Interest On Reserves, Part III

Ben Bernanke, financial crisis, interest on reserves, Lehman Brothers, TED Spread
Emily Levin, harpist:

HarpistWhy do I keep harping on interest on reserves?  Because, IMHO, the Fed's decision to start paying interest on reserves contributed at least as much as the failure of Lehman Brothers or any previous event did to the liquidity crunch of 2008:Q4, which led to a  deepening of the recession that had begun in December 2007.

That the liquidity crunch marked a turning point in the crisis is itself generally accepted.  Bernanke himself (The Courage to Act, pp. 399ff.) thinks so,  comparing the crunch to the monetary collapse of the early 1930s, while stating that the chief difference between them is that the more recent one involved, not a withdrawal of retail funding by panicking depositors, but the "freezing up" of short-term, wholesale bank funding.  Between late 2006 and late 2008, Bernanke observes, such funding fell from $5.6 trillion to $4.5 trillion (p. 403).  That banks altogether ceased lending to one another was, he notes, especially significant (p. 405).  The decline in lending on the federal funds market alone accounted for about one-eighth of the overall decline in wholesale funding.

For Bernanke, the collapse of interbank lending was proof of a general loss of confidence in the banking system following Lehman Bothers' failure.  That same loss of confidence was still more apparent in the pronounced post-Lehman increase in the TED spread:

The skyrocketing cost of unsecured bank-to-bank loans mirrored the course of the crisis. Usually, a bank borrowing from another bank will pay only a little more (between a fifth and a half of a percentage point) than the U.S. government, the safest of all borrowers, has to pay on short-term Treasury securities. The spread between the interest rate on short-term bank-to-bank lending and the interest rate on comparable Treasury securities (known as the TED spread) remained in the normal range until the summer of 2007, showing that general confidence in banks remained strong despite the bad news about subprime mortgages. However, the spread jumped to nearly 2-1/2 percentage points in mid-August 2007 as the first signs of panic roiled financial markets. It soared again in March (corresponding to the Bear Stearns rescue), declined modestly over the summer, then showed up when Lehman failed, topping out at more than 4-1/2 percentage points in mid-October 2008 (pp. 404-5).

These developments, Bernanke continues, "had direct consequences for Main Street America. … During the last four months of 2008, 2.4 million jobs disappeared, and, during the first half of 2009, an additional 3.8 million were lost.” (406-7)

There you have it, straight from the horse's mouth: the fourth-quarter, 2008 contraction in wholesale funding, as reflected in the collapse of interbank lending, led to the loss of at least 6.2 million jobs.

But was the collapse of interbank lending really evidence of a panic, brought on by Lehman's bankruptcy?  The timing of that collapse, as indicated in the following graph, tells a much different story.

Interbank Loans All Commercial Banks

The first of the three vertical lines is for September 15, 2008, when Lehman went belly-up. Interbank lending on the next reporting date — September 17th —  was actually up from the previous week.  Thereafter it declined a bit, and then rose some.  But such variations weren't all that unusual.  As for the TED spread, although it rose sharply after Lehman's failure, the rise reflected, not an actual increase in the effective federal funds rate (as the "panic" scenario would suggest), but the fact that that rate, though it actually declined rapidly, did not do so quite as rapidly as the Treasury Bill rate did:

TED Spread

OK, now on to those other vertical lines.  They show the dates on which banks first began receiving interest payments on their excess reserves.  There are two lines because back then two different sets of banks had different "reserve maintenance periods," and therefore started getting paid at different dates.  (The maintenance periods have since been made uniform.) Those (mostly smaller) banks with one-week reserve maintenance periods began earning interest on October 15th; the rest, with two-week maintenance periods, started getting paid on October 22nd.  The collapse in interbank payments volume coincides with the latter date.  Notice also that the collapse continues after the TED spread has returned to a level not so different from its levels before Lehman failed.

If you still aren't convinced that IOR was the main factor behind the collapse in interbank lending, perhaps some more graphs will help.  The first shows the progress of interbank lending over a somewhat longer period, along with the 3-month Treasury Bill rate and (starting in October 2008) the interest rate on excess reserves:

IOR and Interbank Lending

To understand this graph, think of the banks' opportunity cost of holding excess reserves as being equal to the difference between the Treasury Bill rate and the rate of interest on excess reserves.   Prior to October 15th, 2008, the opportunity cost, being simply equal to the Treasury Bill rate itself, is necessarily positive.  But when IOR is first introduced, it becomes practically zero; and shortly thereafter it becomes, and remains, negative.  Mere inspection of the chart should suffice to show that the volume of interbank lending tends to vary directly with this opportunity cost.

Once the interest rate on excess reserves is fixed at 25 basis points after mid-December 2008, things get simpler, as the volume of interbank lending varies directly with the Treasury Bill rate.  Here is a chart showing that period, with the opportunity cost itself (that is, the Treasury Bill rate minus 25 basis points) plotted along with the volume of interbank lending:

Opportunity Cost

Now, it would be one thing if Bernanke were merely guilty of misunderstanding the cause of the decline in interbank lending, without having actually been responsible for that decline.  But Bernanke was responsible, as was the rest of the Fed gang that took part in the misguided decision to start rewarding banks for holding excess reserves in the middle of a financial crisis.

What's more, it is hard to see how Bernanke can insist that the Fed's decision to pay IOR had nothing to do with the drying-up of the federal funds market given the justification he himself offers for that decision earlier in his memoir, which bears quoting once again, this time with emphasis added:

[W]e had been selling Treasury securities we owned to offset the effect of our lending on reserves… . But as our lending increased, that stopgap measure would at some point no longer be possible because we would run out of Treasuries to sell….The ability to pay interest on reserves…would help solve this problem. Banks would have no incentive to lend to each other at an interest rate much below the rate they could earn, risk-free, on their reserves at the Fed (p. 325).

Yet when he turns to explain the causes of the collapse in interbank lending, just eighty pages after this passage, Bernanke never mentions interest on reserves.  Instead, he blames the collapse on panicking private-market lenders, while treating the Fed — and, by implication, himself — as a White Knight, galloping to the rescue.  "As the government’s policy response took effect," he writes, "the TED spread declined toward normal levels by mid-2009" (p. 405).  What rubbish.  We've already seen why the TED spread went up and then declined again.  And although interbank lending itself revived somewhat during the first half of 2009, it declined steadily thereafter, ultimately falling to lower levels than ever.

And the Fed's "policy response"?  According to Bernanke, it had "four main elements: lower interest rates to support the economy, emergency liquidity lending…and the stress-test disclosures of banks’ conditions” (409).  Let Kevin Dowd tell you about those idiotic stress tests.  As for "lower interest rates," they were proof, not that the Fed was taking desirable steps, but that it was failing to do so, for although the Fed did get around to reducing its federal funds rate target, its doing so was a mere charade: the equilibrium federal funds rate had long since fallen well below the Fed's target, and the subsequent moves merely amounted to a belated recognition of that fact, without making any other difference.  Finally, although the Fed's emergency lending aided the loans' immediate recipients, as well as their creditors, it contributed not a jot to overall liquidity, the very point of IOR having been — as Bernanke himself admits, and as I explained in my first post on this topic — to prevent it from doing so!

As the next chart shows, IOR, besides contributing to the collapse of interbank lending, also played an important part in the dramatic increase in the banking system reserve ratio.  The vertical lines represent the same three dates as those referred to in the very first chart.  Although the ratio did  rise considerably following Lehmans' failure, it rose even more dramatically — and, quite unlike the TED spread, never recovered again — after the Fed started paying interest on excess reserves:


To better understand what went on, here is another diagram, this one showing banks' choice of optimal reserve and liquid asset ratios as a function of the interest paid on bank reserves:


In the diagram, the vertical axis represents the interest rate on reserve balances, in basis points, while the horizontal axis represents the reserve-deposit ratio.  The picture shows two upward-sloping schedules.  The first is for reserve balances at the Fed, while the second is for liquid assets more generally, here meaning (for simplicity's sake) reserves plus T-bills.  The horizontal line shows the yield on T-bills at the time of implementation of IOR, here assumed to be a constant 20 basis points.  The two dots, finally, represent equilibrium ratios, the first (at the lower left) for before the crisis and IOR, the other for afterwards.  Note that, the high post-IOR ratio reflects, not just the interest-sensitivity of reserve demand, but that, with IOR set at 25 basis points, reserves dominate T-bills.  Thus, although the demand for excess reserves may not be all that interest sensitive so long as the administered interest rate on reserves is less than the rate earned by other liquid assets, that demand can jump considerably if that rate is set above rates on liquid and safe securities.

The last chart I'll trouble you with today tracks changes in total commercial bank reserves, interbank loans, Treasury and agency securities, and commercial and industrial loans, from mid-2006 through mid-2009, this time with a single vertical line only, for October 22, 2008, when IOR was in full effect:

Composition of Bank Assets

The chart shows clearly how the beginning of IOR coincided, not only with a substantial decline in interbank lending (green line), but in a leveling-off of other sorts of bank lending, which later becomes a pronounced decline.  For illustration's sake, the chart shows the course of C & I lending only; other sorts of bank lending fell off even more.

Don't get the wrong idea: I don't wish to suggest that IOR alone was responsible for the post-2008 decline in bank lending, apart from overnight lending to other banks.  There's little doubt that that decline mainly reflects the effects of both a declining demand for credit and much stricter regulation of bank lending, especially as Dodd-Frank and Basel III came into play, Nor do I believe that merely eliminating IOR, as opposed to either reducing the regulatory burdens on bank lending, or resorting to negative IOR (as some European central banks have done), or both, would have sufficed to encourage any substantial increase in bank balance sheets, and especially in bank lending, after 2009, when most estimates (including the Fed's own) have "natural" interest rates sliding into negative territory.  But as I noted in my first post in this series, when IOR was first introduced, natural rates were, according to these same estimates, still positive.  And one thing IOR certainly did do, both before 2009 and afterwards, was to allow banks, and some banks more than others, to treat trillions in new reserves created by the Fed starting in October 2008, not as an inducement to expand their balance sheets, but as a direct source of risk- and effort-free income. (Note, by the way, how, just before IOR was introduced, but after the Fed stopped sterilizing its  emergency loans, bank loans and security holdings did in fact increase along with reserves.)

Moreover, it's evident that the FOMC itself, rightly or wrongly, sees IOR as continuing to play a crucial part in limiting banks' willingness to expand credit.  Otherwise, how can one possibly understand that bodies' decision last month to raise the rate of IOR (and, with it, the upper bound of its federal funds rate target range) from 25 to 50 basis points?  That decision, recall, was aimed at making sure that bank credit expansion would not progress to the point of causing inflation to exceed the Fed's 2 percent target:

The Committee judges that there has been considerable improvement in labor market conditions this year, and it is reasonably confident that inflation will rise, over the medium term, to its 2 percent objective. Given the economic outlook, and recognizing the time it takes for policy actions to affect future economic outcomes, the Committee decided to raise the target range for the federal funds rate to 1/4 to 1/2 percent. The stance of monetary policy remains accommodative after this increase, thereby supporting further improvement in labor market conditions and a return to 2 percent inflation.

Bernanke's implementation and defense of IOR would be more than bad enough, were it not also for his particular determination to avoid repeating the mistakes the Fed made during the Great Depression.  "[M]ost of my colleagues and I were determined," he says, "not to repeat the blunder the Federal Reserve had committed in the 1930s when it refused to deploy its monetary tools to avoid a sharp deflation that substantially worsened the Great Depression” (p. 409).  Among the Fed's more notorious errors during that calamity were its failure to expand its balance sheet sufficiently, through open-market purchases or otherwise, to offset the dramatic, panic-driven collapse in the money multiplier during the early 1930s, and its recovery-scuttling decision to double reserve requirements in 1936-7.

Of course, Bernanke's Fed didn't commit the very same mistakes committed by the Fed of the 1930s.  But, as David Beckworth had already recognized by late October 29, 2008, it made remarkably similar ones that also resulted in a collapse of credit.  "History," Bernanke  credits Mark Twain with saying, "does not repeat itself, but it rhymes” (p. 400).  If you ask me, Bernanke himself was a far better versifier — and a far worse central banker — than he and his many champions realize.


Addendum (1-12-2016, 6PM): As Forbes' Frances Coppola, in replying to my criticism of her in an earlier post in this series, claims that I am inconsistent in my various posts regarding to bearing of IOR on bank lending, allow me to clarify my position by means of the following précis:

1. IOR was implemented in October 2008 for the avowed purpose of checking bank credit expansion in response to the Fed's creation of fresh bank reserves.

2. In fact, IOR contributed to the fall 2008 wholesale credit crunch, most obviously by causing a dramatic decline in interbank lending.   Again, this contribution was anticipated by Bernanke and others responsible for the policy.

3. Once the rate of IOR exceeded the yield on Treasuries and other low-risk assets, as it did shortly after the program began, banks had an incentive to accumulate excess reserves instead of attempting to acquire such securities.  Thus the normal process of bank balance-sheet expansion and deposit creation in response to reserve injections was short-circuited.

4. IOR also contributed to the relative decline in risky bank lending by increasing the marginal opportunity cost of such lending.  This portfolio effect of IOR on risky lending was very small relative to that of increasing regulatory burdens, including capital requirements, especially after 2008.  But as at least some banks had both surplus capital and surplus reserves, capital constraints alone did not prevent IOR from also having some influence.

5.  Fed officials, including Bernanke, who would deny that IOR had the consequences I have just outlined, are at least obliged to reconcile their denials with the justifications offered for implementing the program in the first place.