Wednesday, October 12, 2011

FOMC: We Got a Money Demand Problem

The FOMC minutes for the September, 2011 meeting were released today and the first things that stand out are the clear hawk-dove divide, the smorgasbord of additional ad-hoc monetary stimulus policy options that were discussed, and the increased economic pessimism of the members.  Something else, though, really caught my attention in the minutes.  It was this acknowledgement by the Fed staff:
 M2 surged in July and August, as investors and asset managers sought the relative safety and liquidity of bank deposits and other assets that make up the M2 aggregate. Notably, institutional investors, concerned about exposures of money funds to European financial institutions, shifted from prime money funds to bank deposits, and money fund managers accumulated sizable bank deposits in anticipation of potentially large redemptions by investors. In addition, retail investors evidently placed redemptions from equity and bond mutual funds into bank deposits and retail money market funds.
In other words, we have rapid growth in M2 coming from a surge in money demand.  This is a big deal, because money is the one asset on every market and an increased demand for it will thus affect every other market.  The more money demand there is, the less nominal spending there will be on goods, services, and other assets.  This development means the economic slump is being prolonged. 

It is great to see the Fed acknowledge this problem, but the fact is this problem has been going on for the past three years and the Fed has failed to address it in a forceful and systematic manner.  All the Fed's interventions over the past three years, including Operation Twist from this meeting, have not arrested this problem.  

How do we know this?  Well, start with the figure below.  It shows that the share of household's liquid assets (cash, checking account, time and saving deposits, money market accounts, and treasury securities) as a percent of all household's assets is closely tied to the swings in M2 velocity.  

Note that household's share of liquid assets never has returned to its pre-crisis level.  Due to the ongoing economic uncertainty, households still have an elevated demand for these assets and consequently money spending has fallen.  Consequently, velocity too has yet to return to its pre-crisis level.  

This next figure shows the actual dollar amounts.   From the peak of household asset values in 2007:Q2 to the latest data for 2011:Q2, households have lost around $8.8 trillion worth of non-liquid assets. Despite these large losses and the subsequent slump in personal-income growth, households have somehow increased their holdings of money and money-like assets by a staggering $1.6 trillion:

The composition of the increase in liquid assets is also interesting as seen in the next figure. Most of the increase has come in the form of time and saving deposits, though treasuries have been important too.  Money assets alone (cash, checking account, time and saving deposits, and money market accounts) have remained elevated and close to their peak value in late 2008.

Now these graphs only take us through 2011:Q2.  To get a sense of what has happened since then we can look at the weekly M2 data which shows the spike mentioned in the FOMC minutes.  A closer look at the M2 data, however, shows that main growth is in saving deposits. The next figure vividly illustrates this growth:

The growth of saving deposits is even more stark if viewed as percent of personal income:

This last figures shows us that the growth in savings is clearly not an increase in money demand from income growth.  It is all about holding precautionary money balances.  And this is why nominal spending continues to slump.

So what can be done?  My own view is that the money demand problem could be fixed by properly shaping expectations about future spending and inflation via something like a nominal GDP level target.  Unfortunately, the latest FOMC minutes indicate that is not an option.  So for now we lumber on, hoping that in absence of forceful and systematic Fed actions the market will be able to heal itself in a timely fashion.

Update:  This figure from a previous post shows that share of liquid assets has been systematically related money velocity over the past 60 years or so.  


  1. Great post. There are tons of people who are looking at the recent m2 growth as a sign of an inflationary boom being around the corner, but they don't seem to take money demand into account.

  2. "In other words, we have rapid growth in M2 coming from a surge in money demand."

    I think I'd disagree with your translation of the Fed's minutes.

    It seems to me like we are having a shift in demand from one sort of money to another. For instance, as the Fed minutes point out there was a shift from prime money funds to bank deposits. Rather than holding the more risky monies provided by the shadow banking system, people want to hold safer monies in M2. So this isn't a money demand problem; just a shift in investor portfolios.

  3. JP Koning,

    The shift is toward more liquid forms of money assets. That is an increase in demand for assets that are more money-like or more liquid than others. It doesn't matter if one component of M2 falls and the other increases (M2 is an arbitrary definition). It still a shift toward more liquid assets and a reversal of the hot potato process that spurs nominal spending. This is an increase in money demand.

    From an accounting perspective, it is also important to note that there can still be a net increase in money deposits for the public. The public could be selling treasuries to banks that in turn create new deposits for the public. In fact, if you look at MZM + small time deposits (which is the same as M2 plus institutional money market funds), it is growing overall. So it seems there is also some new deposit creation occuring.

  4. David,

    I came across this chart by Bianco Research on corporate (non-financial) liquidity. I have some doubts about the data (liquidity was 40% in the 50's?), but more recently it shows a non-material increase in corporate liquidity positions.

  5. As usual, I am unconvinced. You show lots of evidence that deposit money supply is increasing. How do you know that the increase is not sufficient to fully satisfy the precautionary demand?

  6. Hi David,

    Thanks for your excellent work on this.

    Can you clarify your point just a bit. Are you saying that households have more than enough "money assets" and merely refuse to spend them?

  7. "The shift is toward more liquid forms of money assets. "

    It could just as likely be a shift towards less risky forms of money-assets, and would therefore not be an increase in demand for money, but an increase in demand for safe assets.

    If the July - August jump in M2 that the Fed describes was caused by the public selling treasuries to banks, we'd expect to see a big jump in bank holdings of government securities. But this is not the case:

  8. I am also interested in the answer to JP's question as it ties into my previous question also. The data shows large declines in foreign commercial paper and institutional money funds that equates to the majority of the climb in M2. This implies that it's mostly just banks shifting assets into safer assets.

    I don't doubt that the banks are reducing their risk given the Euro crisis, but I don't see how this ties into the US economy and in particular, the US consumer. There appears to be no evidence that the US consumer is saddled with too many assets and won't spend.

  9. JP Koning, DavidR:

    As the Fed notes in above statement, both institutional and retail investors are moving into deposits. Thus, households are a part of the development. Moreover, the flow of funds data is the household sector so it is being reflected there too. (Yes, hedge funds are included in the household sector data which may dirty the data up a bit. But even then, hedgefunds manage funds for wealthy households and so even here we are capturing household funds to some extent.)

    The bigger point, though, is that portfolios are being rebalanced toward more liquid, less risky assets. Whatever the motive, this is an increase in demand for money--there is no way getting around it. That means the portfolio channel of monetary policy is working in reverse.

    In normal times, investors if they have too many money assets will search out higher yielding, slightly riskier assets like stocks and corporate bonds. As the yield on these go down from the increased demand for them, investors start looking at other assets like capital, real estate, consumer durables, etc. This affects nominal spending indirectly through wealth effects and improved balance sheets and directly as eventually the latter types of assets are purchased. When money demand rises like it is currently doing, this portfolio adjusting process is reversed.

    This is why velocity--the amount of times money is spent--is so closely tied to amount of money assets households are owning. See the figure in the update link in the post.

    Finally, to answer DavidR's question, the precise answer is that households, firms, and other sectors have an excess money demand problem. Their demand for money assets is not satiated given the current economic uncertainty and the current price level. This heightened demand is rational for each entity, but collectively it results in lower nominal spending. Their demand for money balance will fall once the economic outlook improves, but the economy won't improve until money demand falls. So yes, there is a circularity problem here. The one way to break it is to forcefully adjust expectations. The Fed can do that.

  10. David,

    Thanks for the response. You weren't very specific in your answer though. Do you think the US consumer has too many money assets in the aggregate? This seems to conflict with most other data points we see on a regular basis. If anything, consumers can't obtain enough money assets to deal with their every day needs.

    Can you elaborate. Thanks!

  11. DavidR,

    The households hoarding liquid assets are the creditors, the ones with excess funds but are reluctant to invest or spend them. So, yes, I am speaking mostly about wealthier households. I suspect there are less wealthy households too that are very cautious about their funds and acting similarly.

  12. So, isn't this more of a wealth disparity problem and not a problem of money hoarding? If the poor had more money they'd spend it? Why bother with all the monetary policy gimmicks that MIGHT work when we know we can solve your problem by taxing the rich and giving the poor a tax break?

    Why not fix the problem with fiscal policy that we know for a fact will fix the problem you describe?

  13. "The bigger point, though, is that portfolios are being rebalanced toward more liquid, less risky assets. Whatever the motive, this is an increase in demand for money--there is no way getting around it. "

    Sorry David, you don't convince me.

    What you are missing is the possibility that investors are simply switching from certain types of dollars - those dollars most exposed to Europe - into other types of dollars - those most exposed to the US.

    There is, therefore, no reason to assume that has been a net increase in the demand for money, just a switch from one type of dollar to another.

  14. Okay JP, let me explain it this way. The changes you are describing is what is typically called an increase in precautionary money demand. This type of money demand is different than transaction demand for money. When the former happens, investors move to safer, more liquid assets this leads to a move away from riskier assets and reverses the portofolio channel (i.e. hot potato effect in reverse).

    I am not sure why this isn't more clear. Investors are getting out of riskier asset and moving into safer ,more liquid ones. By definition, they are avoiding riskier assets which has a ricocheting affect on even riskier. Wealth is reduced and balances sheets weakened. Spending falls. This is an rise in money demand.

  15. JP,

    One last thing. How do you interpret the fact that if you take M2 and institutional money market funds to it (or equivalently add small time deposits to MZM) you get an increase in the aggregate? In other words, the monetary aggregates are increasing on net, even when we account for retail and institutional money market accounts.

  16. David,

    You never answered my question on fiscal. Touchy political point for you?

    Also, you're just cherry picking the most recent M2 spike. MZM and M3 fell substantially throughout the entire recession and only just recently began to rise again. How can you claim this recession has been due to money demand when MZM and M3 were clearly on the decline from 2008-2010?

  17. You came pretty close to describing my side. I'll repeat your post almost word for word except with a few changes to illustrate where I am coming from:

    Okay JP, let me explain it this way. The changes you are describing are what is typically called an increase in precautionary asset demand. This type of demand is different than the liquidity demand for assets. When the former happens, investors move to safer assets.

    "I am not sure why this isn't more clear. Investors are getting out of riskier asset and moving into safer ones.... This is not an increase in liquid asset demand, or put otherwise, it is not an increase in money demand."

  18. Re: Your point on M2 and institutional money market funds.

    Fair enough, but my understanding of the tally of money-assets in the shadow banking system is that it includes many other categories than just institutional money market funds. Take commercial paper, which was down quite dramatically in July and August.

    See back page chart: