Required reserves: much smaller than you think

The blogosphere has already responded in force to Dean Baker’s strange proposal to destroy the $1.6 trillion in Treasuries held by the Fed and use increased reserve requirements to maintain the Fed’s balance sheet once QE is withdrawn. Greg Mankiw poses it as an “exam question”, and rightly observes that the proposal is a mixture of accounting gimmickry and financial repression. (To be fair, the “accounting gimmick” is part of the intent: the idea is to circumvent the statutory debt limit.)

But there’s also a simpler, quantitative problem with Baker’s analysis: he doesn’t seem to have any clue how small required reserves currently are. Let’s take a look:

Currently, banks need to hold only $80 billion in reserves. And that’s in the midst of an ongoing recession, where consumers and businesses have plowed money into liquid assets and, through interest on reserves, the Fed has eliminated the cost of keeping money in accounts with a reserve requirement. Before the spectacular recent increase, the level of required reserves was closer to $40 billion.

The difference between $40 billion and $80 billion, however, is minimal compared to the $1.6 trillion that Baker proposes to capture using a reserve requirement. Currently, the reserve requirement is set at 10% of checking account balances. Even if we increased the reserve requirement to 100%, we’d only be halfway to $1.6 trillion—and that’s under the extraordinarily unrealistic assumption that these balances would stay at their recessionary highs even after a vast expansion in the cost of holding reserves!

How hard is it to get to $1.6 trillion? Let’s say that we broadened the scope of the reserve requirement to cover everything contained in the M2 aggregate: savings deposits, money market deposit accounts, small-denomination time deposits, and retail money market funds. That’s roughly $9 trillion, or $8 trillion after we subtract currency. $1.6 trillion is 20% of $8 trillion. So yes, we can induce a demand for reserves of $1.6 trillion by doubling the reserve ratio and vastly expanding the pool of deposits covered. But even this ignores the fact that depositors would quickly abandon the assets covered by the new requirement, to the point where the total would be far less than $8 trillion.

Of course, if you cast a wide enough net, it’s quite possible to reach $1.6 trillion: US households have almost $50 trillion in financial assets. But you have to ask why this would be a remotely desirable way to raise money. Reserve requirements are effectively a tax on deposits: at the very least, banks are forced to sacrifice the spread between perfectly safe assets (like T-Bills) and reserves (which have traditionally paid zero). Why does Baker think this is a efficient tax? It seems awfully bizarre to me: not only are you taxing savings, which is already inefficient, but you’re restricting the tax to a certain form of savings, which is all the more arbitrary and inefficient. (Not to mention regressive—Grandma’s savings account is taxed while hedge funds are not.)

To be fair, there are serious proposals to use reserve requirements as a regulatory tool, taxing risky liquidity creation to bring it down to the socially optimal level. But to do this properly, you’d need to completely redefine the requirements’ scope: the real risk to the financial system comes from “shadow banking” like repo and commercial paper, not traditional retail deposits. Is this what Baker is proposing? If so, that’s fine—but he certainly doesn’t give any hint of it.



Filed under macro

21 responses to “Required reserves: much smaller than you think

  1. TC

    Pretty sure that Dean knows reserves are really small. He doesn’t say we should increase RR, only that we could increase RR should excess lending become a problem.

    Part 1 is pure MMT.

    But part 2 is where it gets interesting, because Mankiw fails he own exam.

    Mankiw failed the exam for three reasons:

    1. The lending isn’t a problem right now, the demand is the problem. At least according to the Dallas fed.

    2. He misread Baker. Baker is not advocating for an increase in RR. He’s proposing a method to control the amount of lending, which is change RRs.

    3. Increases in RR are no more financial repression than raising the FF rate would be right now. Increases in RR have similar ending economic impacts to higher FF rates. Of course FF rates are not at market interest rates – that’s the entire point of FF rates! That would be the point of changing RR too!

    The Fed doesn’t pay market rates in the FF market. It doesn’t pay market rates so it can stimulate/destroy lending. But you can get similar effects by keeping FF rates locked and increasing/decreasing RR.

    I mean look at this quote:

    ” Assuming the Fed does not pay market interest rates on those newly required reserves, it is like a tax on bank financing. The initial impact is on those small businesses that rely on banks to raise funds for investment. The policy will therefore impede the financial system’s ability to intermediate between savers and investors. As a result, the economy’s capital stock will be allocated less efficiently. In the long run, there will be lower growth in productivity and real wages.”

    It’s probably the dumbest thing anyone could say in response to Bakers proposal, because Baker was explicitly trying to show another way the fed could impact lending besides meddling with FF or QE.

    • Increases in RR are no more financial repression than raising the FF rate would be right now. Increases in RR have similar ending economic impacts to higher FF rates. Of course FF rates are not at market interest rates – that’s the entire point of FF rates! That would be the point of changing RR too!

      The effects aren’t similar at all. Right now, an increase in required reserves does absolutely nothing—there are already plenty of excess reserves. In the future, if there are no longer so many excess reserves, an increase in required reserves will effectively be a tax on a certain kind of financial intermediation—accepting deposits to a checking account and then lending them out. (If the requirement is expanded to new types of deposits, it is also a tax on them.)

      The federal funds rate, on the other hand, determines the baseline risk-free nominal interest rate, which plays a pivotal role in determining the yield on every loan and asset in the economy. It’s vastly more important.

      One confusing point is the fact that under traditional policy, it was not possible for the Fed to freely set the FFR if there were excess reserves in the system; bankers with extra money on their hands would quickly bid the FFR down to zero. In such an environment, an expanded reserve requirement would be necessary to raise the FFR above zero after the recession ended (assuming the Fed didn’t want to wind down its balance sheet), and would therefore be pivotal in the conduct of monetary policy.

      Now, however, the Fed can simply adjust the interest rate paid on reserves to move the FFR. Reserve requirements are much less relevant.

      • TC

        Both the ECB and the PBOC use raising reserves to reduce the amount of lending in the system.

        The Securities Market Programme (SMP) is the mechanism by which the ECB purchases sovereign debt. It takes these “out of circulation” by raising reserve requirements:

        The PBOC has raised reserve requirements several times in the last year. The widely expected impact? Reduced lending – at least according to Bloomberg and the WSJ.

        The goal is to reduce lending by banks. According to the traditional view, this can be done by either raising interest rates or by raising reserve requirements.

        We’re in a strange place because of the ZLB, and current gigantic excess reserves. But this doesn’t matter to Bakers claims. His only claim was that, yes, we can drain these funds from the banking system if we desire.

        The precise impact on the banks isn’t relevant, the impact on lending is the important part to the Central Bank.

        This is why Mankiw is wrong. He doesn’t understand the purpose of basic bank operations, even in the traditional framework. He knows it causes a tax, but not why the central bank might want to impose a tax on savings, even if it isn’t perfectly efficient. He knows the notes, but not the music.

        Yes, RR are less relevant in the current U.S. situation. I’d also argue they are nearly totally irrelevant relevant due to operational realities of the banking system from an MMT viewpoint. But to argue they don’t matter from within the traditional paradigm is beyond foolish, and this is what Mankiw is doing.

        I don’t think you’re doing exactly that, but you’re close. You’re claiming that we a have huge amounts of excess reserves, so it would take a huge increase to have an impact. To me, this is mistaking the notes for the music. Baker isn’t talking about right now – when we don’t want to raise RR. He’s talking about the future, when we might want to raise RR. If it gets to that point, we can and should use every policy option we have. Raising RR is one of those policy options.

  2. You suggest toward the end the taxing savings is always inefficient. But in a liquidity trap excess savings is the problem confronting policy makers, not insufficient savings. So its not clear why, under liquidity trap conditions, a tax on savings would be inefficient.

    • It is true that a tax on savings can be beneficial during a liquidity trap. But Baker’s proposal doesn’t tax savings during a liquidity trap: reserve requirements are not costly when the federal funds rate is at zero, or equal to the rate paid on reserves.

      Viewed in this light, Baker’s proposal looks even worse: it taxes savings when they should not be taxed (under normal, positive interest rate conditions) yet fails to tax them in the one instance where it’s appropriate (a liquidity trap).

      In fairness, it is very difficult to implement an effective savings tax when interest rates are already at zero: people can always just hoard cash, and it’s awfully hard to enforce a tax on paper currency.

  3. Uh what? Dean didn’t say that you have to increase reserve requirements by the same amount than the debt destroyed. And why would you assume that? What he said is that if you are concerned with the ability of the Fed to control liquidity (given that destroying the bonds will reduce their ability to sell them) one can use other instruments, like reserve requirements. Mind you, there is probably no need to reduce liquidity anyway.

    • I agree that he didn’t explicitly say that, but only because he was so hopelessly confused on this issue that he wasn’t really making any sense.

      If the monetization of debt is to be fiscally beneficial, the interest rate paid on reserves must be lower than the rate on T-bills; otherwise you might as well have kept the debt at the Treasury. The rate on T-bills can only rise above the rate on reserves, however, if there is a liquidity premium on reserves, and that can only happen if the vast supply of excess reserves in the system is soaked up by some kind of new reserve requirement. You need the reserve requirement to match the magnitude of the supply of reserves; otherwise, the liquidity premium will collapse to zero (at the margin, there is no use for reserves at the Fed except to satisfy mandatory requirements). That’s why it’s important to compare the magnitudes.

      • The important thing is not if the rate of interest on reserves is higher or lower than on government bonds, but how much interest the Fed has to pay with the increase in reserves vis-à-vis what the Treasury would have if it didn’t destroy the bonds. If the increase in reserves is small, even with a higher rate of interest, the interest bill would be smaller. The only reason to be against the proposal is if you think that inflationary risks are of such magnitude that a huge increase in reserve requirements would be needed to contract liquidity. I think Dean’s point, with which I would agree, is that this is a very unlikely scenario. Mind you, given the political position of Obama, that seems to be willing to cut spending on Medicare, and reduce Social Security benefits, it is completely impossible to think that Dean’s proposal would have a chance to pass.

      • I am baffled by this statement:

        The important thing is not if the rate of interest on reserves is higher or lower than on government bonds, but how much interest the Fed has to pay with the increase in reserves vis-à-vis what the Treasury would have if it didn’t destroy the bonds. If the increase in reserves is small, even with a higher rate of interest, the interest bill would be smaller.

        The “increase in reserves” (relative to the baseline scenario where the Fed retires many of these reserves) is the same as the debt forgiven: the Fed can no longer sell the bonds it has cancelled, and has to leave an *equivalent* amount of reserves on its liability sheet instead. (Maybe in this scenario it sells other assets more than it otherwise would, but then it’s losing out on the yield from those other assets, which doesn’t help matters any.) Thus the amount of interest paid by the Fed on reserves, relative to the amount of interest paid by Treasury, is directly connected to the interest rates paid by each. And there will only be a gap between these two interest rates if there is a liquidity premium on reserves, which can only happen if the reserve requirement is increased to a level commensurate with the reserves in the system.

        I’m not sure what distinction you’re trying to draw here, but I certainly don’t understand it.

  4. Pingback: Avoiding the word “tax” | Matt Rognlie

  5. Again, you seem to complicate relatively simple things. Perhaps you not Dean is the one “hopelessly confused”, as you said. The government pays interest on the bonds. And so does the Fed on reserves. The overall payments of the government are the sum. The elimination of the bond, as per Dean’s proposal, eliminates the interest payments associated to them. Again, only if you assume that the increase in reserves would have to be of the same size of the eliminated debt there might be a problem. Since that is not the case the interest payments that the Fed would have to do on reserves should be way smaller than the payments forgone by the treasury.

    • First of all, the “overall payments of the government”, which you define as a combination of Treasury and Fed payments, certainly shouldn’t include interest payments the government is making to itself (Treasury to Fed). If you think that the Treasury and Fed should be treated as separate entities with separate budget constraints, that’s fine (though it would make analyzing this policy impossible), but you can’t combine their external payments and include payments from one to the other in a measure of “total payments”. That’s not coherent.

      Since we are discussing issues that are ultimately quantitative, perhaps I should provide an explicit quantitative example. (The numbers are not meant to be exactly right; they’re just to illustrate the conceptual point.) Suppose that the Fed has $3 trillion in assets, of which $1.5 trillion are Treasuries. (For simplicity, let’s suppose that all assets pay the same as T-bills. Since the Treasury assets are mostly long-duration assets, and the non-Treasury assets are mainly MBS, this is not quite true without adjusting for risk and duration, but it’s not so far from the truth either.) The Fed also has $3 trillion in liabilities. Under the existing schedule of reserve requirements and currency demand, the maximum supply of base money that is consistent with the FFR being greater than IOR is $1 trillion (of which reserves will be a very small portion); otherwise, base money will be in such excess that it will no longer carry any liquidity premium, and the FFR, the IOR rate, and T-bill rate will all be roughly the same. (As they are now.)

      The Fed plans to return to the usual situation of FFR > IOR, which means selling $2 trillion in assets and removing $2 trillion in liabilities, until it hits the $1 trillion mark where a liquidity premium reemerges. To be concrete, let’s suppose that in this case it sells off all its Treasury debt. (Its choice about what asset to sell doesn’t actually matter as long as the assets are priced fairly.)

      Now suppose that it destroys the $1.5 trillion in Treasuries. Now it has $3 trillion in liabilities and only $1.5 trillion in assets. Obviously it can’t get all the way down to $1 trillion in liabilities; it’s insolvent and doesn’t have the assets to sell. How low can it go? Well, it needs to retain some assets. For the sake of argument, let’s suppose that it’s risk-loving and is willing to tolerate $500 billion in assets backing $2 trillion in liabilities. (Only 25%!) It sells $1 trillion in assets to reach this point.

      Now we still have $2 trillion in liabilities. That’s way more than $1 trillion, which is the level necessary for a liquidity premium to emerge and to have FFR > IOR. If the situation stays this way, there is no fiscal benefit.

      Why? Well, the previous plan was for the Fed to pay interest on $1 trillion in liabilities while gathering returns on $1 trillion of non-Treasury assets, and for the Treasury to pay interest to some external party on the $1.5 trillion of its bonds that the Fed sold. Ignoring the Treasury’s other debt, the net position of the Fed is $0 (though it has seignorage revenue from the fact that its liabilities pay less than its assets), while the net position of the Treasury is -$1.5 trillion.

      Alternatively, if we destroy the $1.5 trillion in Treasuries, the Fed is now an additional $1.5 trillion in the hole, while the Treasury is up to $0. Is there any net interest savings? Not at all! To be explicit, let’s write:

      Status quo ex-ante:
      Fed pays: (0%)*($1 trillion) = $0
      Fed earns: (FFR%)*($1 trillion)
      Treasury pays: (FFR%)*($1.5 trillion)
      Net external payment: FFR% * $0.5 trillion.

      Cancelling Treasuries scenario:
      Fed pays: (0%)*($1 trillion) + (FFR%)*($1 trillion)
      Fed earns: (FFR%)*($0.5 trillion)
      Treasury pays: $0
      Net external payment: FFR% * $0.5 trillion.

      It’s the same result—you pay the prevailing interest rate (which I write as the FFR%) on the part of your debt that is not cash or interest-free reserves. (In fact, the second situation is slightly worse, since the Fed is not actually paying 0% on the full $1 trillion; a sliver of that is supposed to consist of reserves as well.)

      How do we alter this result? Well, we can’t convince the public to hold more cash without lowering the trajectory of the FFR, which in the long-run must mean lower steady-state inflation; even then, their demand is pretty inelastic. Our only real option here is to change the reserves from “interest-paying” to “interest-free”. But conditional on a particular trajectory for the FFR%, the only way to do this is to raise the reserve requirement so much that a liquidity premium reemerges. For this to happen, we need to create $1 trillion in additional required reserves. And that is what Dean Baker doesn’t seem to understand.

  6. Paul

    Matias you said: “Again, only if you assume that the increase in reserves would have to be of the same size of the eliminated debt there might be a problem.”

    My question and I think Matt’s point here (correct me if I’m wrong) is that they must be the same size or pretty damn close if you don’t want to risk inflation since any debt that is eliminated that is not counteracted by increased reserves is essentially monetized. Since required reserves are currently 80 billion with the RR being 10% there is no RR in the scope of current policy that would ever come close to soaking up the current 1.6 trillion excess in the system. This means Baker’s proposal either requires a drastic redefinition of required reserves (which he doesn’t mention) or a ton of the debt has to be monetized if we are to destroy all 1.6 trillion of the bonds (the more you do of this the riskier it is).

    If you don’t drastically redefine the RR and don’t want to monetize a significant amount of the debt then you are unable to destroy anything more than peanuts of the current excess.

    • Thanks, Paul. This is the essence of my argument, though I am long-winded in my response to Matias because I think that conventional terms like “monetization” are a little ambiguous.

      Basically, if we put a lot of money into the system, then the liquidity premium will fall to zero and we’ll have IOR% = FFR% = T-bill% (approximately). In this world, there is no benefit from having debt in the form of “money” rather than bonds; they pay the same rate.

      The exact level of IOR and FFR is not so clear—we only know that they’ll be (almost) equal. Under conventional, pre-IOR policy, we’d be forced to have FFR = 0%, which would almost surely be inflationary. I think it is more plausible to assume, however, that the Fed will continue to move the Fed Funds Rate to achieve its monetary policy goals in the same way as before, and adjust the IOR as necessary to do so.

  7. Hi Paul. That’s what I implied, that Matt argues that both have to be the same size because of the need to control inflation. I think that Dean (have not discussed it with him) is not concerned with inflation, or at least not to the point that he would believe that reserve requirements would have to increase nearly that much. I wrote more here if you are interested

  8. axiom

    One complication is that it forces the Fed into a negative capital position.

    That’s how its balance sheet balances.

    Which it must, if it remains a stand-alone institution.

  9. Paul

    I think the real problem is that Baker doesn’t really make it clear what his views are with regards to monetization vs actually having to soak up the total amount of excess reserves. He either thinks simply increasing the RR on deposits is sufficient (which it is not) or he doesn’t want to fully develop the idea (i.e. admitting he is either expanding the RR to apply to more than deposits or monetizing the debt) to make the proposal seem less radical. Even in his followup to Mankiw he still only mentions reserves on deposits and does not mention monetization. His example of doubling the RR for twice as many reserves is irrelevant given that excess reserves are practically 20 times what required reserves are and the RR cannot be increased by 20 times. Again if he believes monetizing a significant portion of the debt is not a problem then he needs to come out and say so. If he believes that we need to be able to soak up all of the excess then he needs to explain what kind of reserve requirement would actually be capable of doing it, but he cannot accurately claim that we can have our cake and eat it too (i.e. not monetize any debt, destroy all the bonds and only have to slightly fiddle with the RR on deposits). Matt’s original post I think is based on the assumption that Baker is unaware he would have to make such a choice, but unless he gets more specific we can’t know the answer to that question. Perhaps there is another explanation, but I don’t see it.

  10. vimothy

    Hi Matt,

    I like Mankiw’s exam question. But I’m a bit confused by it—part 1, at least. Surely there are always winners and losers.

    It seems to me that an equivalent question would be: Suppose the govt stops issuing debt; who loses? Or, suppose the govt funds its net spending by having the central bank buy all of its newly issued bonds, Weimar-style; who loses?

    I’m pretty sure that the correct answer is not “no one”. On the other hand, I am not at Harvard so perhaps I am used to a lower standard of exam question. Am I missing some subtle point of theory?

    If the govt does not in fact spend the cash it acquired via this “accounting gimmick”—perhaps it is buried under the White House lawn, or used to wallpaper the Fed—then I can agree that it is indeed no more than an accounting gimmick. On the other hand, if spend it does, then it follows that real resources have been transferred as a result and that the policy has some definite consequences for the economy and its participants.

    But perhaps my assumption that the govt spends the funds it has so acquired is unwarranted and not made by Mankiw. Or perhaps there is something else. What say you?

    • vimothy

      Obviously I’m assuming here that the Fed wasn’t going to hold the bonds to maturity anyway. I agree that there’s no difference between the Fed shredding the bonds today and the Fed holding the bonds to maturity and then handing the principal and interest back to the Treasury.

  11. Pingback: The reserve requirement | princetonseoul

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )


Connecting to %s