mid-2008 the deflationary fed policy clearly had a devastating impact on bank profits. But I also understand that at this point in the “recovery” things are a bit different from mid-2008, so I would blame anyone for thinking my reasoning was inadequate for current circumstances.