Wednesday, June 24, 2015

The Fed's game plan: it's all about the demand for money

It's no secret that the Fed will begin raising short-term interest rates before too long. However, the method they will use has never been tried before, so there naturally exists a degree of confusion and uncertainty surrounding the future course of monetary policy and how it will affect the economy. The purpose of this post is to simplify the issue in the hope that leads to better understanding. 

As I've argued in numerous posts over the past 5-6 years, the most important driver of monetary policy from late 2008 until recently has been unusually strong money demand. As I see it, the Fed was almost compelled to resort to Quantitative Easing in order to satisfy the world's demand for money and near-money substitutes. QE was not about pumping money into the economy, it was all about satisfying the economy's demand for liquidity. QE was erroneously billed as "stimulative," since printing money in excess of what's needed only stimulates inflation. Instead, QE was designed to accommodate intense demand for money, without which the economy might well have stumbled.

It all started when financial markets teetered on the edge of the abyss in the third quarter of 2008. The problem was that the world suddenly wanted a lot more money and money substitutes (e.g., cash, bank savings deposits, T-bills) than were available. Debt that made sense when housing prices were rising suddenly became suspect as prices collapsed. Too many people tried to rush for the exits, but the financial system lacked the necessary liquidity to allow so many to sell so much. Even before the meltdown that followed the collapse of Lehman, demand for T-bills was so intense that the Fed all but exhausted its holdings of T-bills by mid-2008 in an attempt to increase their supply on the market.  

Such was the shortage of money in late 2008 that the dollar soared, commodity prices collapsed, and the prices of TIPS and Treasuries signaled impending deflation on a massive scale, as the charts above show.

The increase in the demand for money can also be seen in the chart above, which measures the ratio of M2 (cash, demand deposits, retail savings accounts) to nominal GDP. Think of it as the amount of cash and cash equivalents that the average person wants to hold relative to his annual income. Between the end of 2007 and today, that ratio has jumped by over 30%. It's as if collectively we wanted to stuff almost $3 trillion under the mattress for a rainy day. The shock of the 2008 financial collapse and the deep recession that accompanied it resulted in years of risk aversion.

The most important policy change that QE brought with it was the Fed's decision to pay interest on bank reserves (IOR). Prior to that, bank reserves paid no interest, but banks were compelled to own them in order to collateralize their deposit base (our fractional reserve banking system required banks to hold about $1 in reservers for every $10 in deposits). This meant that reserves were a deadweight asset, so banks always tried to minimize their holdings of reserves. With QE, the Fed had to encourage banks to hold reserves, and IOR was the key to doing that.

Controlling and limiting the supply of reserves was the Fed's primary policy tool prior to QE. The Federal Open Market Committee controlled the supply of reserves indirectly, by targeting the interest rate that banks paid each other to borrow or lend the relatively fixed supply of reserves in existence. If the interest rate on the Fed funds market slipped below the FOMC's target, then the FOMC would deduce that reserves were in over-supply, and they would sell bonds in order to reduce the supply of reserves. If the Fed funds rate exceeded the FOMC's target, they would buy bonds in order to alleviate the shortage of reserves. (The Fed can only buy bonds from participating banks, and it can only pay for them by crediting the banks' reserve account at the Fed.) If the Fed wanted to ease monetary conditions, it would lower the funds rate target, and sell bonds to add reserves in order to make reserves more plentiful. By targeting the Fed funds rate and indirectly the supply of bank reserves, the Fed was thereby able to control the money supply and in turn inflation.

But paying interest on reserves changed everything. Suddenly banks viewed reserves as functionally equivalent to T-bills: a very safe, short-term, very liquid, and interest-bearing asset. If the world wanted tons of safe "money," the Fed now had the means to deliver what the world wanted. With QE, the Fed effectively transmogrified trillions of notes and bonds into T-bill equivalents. IOR allowed the Fed to create virtually unlimited supplies of the "money" that the world so desperately wanted—without being inflationary. Inflation, as Milton Friedman taught us, only happens when the supply of money exceeds the demand for it. QE was not about artificially pumping up the supply of money, it was about providing the money that the world wanted.

Banks have taken in almost $4 trillion of savings deposits since late 2008, and instead of lending all that money to the private sector, they lent most of it to the Fed, receiving bank reserves in exchange. (In practice, banks used their deposit inflows to purchase notes and bonds which they then sold to the Fed, which the Fed paid for with bank reserves.) Banks were very reluctant to lend to the private sector, but they loved lending to the Fed, since that involved zero risk. 

Today we know that all of the above is true, because the Fed has purchased some $3 trillion of notes and bonds yet inflation has remained relatively low and stable. The Fed supplied just enough money to satisfy the world's demand for money, so it wasn't inflationary. 

So now we look ahead to the unwinding of all this. Since a huge demand for money is what led to QE, the reversal of QE will (or at least should) be led by a reduction in the world's demand for money. If and when the banking system's desire to hold trillions of excess reserves declines, the Fed is going to have to reduce the supply of reserves—by selling its holdings of Treasuries and MBS, or allowing them to mature—and/or take steps to increase banks' desire to hold those reserves—by raising the interest rate it pays on reserves. As the Fed has told us, they will almost certainly do both, with an initial emphasis on raising the interest paid on excess reserves (IOER). If they don't, then the supply of money will exceed the demand for it, and higher inflation will ensue. Given that banks currently hold about $2.5 trillion of excess reserves, they have a virtually unlimited ability to increase their lending activities. 

There's already mounting evidence that banks have returned to the lending business with gusto—which implies that the demand for money is declining and risk aversion is receding. As the chart above shows, bank lending to small and medium-sized businesses has been increasing at strong double-digit rates since early last year. Over the same period, total bank credit has increased by more than $1.1 trillion, and is growing at a 7-8% annual pace. Prior to the beginning of last year, it took almost six years for bank credit to increase by $1 trillion. From the end of 2007 to the end of 2013, it was a time of strong money demand and pervasive risk aversion. For the past 18 months, the tide has begun to turn: money demand is down and risk aversion is receding.

When banks lend more and the private sector borrows more, that is by definition a decline in the demand for money, and it goes hand in hand with a decline in risk aversion. As banks realize that lending to the private sector produces risk-adjusted returns that exceed the IOER (now a mere 0.25%), their desire to sit on mountains of excess reserves that could be used to collateralize more lending is going to decline and lending is going to accelerate further. Raising IOER makes bank reserves more attractive, offsetting banks' growing lack of interest in sitting on excess reserves.

The Fed is going to have to react to the decline in money demand which is already underway, no question. They are going to have to slowly drain reserves from the system, and they are going to have to raise the interest they pay on reserves. But it needn't be scary.

Raising interest rates a few notches at this point is not equivalent to "tightening" monetary policy. It's more like easing off the accelerator, having reached the speed limit. As the chart above suggests, "tightening" monetary policy involves increasing real interest rates. Every recession in the past 50 years has been preceded by a significant rise in real short-term interest rates and a flattening or inversion of the real and nominal yield curves. We're still years away from that happening, and real short-term interest rates (the red line in the chart) are still negative.

The Fed can raise rates by hundreds of basis points without damaging the economy or threatening the health of financial markets, because higher rates will be a natural response to a stronger economy and a decline in the demand for money. If they don't they risk a potentially painful increase in inflation.


Cabodog said...

Scott, as always, thank you for taking the time to share your thoughts.

Benjamin Cole said...

Scott- nice post but you make a fundamental error.

"Banks have taken in almost $4 trillion of savings deposits since late 2008, and instead of lending all that money to the private sector, they lent most of it to the Fed, receiving bank reserves in exchange. (In practice, banks used their deposit inflows to purchase notes and bonds which they then sold to the Fed, which the Fed paid for with bank reserves.)"

Actually commercial banks do not sell notes or bonds to the Fed. Only the 22 primary dealers do.

You may wish to contend that the commercial banks bought $4 in notes and bonds with deposit inflows, and then sold those securities to the 22 primary dealers, who then sold the $4 trillion to the Fed.

However, you run into a problem.

As explained at the NY Fed website, when the Fed buys bonds from the primary dealers, it credits the commercial bank accounts of those primary dealers.

Okay, but in your example, the commercial banks sold the bonds to the primary dealers, and got paid for those bonds by the primary dealers. So we end up with $4 trillion in cash in commercial banks, and another $4 trillion in the commercial bank accounts of the 22 primary dealers.

In fact, the Fed prints or digitizes money and buy bonds in QE. I see no shame in that. That is what they did, and they did $4 trillion of it, and they should have done more, and they probably should still be doing it. See the Bank of Japan.

Although QE operates like a OMO, it is not the same. The stimulus does not come from increasing bank reserves that can be lent out; but by printing and placing $4 trillion new cash into the hands of the ultimate bond sellers.

The bond sellers can then invest in other assets, deposit the money, or spend it, or stuff it under a mattress.

There is another reason demand for cash is soaring: low inflation or deflation. Cash in circulation soars when advanced economies go into deflation or low inflation. In Japan there is $6,000 in circulation for every resident. It makes sense to "save" in the form of cash, and then transact in cash to avoid onerous taxes in deflation.

This reality makes long-term low-inflation or deflationary economies probably not practical. A large and growing underground economy will eviscerate the tax system, placing more and more burdens on those aboveground.

Yes, that is appealing on some levels, but 3rd-World solutions are...well. not the solutions....

Scott Grannis said...

Benjamin: the Fed does not "print or digitize money" when it buys bonds. It buys bonds from primary dealers and credits their account (at the Fed) with bank reserves. To the extent that the primary dealers buy bonds from other banks, they can pay for those bonds with reserves.

John said...


"The problem was that the world suddenly wanted a lot more money and money substitutes (e.g., cash, bank savings deposits, T-bills) than were available"

Why weren't they available? Is this because of fractional reserves?


Thinking Hard said...

“The Federal Reserve’s primary tool for conducting monetary policy prior to the financial crisis was to raise or lower the federal funds rate via open market operations that changed the amount of reserves outstanding. However, this mechanism became inadequate during the Great Recession, as both employment and inflation remained below targets even though the FOMC had reduced the federal funds rate to near zero (0.25 percent). To stimulate the economy further, the FOMC sought to lower longer-term interest rates by engaging in large scale asset purchases, buying quantities of U.S. Treasury and Agency obligations.” Larry D. Wall, Federal Reserve Bank of Atlanta, please see

The above article does a great job of explaining the difficulties of raising rates during this cycle. Congress will likely need to act and exempt excess reserves from the FDIC assessment for domestic banks. Currently, the net benefit of holding reserves is lowered by the FDIC assessment rate, which varies from bank to bank. Costly regulations creates a problem with IOER arbitrage and will force the FOMC to pay a higher rate on IOER or rely more on ON RRP.

Please note, higher payments on IOER will cost the U.S. Treasury. Federal Reserve earnings payments to the U.S. Treasury will be reduced as IOER increases and will likely lead to higher interest being paid to foreign banks, which hold a significant amount of excess reserves. The above referenced article does a great job of explaining why Congress needs to act sooner rather than later. Maybe this is what the Federal Reserve is waiting on?

On a side note, it seems the Fed is in a corner. With numbers coming out today showing higher wages and consumer spending ( we are seeing continual numbers that will likely cause the Fed to act sooner rather than later. Now, to what effect does raising rates have on the strength of the U.S. dollar as the ECB eases, BoJ eases, and PBOC is attempting to ease? It will likely strengthen the dollar which will put downward pressure on U.S. large cap profits and exports. So, we raise rates and put downward pressure on profits and downward pressure on equity prices? On the other hand, we can wait and see if inflation actually picks up steam going into 2016. (March 2015)
vs (June 2015)

I like the shift down in projected interest rates. Now, to what effect will raising IOER and indirectly the Fed Funds Rate have on the overall economy? Raising IOER and the Fed Funds Rate will increase the disinflationary impact of a rate increase, but to what degree?

This impact is one reason we are likely to see lower rates than normal for a considerable length of time. Please see ( with Ben Bernanke saying we will not see rate normalization during his lifetime.

Scott Grannis said...

Alter: Conceptually, I suppose what happened was similar to a run on the banks. Too many people wanted to sell, and there were not enough buyers. Too many people wanted to shift from holding risky assets to holding cash and cash equivalents, and there weren't enough people willing or able to do the opposite. Almost overnight, the world wanted more cash and cash equivalents than existed. Supplying that cash is something only the Fed can do, as "the lender of last resort."

Thinking Hard said...

Scott - What impact do you see overseas from raising IOER? Raising it too far or too fast will likely create an arbitrage opportunity that outweighs any short to mid term sovereign debt in many areas. Will this impact create a theoretical ceiling on IOER because of the potential negative externalities in overseas economies and markets?

William said...

Scott thank you again for the detailed explanation. You wrote: "The Fed can raise rates by hundreds of basis points without damaging the economy or threatening the health of financial markets, because higher rates will be a natural response to a stronger economy and a decline in the demand for money."

So what do you believe are the real reasons that the FED hasn't raised the Federal Funds Rate? Publically FED members speak of unemployment and inflation rates. But no where are either mentioned in your discussions.

On the one hand it is mind boggling why the FED is unwilling to raise rates by 1/4 percent? On the other it is concerning that the FED appears to think something negative would happen to the economy if they move too soon.

Do you think they basically are monitoring what US banks are doing with their reserves parked at the FED?

Scott Grannis said...

William: It's unfortunate, but I believe one reason the Fed is reluctant raise rates is that it is composed of humans, and humans aren't perfect. I'm sure they worry that if they raised rates at a time when the economy is not particularly strong and inflation is quite low, and then something went wrong, they would be blamed. They would dearly like to see some stronger economy news to give them cover. I don't think their hesitation to date is a serious mistake, but they are taking risks by waiting.

Similarly, I think the Supreme Court today ruled in favor of Obamacare not on the merits but on the justices' unwillingness to take the blame for disrupting the healthcare industry.

Scott Grannis said...

Thinking Hard: I suspect the forex markets are already priced to the expectation that the Fed will raise rates sooner, and perhaps more than, other central banks. That would explain the dollar's substantial strengthening in the second half of last year. Is this a bad thing? I don't think so. The dollar is only moderately stronger, adjusted for inflation, than its long-term average against a large basket of currencies, and approximately equal to its long-term average against major currencies.

In any event, I don't think there is anything necessarily wrong with a strong currency, and I would always want the dollar to be stronger rather than weaker, given the choice.

What I worry more about is the Fed waiting too long to raise rates because of worries about other economies, etc. A mistake (e.g., delaying rates hikes) can compound over time, and the inevitable adjustments and corrections that later need to be made can seriously disrupt markets and economies.

Benjamin Cole said...


Earlier you posted charts showing the market expects inflation below target for the next 10 years, at less than 2% on the CPI (although the Fed targets the PCE, which is running even lower). If that is true, is the Fed being "too easy"?

If markets expect inflation running below target for the next 10 years, how can the Fed be "too easy"?

Obamacare: The Supremes voted 6-3 in favor, led by Chief Justice and Bush nominee Chief Justice Roberts. If Obamacare is unconstitutional, we are in a mess. Our top legal talent and tip-top judges are unable to understand the Constitution. Maybe so.

Of course, we already have a huge federal medical program, staffed by more than 300,000 federal doctors, nurses and administrators, housed in federal facilities, for the free and exclusive benefit of former federal employees, and paid for by federal income and capital-tax payers.

That is Constitutional? I just described the VA. I see no allowance for the VA in the US Constitution.

Politics trumps the Constitution, which trumps ideology which trumps logic.

Benjamin Cole said...

Cleveland Fed Estimates of Inflation Expectations

News Release: June 18, 2015

The Federal Reserve Bank of Cleveland reports that its latest estimate of 10-year expected inflation is 1.82 percent. In other words, the public currently expects the inflation rate to be less than 2 percent on average over the next decade.

The Cleveland Fed’s estimate of inflation expectations is based on a model that combines information from a number of sources to address the shortcomings of other, commonly used measures, such as the "break-even" rate derived from Treasury inflation protected securities (TIPS) or survey-based estimates. The Cleveland Fed model can produce estimates for many time horizons, and it isolates not only inflation expectations, but several other interesting variables, such as the real interest rate and the inflation risk premium.


the above from freshwater Cleveland Fed....

Unknown said...

Why did the dollar go so high from 94 to 2002. What made it decline after 2002. I would like to read about it if someone can direct me to a historical record. Thanks.

Scott Grannis said...

Tim: Explaining, much less predicting, currency movements is difficult. But here's my two cents. The dollar strengthened from '94 to '02 because 1) the U.S. economy enjoyed rather healthy growth, and 2) the Fed was actively trying to slow the economy (because it worried that strong growth might prove inflationary) throughout the period, keeping the real Fed funds rate between 3 and 4.5%. Strong growth and relatively high real yields are a potent combination, and very attractive to capital. Simply put, the dollar and the U.S. economy were great investments.

Both of these conditions reversed beginning in '02. The Fed cut interest rates dramatically and the economy lost a lot of its oomph. Not surprisingly, the dollar fell.

Scott Grannis said...

Benjamin: Yes, inflation expectations in the bond market are subdued. But the Fed can't always drive monetary policy by looking in the rearview mirror. It needs to anticipate things. As Milton Friedman explained, monetary policy acts on the economy with long and variable lags. The Fed can't wait for inflation to rise in order to then tighten policy.

I'm focused on the signs of declining money demand (rising real yields, increased bank lending, a slowing in the decline of the velocity of money) because I think those are early warning signs that inflation pressures are beginning to build. I think they justify making monetary policy slightly less accommodative by raising short-term interest rates a few notches. Nobody is talking about tightening policy.

I also think that keeping short-term interest rates at or near zero for many years has the potential of producing a perverse outcome—reinforcing weak growth expectations. I'd like to see short-term rates be a little more normal.

Thinking Hard said...

When I look at this rate hike cycle I see a lot of political risk that was not present before. With IOER there is a high probability we will see populist politicians on the left and right arguing over how much interest can be paid through IOER, especially to foreign institutions. At what point does the rate on IOER pose a significant political risk? Hard to pin point for me but there is likely a line that will present PR problems and substantial political fallout.

This could become very apparent during the next presidential cycle. Could make for some interesting political debates, but probably not.