Does liquidity matter? Wrong question.

Josh Hendrickson has an interesting post discussing how he approaches monetary theory:

An additional insight of the New Monetarist literature is that the way in which we define “money” has changed substantially over time.  A number of assets such as bonds, mortgage-backed securities, and agency securities are effectively money because of the shadow banking system and the corresponding prevalence of repurchase agreements.  As a result, if one cares about quantitative targets, then one must expand the definition of money.  David Beckworth and I have been working on this issue in various projects.  In our paper on transaction assets shortages, we suggest that the definition of transaction assets needs to be expanded to include Treasuries and privately produced assets that serve as collateral in repurchase agreements.  In addition, we show that the haircuts of private assets significantly reduced the supply of transaction assets and that this decline in transaction assets explains a significant portion of the decline in both nominal and real GDP observed over the most recent recession.

I prefer to define “money” as the base.  Before explaining why, let me point out that this is purely a question of semantics—there is no generally accepted definition of money, and hence the real question is; which definition is more useful?  I like the base for two reasons; it is the medium of account, and the quantity of base money can be directly controlled by the Fed.  Let’s call close substitutes like T-bills “highly liquid assets” (HLAs.)  These assets have market prices that change as market conditions change.  In contrast, the price of base money is always one, and when the market for base money is hit by a shock, the nominal price doesn’t change. Instead, the prices of all other goods adjust.

In fairness to the other side, I can imagine scenarios where the stock of HLAs might be a more interesting variable than the stock of base money.  For instance, suppose the stock of HLAs is much more closely correlated with NGDP than the stock of base money.  Also assume that policymakers can easily control the stock of HLAs.  In that case it might make sense to focus on HLAs.

Now let’s consider how a shortage of HLAs would impact the macroeconomy, from the perspective of a base-obsessed economist like me.  We know that by definition:

P = M/(M/P)   Where M is the monetary base and P is the price level.

This means that changes in the price level results from either changes in M, or from M/P.  In equilibrium, M/P is the real quantity demanded of base money.  Thus if a shortage of HLAs is deflationary, for any given stock of base money, it is presumably because the shortage pushes AD and the price level lower by increasing the demand for base money.  Which makes sense; banks unable to get their hands on enough T-bills might hold more reserves.

Many debates about “causality” are disguised debates over policy counterfactuals.  Consider people arguing whether the actual cause of deflation is too little base money or too little HLAs.  They can be implicitly seen as debating which policy counterfactual would have been more effective; lots of open market operations (OMOs), or the issuance of lots more HLAs.

However I find these debates to be frustratingly incomplete.  It’s simply not possible to model the impact of a particular policy “gesture,” without embedding that action in a complete policy regime.  Here’s Hendrickson:

Seemingly lost in the discussion of monetary policies various QEs is a meaningful resolution of our understanding of the monetary transmission mechanism.  Sure, New Keynesians argue that forward guidance about the time path of the short term nominal interest rate is the mechanism, Bernanke argues that long term interest rates are the mechanism, and skeptics of the effectiveness of QE argue that it is the interest rate on excess reserves that is the mechanism.  I actually think that these are not the correct way to think about monetary policy.  For example, there are an infinite number of paths for the money supply consistent with a zero lower bound on interest rates.  Even in the New Keynesian model, which purportedly recuses money from monetary policy, the rate of inflation is pinned down by the rate of money growth (see Ed Nelson’s paper on this).

This is an important point.  Before discussing the implications, consider an analogy using the following two policy options for a flu outbreak:

Option A:  Have a doctor and nurse visit everyone and give them . . .

Option B:  Broadcast to everyone the importance of frequently . . .

Is it possible to evaluate those options?  Clearly not.  They are incomplete.  If we added “a flu shot” to the first, and “washing hands” to the second, we could evaluate the two policy options.

I’d say the same about OMOs vs. the creation of more HLAs.  They are not embedded in a fully fleshed out policy regime.  For example, the creation of more HLAs might be very effective in a gold standard regime, where the supply of base money is relatively stable.  In contrast, it might be completely ineffective in a regime where the central bank is targeting inflation at 1.5%.

If we have a fully fleshed out policy regime where the central bank’s target is also the goal variable (say NGDPLT with a 4.5% trend line) then neither transitory OMOs nor the creation of HLAs will have any significant effect on the long run path of NGDP, and hence very little impact on current NGDP.  On the other hand if policy is “adrift” with no clear target, either OMOs or the creation of HLAs might boost NGDP expectations, because they would send a signal about future expected monetary policy, and hence the future path of NGDP.

However there is a asymmetry, which explains why I focus on the base.  The Fed can actually control the base quite effectively, whereas they cannot control the total stock of HLAs nearly as precisely.  So that raises another question; if the Fed is using its control of the base (as well as expected future changes in the base) to target NGDP, can one meaningfully talk about a government policy aimed at addressing the “shortage” of HLAs?  One answer might be that while more HLAs would not impact NGDP, it would allow the Fed to do its NGDP targeting with a smaller balance sheet.  Recall that the Fed is reluctant to do aggressive base expansion because they worry about the long term effects of a bloated balance sheet.  In that case an expansionary HLAs policy would allow the Fed to hit its target with less worry about excessive balance sheet expansion.  Some Keynesians would go even further, arguing that fear of a large balance sheet would cause the central bank to fall short of its NGDP target, and hence producing more HLAs would help to achieve that target.

I see that as the best argument for the creation of more HLAs, although not necessarily “fiscal stimulus” in the ordinary sense.  The Treasury could simply issue lots of debt and use the money to set up a sovereign wealth fund.  (Perhaps with European, Canadian, Australian and Japanese equities, if you are worried about socialism.)  I seem to recall Miles Kimball making a similar argument.  I still regard that sort of approach as “banana republic” policy, not worthy of a great nation.  I’d prefer a simple OMO technique, combined with NGDPLT.  But then maybe I’m too old-fashioned.

PS.  That last point relates to my often misunderstood opposition to (demand-side) fiscal stimulus.  Many people seem to believe that I feel very strongly that the fiscal multiplier is zero.  Not so, I feel very weakly that the fiscal multiplier is roughly zero.  I feel very strongly that the fiscal multiplier should be zero.  Banana republics have non-zero multipliers.  Is that what we want to be?


Tags:

 
 
 

23 Responses to “Does liquidity matter? Wrong question.”

  1. Gravatar of Joe S Joe S
    3. June 2013 at 05:50

    “HLA” became “HSA” in a couple of early paragraphs

  2. Gravatar of Jim Ancona Jim Ancona
    3. June 2013 at 06:00

    Should the references to “HSAs” in paragraphs 3 and 4 be changed to “HLAs”?

  3. Gravatar of OhMy OhMy
    3. June 2013 at 06:24

    “quantity of base money can be directly controlled by the Fed.”

    Incorrect. The CB cannot even attempt to set the base lower than what the market demands – failures of solvent banks would the outcome. Setting the base higher has very little consequences apart from crushing the rate. And banks look at the rate not the base. It follows that the rate and not the base is the CB’s tool. Bankers note that the volume of the base has almost no information content:
    http://www.mnb.hu/Root/Dokumentumtar/ENMNB/Kiadvanyok/mnben_mnbszemle/mnben_szemle_cikkei/bulletin_2007june_komaromi.pdf

  4. Gravatar of ssumner ssumner
    3. June 2013 at 06:34

    Thanks Joe and Jim, I corrected it.

    OhMy, A permanent increase in the base leads to a proportional increase in the long run expected level of NGDP.

    I agree that some very contractionary monetary policies might be politically infeasible, but that has no bearing on my post.

  5. Gravatar of OhMy OhMy
    3. June 2013 at 07:24

    “OhMy, A permanent increase in the base leads to a proportional increase in the long run expected level of NGDP.”

    I understand that this is an article of faith, since there has not been an even in history which we would call a “permanent increase in the base”, so you are not speaking from experience, right?

  6. Gravatar of Philo Philo
    3. June 2013 at 07:35

    “. . . the Fed is reluctant to do aggressive base expansion because they worry about the long term effects of a bloated balance sheet.” Is this a realistic worry? I have never understood what, exactly, those (possible) deleterious long-term effects are supposed to be.

    To inject more HLAs into the economy (as an alternative to injecting more base money, as would be done through ordinary OMOs): “The Treasury could simply issue lots of debt and use the money to set up a sovereign wealth fund.” Or the Federal Government could simply offer its guarantee of various bonds being issued by private corporations (with the corporation paying a fee for the guarantee); this would turn the bonds into HLAs. But why would anyone think that injecting HLAs was *better* than injecting base money? HLAs are important only for their contribution to liquidity, and base money constitutes a purer contribution to liquidity. In short, I do not at all understand the Hendrickson/Beckworth obsession with HLAs.

  7. Gravatar of Bill Woolsey Bill Woolsey
    3. June 2013 at 07:39

    Scott:

    I agree. But..

    What is missing from your analysis is what the Fed purchases with its OMO. If it purchases HLA, it is increasing the base while reducing other HLA.

    I don’t think this point makes your argument wrong, but rather it is something you should deal with.

    I also think that the Divisia measures of the quantity of money have great value under current conditions. The broad Divisia measures of the quantity of money have shifted to lower growth paths, just like nominal GDP. The monetary base shifted to a much higher level. You and I have long been quite open to the notion that the demand for base money could be subject to large changes. Those with a more traditional monetarist focus instead want to assume that the demand for money is very stable. Using a better measure of the quantity of money than either the traditional M2 or the base, the quantity of money is on a lower growth path. It fell, it is rising, but it has not recovered to its past trend growth path. There has not been an inflationary increase in the quantity of money using a more appropriate measure of the quantity of money. To a large degree the huge drop off in overnight repurchase agreements, that were not measured by the Fed, involved a major decrease in the quantity of money.

    Again, all of this, even the implications of open market purchases of other highly liquid assets, can be analyzed as impacting the demand for base money.

    Also, rather than the “problem” being a 1.5% inflation rate, I think the problem is that if the target requires nominal GDP remain on a lower growth path and that nominal wages and other sticky prices shift to a lower growth path, then expanding the quantity of highly liquid assets or base money will do no good.

  8. Gravatar of Mark A. Sadowski Mark A. Sadowski
    3. June 2013 at 07:59

    Scott,
    Your use of terminology is sometimes variable. When you say in this context “base money” do you mean the “monetary base” in the conventional sense, as in:

    Currency Component of M1 (i.e. “currency held by the public”)

    +

    Reserves (required and excess)

    +

    “bank vault cash” (which together with the Currency Component of M1 forms Currency in Circulation)

    Or do you mean some subset of the above?

  9. Gravatar of Josh Josh
    3. June 2013 at 08:49

    Scott,

    I agree that the monetary base is the only thing that the Fed directly controls, but over short horizons I don’t think that the base is the best indicator of the stance of monetary policy. The stock of transaction assets as David and I define in our work together is a very good predictor of the variation in NGDP.

    I would echo Bill’s point above: “I also think that the Divisia measures of the quantity of money have great value under current conditions. The broad Divisia measures of the quantity of money have shifted to lower growth paths, just like nominal GDP. The monetary base shifted to a much higher level.”

    David and I actually argue in a paper that we are currently writing that the effect of monetary policy in real economic activity is through expectations about the future path of both the supply of transaction assets and the expected path of nominal income. This is explicitly derived in the context of a model in which (1) transaction assets are essential (contra New Keynesian land), (2) consumers smooth their consumption, (3) households face borrowing constraints, (4) the only thing that Fed controls is the monetary base. It is hard to describe in a blog comment. Perhaps I will do a post on it or just send you the paper when its finished.

  10. Gravatar of Doug M Doug M
    3. June 2013 at 08:52

    The government can’t really control the supply of HLAs.

    If the treasury issued a trillion dollars in t-securites (increasing HLA’s) to create a sovereign wealth fund (buying less liqid assets), they would be crowd out the banks, SIVs, and CLOs, that have been the primary provider of HLA’s to the market.

    While the Fed should be keenly aware of the quantity of HLAs supplied to the market, and may think about how to manipulate banks incentives to create HLA’s, they should not try to target the supply directly.

  11. Gravatar of Doug M Doug M
    3. June 2013 at 08:57

    A permanent increase in the base leads to a proportional increase in the long run expected level of NGDP.

    In which case you may say that the Fed should follow a “Friedmanesque” policy of growing the base at a constant growth rate.

    Of course, the long-run doesn’t matter for a hill of beans. If, as a Fed governor, you don’t get the economy on track in the relatively short term, you will be replaced by someone advocating a more activist policy.

  12. Gravatar of John John
    3. June 2013 at 09:26

    I think the best way to decide if something is money or not is to ask yourself, “Does it function as a final payment for goods and services?” In this sense, credit is not money because you as the purchaser have to pay for your purchase at a later date from the lender. However, all demand and checking accounts do act as currency. If it isn’t available for immediate use, it isn’t really money yet.

    You can’t directly buy anything with a MMMF or a treasury bill without first selling your holdings for real money first.

  13. Gravatar of Saturos Saturos
    3. June 2013 at 10:43

    Any thoughts on this Scott? http://www.ft.com/intl/cms/s/0/58c0cd38-c852-11e2-8cb7-00144feab7de.html#axzz2UkKN44ef

    John, I tend to agree with you, but Scott is determined to hold that whatever prices are generally measured in is a necessary and sufficient definition of money. We’ve had this debate already, ad nauseam.

  14. Gravatar of Rob Rawlings Rob Rawlings
    3. June 2013 at 11:17

    Picking up on what Bill said: “What is missing from your analysis is what the Fed purchases with its OMO. If it purchases HLA, it is increasing the base while reducing other HLA.”

    If the fed used “helicopter drops” rather than OMO (say financing tax cuts with newly created money) couldn’t this issue be avoided?

  15. Gravatar of Jeff Jeff
    3. June 2013 at 11:28

    There are more HLA out there now than ever before. If the Treasury wants to create even more of them, it can do so by shortening the maturity of its debt.

    However, maybe we should stop and ask just why all these HLA are needed. Ten years ago the economy was performing pretty well with many fewer HLA than now. So what’s different now?

    HLA are necessary to grease the ever-faster transactions generating machines on Wall Street. My guess is that what is really going on is an attempt to bury rent seeking in enough transaction levels to bamboozle both the regulators and the customers. Zebras don’t change their stripes, and we should all know by now how Wall Street operates.

  16. Gravatar of Saturos Saturos
    3. June 2013 at 11:39

    So apparently we’re the biggest success of “reformist conservatism”: http://au.businessinsider.com/the-reformists-and-me-2013-6

    Josh Barro supports NGDP targeting, but doesn’t get the Sumner critique. And apparently Jim Pethokoukis got this quote from a GOP Senate staffer: “It’s not that we don’t know about your ideas or understand them. We just don’t think they’re any good.”

  17. Gravatar of jknarr jknarr
    3. June 2013 at 14:02

    The notion that HLAs are cash equivalents is just not smart. The monetary base is a perpetual, zero-yielding (currency for you or I), fixed-price liability of the effectively-self-funding central bank.

    HLAs are extinguishing-at-maturity, yield-bearing, variable-price liabilities of any number of credit structures — all of which do not have the power to create base money and must obtain cash to service obligations elsewhere.

    All HLAs are derivatives on base money — with base money dollars the deliverable. Accordingly, the scarcity of base money determines the underlying value of the notional derivatives. If base money volume was doubled, HLA value would halve. If HLA volume was doubled, then base money would be unaffected — but in the lender-of-last-resort system of the Fed, the base very likely would have to expand rapidly into a HLA-liquidation/base-demanding event.

    That is, every HLA that the US Treasury creates is borrowed/taken from the private sector, and spent back into the private economy: effectively zero sum splitting of the monetary pie. The creation of base money expands the entire monetary pie for both the public- and private- econonomy.

    HLA shortages are a function of lower balance sheet quality — due to higher debt loads. Higher debt loads are a consequence of tight base money — ie lower NGDP, lower incomes, lower interest rates, higher debt. Can’t get people to borrow, even at close to zero rates because of over-leverage — ergo, HLAs are scarce.

    Base is base, debt is debt — they are not equivalent by a long mile. Are IOR reserves still base money? That’s a better question –they may now be 3m bill equivalents. Then, its perhaps better to ask whether they still have the essential characteristics of base money, and whether they ought to still count in the base measurement.

  18. Gravatar of ssumner ssumner
    3. June 2013 at 16:55

    OhMy, I generally assume that my readers know what the concept of “the long run” means in macroeconomics.

    Philo, I’m as mystified as you are.

    Bill, I don’t think that’s missing, as I acknowledge that the short run effects of a base expansion may be small (partly for reasons you indicate.) But I also indicate that it’s the long run effects that are most important. In the long run the base and T-securities are not close substitutes (unless we are in a permanent zero bound, in which case we should replace government debt with currency.)

    Regarding your other point about the aggregates, obviously I think a NGDP futures market is better than the aggregates. If we don’t have such a market, then I have an open mind on the utility of the newer divisia indices.

    You said;

    “Also, rather than the “problem” being a 1.5% inflation rate, I think the problem is that if the target requires nominal GDP remain on a lower growth path and that nominal wages and other sticky prices shift to a lower growth path, then expanding the quantity of highly liquid assets or base money will do no good.”

    Yes, that’s what I meant to imply.

    Mark, Yes I mean what you describe. But one correction—it makes no sense to talk about vault cash plus reserves, as vault cash is a part of reserves. You meant vault cash plus bank deposits at the Fed.

    Josh, You said;

    “I agree that the monetary base is the only thing that the Fed directly controls, but over short horizons I don’t think that the base is the best indicator of the stance of monetary policy.”

    No need to use the qualifier “but;” I agree that the base is a horrible indicator of the stance of monetary policy. I prefer NGDP expectations, or some other proxy. It’s an open question as to whether market prices or monetary aggregates are a better proxy for NGDP expectations. I suspect market prices, but can’t be sure.

    I look forward to seeing your paper.

    Doug, I agree it’s hard to control the stock of HLAs.

    You said;

    “In which case you may say that the Fed should follow a “Friedmanesque” policy of growing the base at a constant growth rate.”

    Actually this doesn’t follow at all from what I said about the base, as V is very unstable.

    John, In my view the most important attribute of money is that it is the medium of account. It’s role as medium of exchange is secondary.

    Rob, See my reply to Bill.

    Jeff, Along the same lines, if NGDP growth were higher, there’d be less demand for liquidity.

    Saturos, Yes on the “Sumner Critique,” but overall it’s a very good article by Barro.

    jknarr, Good points.

  19. Gravatar of JoeMac JoeMac
    3. June 2013 at 19:28

    Scott,

    What factors determine the demand for base money? For example, is the money demand for reserves different from the money demand for currency?

  20. Gravatar of Mark A. Sadowski Mark A. Sadowski
    3. June 2013 at 19:52

    Scott
    You wrote:
    “But one correction””it makes no sense to talk about vault cash plus reserves, as vault cash is a part of reserves. You meant vault cash plus bank deposits at the Fed.”

    Or in FRED speak, Vault Cash plus Reserve Balances with Federal Reserve Banks.

    As a result of the Koo post I’ve been trying to keep track of the monetary base components and the Federal Reserve doesn’t make it easy with all of the various adjustments for seasonal variation and for reserve requirements.

    To be technical, bank vault cash was not permitted to statisfy reserve balances until November 1959. And even then not all bank vault cash has been used that way since then. For example as of April only 80.3% was used that way, and on average 73.5% has been used this way since it was allowed.

    Interestingly the only way I’ve found to make the FRED components of the recent monetary base add up absolutely perfectly is when you add the weekly frequency of Reserve Balances with Federal Reserve Banks and Currency in Circulation together (neither of which is seasonally adjusted or adjusted for reserve requirements). This sum is exactly equal to Source Base (which is also not seasonally adjusted or adjusted for reserve requirements).

    I’d be interested in knowing if anyone can add it up perfectly in any other way.

  21. Gravatar of ssumner ssumner
    4. June 2013 at 11:49

    JoeMac, The interest rate on reserves, as compared to the interest on close substitutes.

    Mark, Thanks, That sounds right to me.

  22. Gravatar of OhMy OhMy
    5. June 2013 at 06:59

    “OhMy, I generally assume that my readers know what the concept of “the long run” means in macroeconomics.”

    That is a dodge. Which QE lasted long enough that you would consider it a long run or “permanent”? All QEs I know of coincided with falling inflation. Therefore my question: your pronouncement is an article of faith, right?

  23. Gravatar of Steve Roth Steve Roth
    5. June 2013 at 19:28

    “I prefer to define ‘money’ as the base.”

    So when we you say “money” — the M in MV = PY (or Q, or T) — you mean base money.

    And base money is defined as banks’ Fed reserve balances + currency + coins.

    So checking/saving account deposits, money market deposits, etc. aren’t money, or at least they’re not part of M in the equation of exchange.

    So when deposits change but reserves/coins/currency don’t (for instance due to bank lending), there’s no change in M?

    Is that right?

Leave a Reply