Update on money demand, currently stable

(This is an updated and rewritten version of a post I wrote about seven months ago.)

I've mentioned the demand for money countless times in the 10-yr history of this blog because it's a very important and under-appreciated macro variable. The Fed controls the supply of money, but the demand for money is a function of a variety of factors, some of which are beyond the Fed's ability to control. The secret to any central bank's ability to deliver low and stable inflation is to keep the supply and demand for money in balance, even though it is difficult to track the demand for money without the benefit of hindsight. As Milton Friedman taught us, inflation is always and everywhere a monetary phenomenon; inflation results from an excess of money relative to the demand for it. It's that simple. Unfortunately, you don't see many people, including the Fed, talking about this, because it's not easy or straightforward. That's one of the things this blog tries to bring to the table.

From 2008 through the summer of 2017, the demand for money (which I define as the M2 money supply divided by nominal GDP) was in a pronounced uptrend, because M2 growth exceeded nominal GDP growth. This in turn was driven by a widespread desire on the part of individuals, investors, and corporations to increase their holdings of money relative to their incomes and as a share of their portfolio, which in turn was a natural reaction to the profound shock of the global financial crisis and the Great Recession of 2008-9. Everyone wanted the safety of money, and they wanted more of it.

During this same period of rising money demand, the Fed engineered an astounding and unprecedented increase in the monetary base, expanding its balance sheet by buying trillions of dollars of notes and bonds. The fact that inflation remained low (~1.5 - 2%) and relatively stable throughout this period can only mean that the Fed correctly responded to strong money demand by dramatically expanding the supply of money. Inflation was not a problem because the Fed kept the supply and demand for money in balance. 

The Fed was unfortunately a bit slow to react to the surge in money demand, but by late 2008 they adopted a radically new monetary policy which they dubbed Quantitative Easing (QE). Most people erroneously believed, and still believe, that the objective of QE was to stimulate the economy by printing money and otherwise making money cheap (i.e., by keeping interest rates artificially low). I have argued in numerous posts over the years that this was most definitely NOT the case. The purpose of QE was to accommodate the market's almost-insatiable demand for risk-free, liquid assets.

In the wake of the near-collapse of the global financial system, the world was desperate to acquire more T-bills, the gold standard for safe money. The demand for T-bills was so extreme that there were not enough in supply—in fact, the Fed was essentially forced to sell virtually all of its T-bill holdings in the first half of 2008 in an attempt to satisfy demand. Without enough safe and liquid securities to satisfy the demand for such, the financial system was in serious danger of imploding. The Fed solved this problem by buying trillions of dollars of notes and bonds and paying for them with bank reserves, which they also announced would for the first time ever begin to pay a risk-free rate of interest. (Bank reserves, crucially, are not "money" that can be spent like dollars can.) The banking system was happy to use their strong inflows of savings deposits to invest in and hold all these new bank reserves, which had suddenly become a valuable asset, being risk-free and interest-bearing, just like T-bills. Banks were risk-averse too, just like the general population. Bank reserves today are best thought of as T-bill equivalents.

Now, however, it makes sense for the Fed to reverse the QE process at least in part, because the crisis of confidence has begun to reverse, risk aversion has begun to fade, and the demand for money appears to have stabilized, at least for now. The Fed already has reversed some of its QE, by reducing excess reserves from a high of $2.7 trillion in 2014 to now $1.5 trillion. In fact, if the Fed had NOT reversed at least some of its QE, they would have run the serious risk of allowing there to be too much money relative to the demand for it, and that would have led to a destabilization/devaluation of the dollar and an unwanted rise in inflation.

If the demand for money weakens from its current level, as it likely would if confidence rises and economic growth remains healthy, then the Fed will be justified in further reducing the size of its balance sheet and raising its short-term interest rate target. And while the market may fret again about rising rates, it needn't be problematic if rising rates are tied to rising confidence, and declining risk aversion.

As always, here are charts that provide the evidence for my story:

Chart #1

The M2 measure of money supply is considered by most economists to be the best measure of "money." As Chart #1 shows, M2 consists of currency, checking accounts, bank savings deposits, and retail money market funds, all of which are liquid, safe, and easy to spend. The largest component by far is bank savings deposits, which grew from $4 trillion at the end of 2008 to now $9.4 trillion. In fact, the growth of savings deposits far outstripped the demand for the other components of M2: savings deposits were about 50% of M2 at the end of 2008, and now they are almost 65%. This is significant, because until 2017 bank savings deposits paid almost nothing in the way of interest, yet people were happy to increase their savings deposits because they offered safety and liquidity. Bank savings deposits in an era of zero interest rates were thus an excellent measure of money demand. 

Chart #2

As Chart #2 shows, the growth rate of bank savings deposits has slowed significantly in the past two years, beginning in early 2017, despite the fact that banks have increased—albeit slowly—the rate they pay on savings accounts over this same period. Prior to 2017, the rate on savings accounts was essentially zero, whereas now most banks are paying around 2%. That's significant: savings accounts on the surface have become more attractive (because they pay a higher rate of interest), yet demand for them has weakened. Why? Because risk aversion is declining and confidence has improved. In other words, slower growth in bank savings deposits today is a direct reflection of a decline in money demand.

It's noteworthy that the big decline of money demand in 2017 coincided with a surge in equity prices. People wanted less money, and they evidently wanted to put some of their money to work in the stock market. Money demand has been flat for the past year or so, so it's not surprising that equity prices have yet to meaningfully exceed their 2017 highs. (See Chart #9 below for a detailed look at equity prices in recent years.)

Chart #3

As Chart #3 shows, it's not just the demand for savings deposits that has declined—the growth rate of the larger M2 measure of money has also declined. M2 currently is growing at a 3-4% annual rate, which is meanin agfully less than the 6.5% annualized rate which has prevailed for most of the past several decades.

Chart #4

As Chart #4 shows, the growth rate of nominal GDP (shown in the blue bars) has picked up a bit in the past few years. Meanwhile, as Chart #3 shows, the growth in M2 has declined. This means that people are unwinding their holdings of cash. Since money can't disappear without help from the Fed (i.e., tight monetary policy), the only way that we collectively can reduce our money balances is by pushing up the price and the volume of transactions of everything else.

Chart #5

Chart #5 is the main exhibit, and it shows the ratio of M2 to nominal GDP. To understand this chart, think of M2 as a proxy for the amount of cash (or cash equivalents) that the average person, company, or investor wants to hold at any given time. Think of GDP as a proxy for the average person's annual income. The ratio of the two is therefore a proxy for the percentage of the average person's or corporation's annual income that is desired to be held in safe and relatively liquid form (i.e., cash or cash equivalents). In times of uncertainty, it stands to reason that most people would want to hold more of their assets in cash, and in times of optimism they would want to hold less.

The demand for money appears to have peaked in the summer of 2017, after rising strongly for the previous decade. Money demand has since fallen a bit, and that is good evidence that risk aversion and caution are slowly giving way to increased risk-seeking. The last two data points in this chart are Q4/18 and Q1/19, the latter of which I've estimated using known data for M2 growth (currently running at a 3-4% rate) and the market's current estimate for nominal GDP, which is roughly 4%. Money demand didn't change much last quarter, which I view as a function of the severe downdraft in equity prices that occurred late last year, which in turn was driven by concerns that the Fed might over-tighten monetary policy.

Going forward, however, I suspect that money demand will resume its decline, since the economic and financial fundamentals remain healthy. If this trend continues, it has profound implications for future economic growth and the conduct of monetary policy.

I believe that, beginning one year or so ago, the dominant narrative switched from one in which people were willing to pay up for safety and liquidity (by accumulating cash) to now one in which the average person (or company, or investment manager) on the margin wants to reduce their holdings of "money" in favor of increasing their holdings of risky assets or just spending it. I think that narrative remains the operative one that will continue, barring, of course, unforeseen disasters.

Chart #6

So it will be very important to monitor consumer confidence, among others. As Chart #6 shows, consumer confidence surged beginning in late 2016 (not coincidentally this was shortly after Trump's election). With increased optimism naturally comes a reduced desire to accumulate cash, and an increased desire to spend money and/or invest it. The rise in confidence was interrupted late last year however, and that is why economic growth in the first quarter of this year was almost certainly mediocre. 

This all has important implications for the Fed, because the Fed will need to take actions to offset the decline in the demand for money, or else it will risk igniting an unwanted increase in inflation. The Fed will need to raise short-term interest rates, and probably by more than the market currently expects (higher short-term rates have the effect of making savings deposits more attractive, thus counteracting the declining demand for same). The Fed will also need to resume the reduction in its balance sheet in order to reduce the supply of bank reserves as banks' demand for those reserves declines. This may cause the market consternation, but it will be exactly what is needed to ensure continued low and stable inflation and in turn a strong economy.

Chart #7

Chart #8

Chart #9

For now, things look fairly stable. Money demand and money supply are likely in balance. The economy is growing at a modest/moderate pace, and few expect that to change. Real and nominal interest rates are priced to the expectation that economic growth will be 2.5% (see Chart #7) and inflation will be about 2% for the foreseeable future (see Chart #8). Fear, uncertainty and doubt have fallen almost to normal levels (see Chart #9), and the bear market of late last year has been reversed. But a resumption in the decline of the demand for money, which would likely accompany a pickup in confidence and stronger growth, would change that dynamic, and almost certainly result in higher-than-currently-expected interest rates and higher equity prices.