The Fed Predicts Inflation, But Doesn't Know Why

Regulation D was the part of the original Federal Reserve Act designed for when physical money was at the center of all transactions. Altered several times during the thirties, it made (some) sense in the context of what had happened starting in late 1930. People began removing cash from the banking system to hoard it lest it become tied up, and lost, in a failing bank. These runs were of the physical variety.

Long after the Great Depression had concluded, banks quite naturally and legitimately wished to move out from under the more onerous restrictions. They argued over the years, successfully, that different types of customer accounts should be handled under different reserve restrictions. A personal transaction account, such as checking or demand deposit, would be accessed regularly and therefore would be properly subjected to higher requirements.

Savings accounts and time deposits, on the other hand, were used far less frequently and therefore some consideration should be given to reducing the reserves attached to them. This was granted so long as these particular accounts met certain thresholds for use, or, in this case, disuse.

Once this disparity was set into the reserve infrastructure, almost immediately the banking system set out to evolve to it. The goal for any bank is always efficiency, in this case to be able to reduce reserve requirements as much as possible for a given set of liabilities; all the while, of course, keeping the depositor happy along the way.

The trick, then, would be how to make a transaction account look and act like a non-transaction one. The former carries with it both a much higher reserve requirement and also the consumer preference for every day, and therefore widespread, use. The latter has much lower reserve requirements, but not as many people readily accept the transaction restrictions imposed on them.

Necessity being the mother of invention, the Great Inflation was one of those historical periods featuring so many. Given the nature of the malady, monetary disease, it’s not surprising that most of these involved banking and monetary form.

Suppose a plain vanilla depository institution operating in the seventies wished to hold on to its base of customers. This wasn’t easy given both Regulation D as well as Regulation Q, and its similar Depression era changes, which, for its own (ultimately wrong) reasons, imposed a ceiling on the rates banks could pay deposit customers. As interest rates in general rose with inflation (the interest rate fallacy), the banking system was forced to invent new ways to stay competitive.

To do so meant overcoming both hurdles simultaneously; to pay customers a better interest rate than regulations allowed, while also getting much more cash efficient on the liability side.

Repurchase agreements had been around for a very long time before the Great Inflation, before the Great Depression, even. Many people are often confused by them simply because they are not what they are supposed to be. The term itself does us no favors in this regard, suggesting a transaction where someone sells a security from someone else at the same time agreeing to buy it (repurchase) back at an agreed upon future date.

The evolution of repo, as it has come to be known, was fraught with unusual difficulty right down to this very meaning. Was it an actual transaction where title of the underlying asset shifted from one hand to the next and back again? No one really knew especially during the early days of the repo market’s rapid growth in the fifties and sixties. The government, often in tax cases, treated it differently depending on its ultimate goal (as did those engaged in them who often classified the thing based on whether it conferred special tax treatment of the security collateral).

In practice, a repo is a collateralized loan. It is often short-term mostly overnight. But because of its history and the lack of clarification while the repo market grew steadily, there remained this opportunity to use what was a collateralized lending arrangement as a repurchase agreement to satisfy both customer demand as well as regulatory efficiency.

As the Great Inflation roared, it became more common for banks to offer first their largest (often corporate) customers the option of a “sweep” account. At the close of each business day, the institution would sweep an agreed percentage of that customer’s transaction deposit balance into an overnight repurchase agreement.

Repos were not subject to Regulation Q and could offer participants a more competitive interest rate. For the bank, it meant shrinking the transaction deposit liability, indeed the entire balance sheet presented for those purposes, therefore reducing required reserves and applied vault cash while raising excess reserves and surplus vault cash by comparison.

In many cases, this increase in excess reserves relative to a non-repo sweep paradigm was lent out in money markets such as federal funds. Rather than hold the former balance on reserve account with the local Federal Reserve branch, the bank was free to sell federal funds in the excess that it had gained by reducing its balance sheet exposures. It provided a direct link and point of integration between these money markets.

To be able to do this, the bank was required to hold a substantial inventory of liquid Treasury securities (and, as the repo market evolved, other fixed income substitutes like agency paper). In the seventies, this wasn’t much of a problem as interest rates provided an incentive on the asset all its own.

On the bank’s balance sheet, the amount swept into the repo simply disappears on both sides; the holdings of UST securities on the asset side is reduced (treated as a repurchase agreement) by the same amount as what was swept out of transaction deposits. The money goes missing especially in conventional monetary categorization because at the time (and even now) repurchase agreements weren’t defined as money.

Even as the Great Inflation waned into the eighties, monetary evolution kept right on with banking. Money market mutual funds had been another increasingly popular answer to the regulation distortions. That led in 1982 to the Garn-St. Germain Act and the creation of the money market deposit account (MMDA). These are treated as savings deposits rather than transaction deposits, and beginning in December 1990 the reserve requirement on them (as well as eurocurrency liabilities) was set to zero.

Thus, starting in 1991 and after the bank could sweep transaction deposits into an MMDA. This had the effect of reducing even further reserve requirements on it, largely because this sort of accounting could be applied to a much wider depository base. It would have been very difficult if not impossible to repo sweep the accounts of regular folks holding small transaction balances, but far easier to MMDA sweep.

The introduction of computer programmed sweep accounts around 1994 made it a no-brainer for widespread adoption.

M2 money velocity around 1990 and 1991 started to rise precipitously. It had been in a rather narrow range between 1.7 and 1.8 going back to the late 1950’s.  By Q2 1992, however, it rose above 1.9 and kept going. It wasn’t necessarily a problem for Federal Reserve officials trying to conduct monetary policy, but it was an issue. A lot of how monetary policy worked depended upon dependable correlations – like those between the stock of money and the aggregates of the real economy.

As policymakers would much later disclose, they did not see this shift coming even though in the end it was very easily explained. Banks had started to repo and MMDA sweep in massive amounts, effectively reducing the growth of M2 since that measure doesn’t include either practice. Economic output, real GDP, accelerated in the middle nineties as money stock M2 lagged, therefore a sharp, historical rise in velocity.

Around late 1997, however, it started to reverse. The FOMC in the immediate aftermath of LTCM and the Asian flu in February 1999 discussed what this might have meant.

“CHAIRMAN GREENSPAN. Tom, is there any evidence to suggest that the big surge in velocity that occurred, contrary to expectations, from opportunity costs in the early 1990s may be in the process of reversing itself?

MR. SIMPSON. That’s a possibility, but it is rather hard to come up with the economic intuition as to why that might be happening.

CHAIRMAN GREENSPAN. Any more so than was the case with respect to the economic intuition in 1990? What happened subsequent to 1990 was far more of a surprise than a reversal would be today, if I may put it that way.”

It was much more than just M2, particularly in that one economic context of global turmoil. Things were not good in the world, and though it appeared, at least to Federal Reserve officials, that the US had dodged a bullet with LTCM they missed the monetary evolution associated with its rise, and fall. The Asian flu was, to put it in recent terms, the first “rising dollar” complete with multiple shocking currency “devaluations.”

The thing is, Alan Greenspan had the answer all along; he just never did anything with it. As he was quizzing Tom Simpson, Associate Director of the Division of Research and Statistics, about these Humphrey-Hawkins monetary targets, the Chairman admitted again:

“CHAIRMAN GREENSPAN. I must say that I have not changed my view that inflation is fundamentally a monetary phenomenon. But I am becoming far more skeptical that we can define a proxy that actually captures what money is, either in terms of transaction balances or those elements in the economic decisionmaking process which represent money. We are struggling here. I think we have to be careful not to assume by definition that M1, M2, or M3 or anything is money. They are all proxies for the underlying conceptual variable that we all employ in our generic evaluation of the impact of money on the economy. Now, what this suggests to me is that money is hiding itself very well.”

A clear change in the velocity of one or another of the M’s without a corresponding change in the overall nature of the economy or markets would indicate exactly what the “maestro” was talking about. In short, “money is hiding itself very well.” But it is also, in these changes, revealing the evolution while hidden, or in the shadows.

There was, of course, a whole bunch of reasons to suspect that was the case that had nothing to do with M2 velocity. In early 1999, there were all sorts of irregularities in markets if not yet the US economy (the global economy was a different story). 

Most people are immediately drawn in on the dot-com bubble, and for good reasons, but coincident to it was this “rising dollar” period complete with a full-on bond market conundrum – up and down.

As the Asian flu, really “dollar” irregularity, progressed, the entire US Treasury curve fell underneath the monetary policy target. It sounds preposterous, but for a time in 1998 when LTCM was in its final death throes (clue) every Treasury yield was less than the policy rate; the 30-year long bond yielded as little as 4.70% on October 5, 1998 even though the federal funds target was on that day 5.25%. The yield curve was not inverted, but it was collapsed (the 2s10s was just 12 bps on October 5; 5s10s only 21 bps).

The FOMC had voted to reduce the federal funds target by 25 bps just a few weeks before, and would vote to do so again a little over a week after (followed by a third in November 1998).

These low bond yields were puzzling to the Committee members. Why would banks, in particular, be paying such a high premium for US official paper? Treasuries are, particularly at the short end, monetary substitutes. There was perceived little or no risk in federal funds, so why hold a UST at any maturity paying a premium let alone a large one to do so?

They got a sense of it in their September 1998 meeting, the one held just after the LTCM debacle came to light.

“MR. KOHN. I do think that the market has put a greater premium on holding liquid assets in these very uncertain times. We can see that in the spread between Treasury securities and federal agency securities, even the benchmark agency securities, which are extraordinarily liquid. That spread has widened out by 15 basis points. The on-the-run, off-the-run Treasury spreads also have widened out.”

It was an obvious liquidity premium where the difference OTR (on-the-run) to OFR (off-the-run) is an indication of repo preferences. Despite everything the Fed had done (mostly ineffective lip service up to that point) and would do (mostly ineffective rate cuts that did nothing to help Russia, Thailand, Japan, Brazil, etc.) the market “for some reason” was pricing dollar illiquidity that the US central bank was not comprehending.

As has become normal, such distortion was taken as an anomaly with which to disregard all important signals aligned with it. To be somewhat fair, though only to a very limited extent, the Federal Reserve’s mandate is not terrestrial. Their concerns, as would be the case in 2007 and 2008, was for the US end of the system alone.

That did not, however, dismiss them of all responsibility. Even if they had no authority over a global dollar regime they should have appealed to Congress for one, or at least have started the process for studying what that actually meant.

By later 1999, instead, they were right back to normal as if none of this had happened; no one the wiser, and in short order, a great many quite poorer. In June of that year, the FOMC voted for the first “rate hike”, to be followed in immediate succession by five more, the last for 50 bps, bringing the federal funds target all the way from 4.75% to 6.50% in the middle of 2000.

And all along the way, the bond market resisted. The benchmark 10-year UST had sold off from 5.93% in yield the day the Fed started to 6.58% by the first trading day of 2000.  By the next “rate hike” in early February 2000, the yield was still 6.60%; at the next in March, it had dropped to 6.13% and falling below 6.00% in early April. Even at the 50 bps hike in May, the 10s were still yielding ~6.40% and more continuously underneath again.

By the time the Fed realized what was really going on at the beginning of 2001, the federal funds target was still 6.5% while the 10s were all the way at 4.92%.  The dot-com bubble had burst and more than that risk was, for a short while, alive again. The US economy never really recovered, though the dot-com recession was the mildest on record.

The reasons for that aren’t perfectly straightforward, but they have a lot if not mostly to do with monetary evolution. The stock bubble was in many ways a sideshow, a symptom of bigger transformations (and also the reason why the stock crash didn’t produce another Great Depression as Greenspan by 2002 started to fear). 

And it was the bond market which was telling monetary officials to look out. But to heed that warning would have meant answering Alan Greenspan’s admission about “hiding” money. To more completely explain UST yields in the context of an actually effective monetary policy would have required delving into the shadows to determine exactly what was going on in them. As if LTCM wasn’t enough, the biggest and deepest market in the world was trading for years on end in fear of everyone finding out.

Though that wouldn’t happen for another business cycle, and conundrum, further on, by 2007 that’s just what occurred.

You can’t deny the similarities now to the Asian flu and its aftermath. That “rising dollar” looked a whole lot like the most recent one 2015-16. The bond market after it is behaving almost exactly in the same way, too, with the exception of nominal placements (the yield curve isn’t *yet* underneath federal funds policy targets RRP or IOER). Nominal rates are rising but resistance to them is palpable in the collapse of the yield curve. Oh, and that stock bubble thing.

The Fed says inflation is coming, though even they have been forced to admit fairly regularly (including their latest policy statement released this week) they don’t really know why. It is, at this point, more an article of faith in the mainstream than rational and reasoned analysis. The bond market is once more politely (for now) suggesting they take the time to figure out what they don’t know. Money isn’t hiding, it’s constantly evolving.