Five Thoughts on Monetary Policy

1. Monetary policy may operate on (a) the quantity of bank liabilities (money); (b) the quantity of bank assets (credit); (c) the price of one or more assets relative to money (an interest rate);  and/or (d) the price of money, normally relative to some other money (an exchange rate). Which of these should be considered the most immediate target of central bank policy, both practically and conceptually, has been debated for over 200 years. All four positions are well-represented in both academic literature and central bank policymaking. For the US over the past 50 years, you could say that the center of gravity — both in policy and in the economics profession — has shifted from the quantity of credit to the quantity of money, and then from the quantity of money to the price of credit. [*] I don’t know of any good historical account of these recent shifts, but they come through dramatically if you compare contemporary articles on monetary policy, ones from 20 years ago, and ones from 50 years ago.

Lance Taylor has a good discussion of the parallel debates in the 19th century on pages 68-84 of Maynard’s Revenge, and a somewhat more technical version in chapter 3 of Reconstructing Macroeconomics. Below, I reproduce his table classifying various early monetary theorists in the four categories above, and on the orthogonal dimension of whether the money/credit system is supposed to be active or passive with respect to the economy. Obviously, confidence about the usefulness of monetary policy implies a position on the lower half of the table.

From Lance Taylor, Reconstructing Macroeconomics

It would be foolish to debate which of these positions is the correct one — though the monetarist view that the quantity of money plays an important causal role is clearly inapplicable to modern economies. It also seems possible that we may be seeing a shift away from the focus on the price of credit, and specifically the single policy interest rate — a position that is presented in many recent textbooks as the only possible one, even though it has been dominant only since the 1990s. In general what we should be doing is recognizing the diversity of positions and exploring the historical contexts in which one or another comes to dominate.

2. Regardless of which margin it operates on, monetary policy in its modern sense typically targets a level of aggregate output. This means changing how tightly liquidity constraints bind current expenditure. In other words, how easy is it for a unit that wants to increase its spending to acquire money, either by selling additional current output, selling an asset, or issuing a new liability? So regardless of the immediate target of monetary policy, the intermediate target is liquidity. (So what’s the point? The point is liquidity. The point is liquidity. The point is liquidity.) This may seem obvious, but keeping this idea in mind helps, I think, to cut through a lot of confusion. Expansionary policy makes it easier for someone to finance increased spending relative to income. Contractionary policy makes it harder.

3. Orthodox macroeconomics confuses the issue by assuming a world of infinite liquidity, where anyone can spend as much they like in any given period, subject to an intertemporal budget constraint that their spending over the infinite future must equal their income over that same infinite future. This condition — or equivalently the transversality or no-Ponzi condition — is coherent as a property of mathematical model. But  it is meaningless as applied to observable economic behavior. The only way my spending over my whole lifetime can be limited, is if my spending in some particular period is limited. Conversely, if I can spend as much as I want over any finite horizon, then logically I can spend as much as I want over an infinite horizon too. The orthodox solution is literally to just add an assumption saying “No you can’t,” without any explanation for where this limitation comes from. In reality, any financial constraint that rules out any trajectory of lifetime spending in excess of lifetime income will rule out some trajectories in which lifetime spending is less than lifetime income as well.

More concretely, orthodox theory approaches monetary policy through the lens of a consumption loan, in which the interest rate represents not the terms on which increased expenditure today can be financed, but the terms on which expenditure today trades off against expenditure in the future. In reality, consumption loans — while they do exist — are a very small fraction of total debt. The vast majority of private loans are taken to finance assets, which are expected to be income-positive. The models you find in graduate textbooks, in which the interest rate reflects a choice between consumption now and consumption later, have zero connection with real-world interest rates. The vast majority of loans are incurred to acquire an asset whose return will exceed the cost of the loan. So the expectation is that spending in the future will be higher, not lower, as a result of borrowing today. And of course nobody in the policy world believes in consumption loans or the interest rate as an intertemporal price or the intertemporal budget constraint or any of that. (Just compare Bernanke’s article on “The Credit Channel of Monetary Policy Transmission” with Woodford’s Interest and Prices, the most widely used New Keynesian graduate textbook. These are both “mainstream” economists, but there is zero conceptual overlap.) If you are not already stuck in the flybottle of academic economics there is no reason to worry about this stuff. Interest is not the price of consumption today vs. consumption tomorrow, it’s the price of money or of liquidity.

4. The fundamental tradeoff in the financial system is between flexibility and stability. The capacity of the financial system to delink expenditure from income is the whole point of it but also why it contributes to instability. Think of it this way: The same flexibility that allows an entrepreneur to ignore market signals to introduce a new product or process, allow someone to borrow money for a project that will never pay off. In general, it’s not clear until after the fact which is which. Monetary reforms respond to this tension by simultaneously aiming at making the system more rigid and at making it more flexible. This fundamental conflict is often obscured by the focus on specific mechanisms and by fact that same person often wants both. Go back to Hume, who opposed the use of bank-credit for payments and thought a perfect circulation was one in which the quantity of money was just equal to the amount of gold. But who also praised early banks for allowing merchants to “coin their whole wealth.”

You could also think of liquidity as providing a bridge for expenditure over dips in income. This is helpful when the fall is short-term — the existence of liquidity avoids unnecessary fluctuations in spending (and in aggregate income). But it is a problem when the fall is lasting — eventually, expenditure will have to confirm, and putting the adjustment off makes it larger and more disruptive when it comes. This logic is familiar in the business press, applied in particular, in a moralizing way, to public debt. But the problem is more general and doesn’t admit of a general solution. A more flexible credit system smooths over short-term fluctuations but allows more dangerous long-term imbalances to develop. A more rigid system prevents the development of any large imbalances but means you feel every little bump right up your spine.

(EDIT: On Twitter, Steve Randy Waldman points out that the above paragraph sits uncomfortably with my rejection of the idea of consumption loans. I should probably rewrite it.)
5. Politically, the fundamental fact about monetary policy is that it is central planning that cannot speak its name. The term “natural interest rate” was introduced by Wicksell, introduced to the English-speaking world by Hayek, and reintroduced by Friedman to refer specifically to the interest rate set by the central bank. It becomes necessary to assert that the interest rate is natural only once it is visibly a political question. And this isn’t only about the rhetoric of economics: Practical monetary policy continues to be constrained by the need for the outcome of policy choices to be disguised in this way.

Mike Konczal has a good discussion of how this need to maintain the appearance of “natural”market outcomes has hamstringed policy since 2008.

Starting in December 2012, the Federal Reserve started buying $45 billion a month of long-term Treasuries. Part of the reason was to push down the interest rates on those Treasuries and boost the economy. But what if the Fed … had picked a price for long-term securities, and then figured out how much it would have to buy to get there? Then it would have said, “we aim to set the 10-year Treasury rate at 1.5 percent for the rest of the year” instead of “we will buy $45 billion a month of long-term Treasuries.” This is what the Fed does with short-term interest rates… 

What difference would this have made? The first is that it would be far easier to understand what the Federal Reserve was trying to do over time. … The second is that it might have been easier. … the markets are unlikely to go against the Fed … the third is that if low interest rates are the new normal, through secular stagnation or otherwise, these tools will need to be formalized. … 

The normal economic argument against this is that all the action can be done with the short-rate. … the real argument is political. … the Federal Reserve would be accused of planning the economy by setting long-term interest rates. So it essentially has to sneak around this argument by adjusting quantities. … As Greta R. Krippner notes in her excellent Capitalizing on Crisis, in 1982 Frank Morris of the Boston Fed argued against ending their disaster tour with monetarism by saying, “I think it would be a big mistake to acknowledge that we were willing to peg interest rates again. The presence of an [M1] target has sheltered the central bank from a direct sense of responsibility for interest rates.” 

I agree with Mike: The failure of the Fed to announce a price target for long bonds is a clear sign of the political limits to monetary policy. (Keynes, incidentally, came to support fiscal policy only after observing the same constraints on the Bank of England in the 1920s.) There is a profound ideological resistance to acknowledging that monetary policy is a form of planning. For a vivid example of this ideology in the wild, just go to the FRED website and look up the Federal Funds rate. Deciding on the level of the Fed Funds rate is the primary responsibility of the Federal Reserve, it’s the job of Janet Yellen and the rest of the FOMC. But according to the official documentation, this rate is “essentially determined by the market” and merely “influenced by the Federal Reserve.” There is a profound resistance, inscribed right in the data, to the idea that interest rates are consciously chosen consciously rather than somehow determined naturally in the market.

[*] This is a better description of the evolution of monetary theory than the evolution of monetary policy. It might be more accurate to say that policy went directly from targeting the quantity of credit to the price of credit, with the transitional period of attention to monetary aggregates just window dressing.

14 thoughts on “Five Thoughts on Monetary Policy”

  1. Two toughts:
    On point 3, it seems to me that the orthodox view is not only that all loans are consumption loans (and so the interest rate is the tradeoff between consumption today and consumption tomorrow), but also that all capital investiments (including lending) are a tradeoff between consumption today or consumption tomorrow.
    If one assumes this, then there is a finite optimal level of capital intensity and of leverage to which the market does/should converge.
    However, if one doesn't assume that interest and profits represent a tradeoff between consumption at differnt times, but takes the different approach that profits and interest are just what investors are able to accumulate, there is neither a finite optimal level of leverage nor of capital intensity. (I'm thinking about this stuff because of Piketty and Marx's "tendential fall in the rate of profit", that in reality just means that there isn't a finite level of capital intensity to which the market tends, but capital just is continuously accumulated).

    On point 4, I think that there is a problem when we apply the idea of lending as a bridge over dips of income to the economy as a whole, because when one person can certainly consume more than s/he produces, the whole economy cannot, as it isn't possible to consume something that hasn't been produced yet.
    Thus, while a fall in income for John can cause an increase in John's debt level, a generalized fall in income for the whole economy (for example, because of peak oil) cannot logically cause an increase in the overall leveragedness of the whole economy (at least not in the same straightforward way).
    For example, suppose that Germans sell cars to the Italians for 100, while Italians sell olive oil to Germans for 50, and this imbalance causes a debt of 50 to the Italians. But all cars and oil were already produced, so there isn't any global overconsumption, and the only way to balance trade is either for car's price to fall to 50 or to olive oil price to rise to 100, because the Italians can't produce cars (or they would already be doing it) and the Germans have no reason to produce oil, that has a lower added value, neither to consume themselves all the cars (or they would already be doing it instead of lending to the Italians).

    I personally believe that there is a sort of "permanent saving glut" in capitalist economies due to the fact that some actors tend to accumulate wealth and not to consume, so an increase in leveragedness is needed to match aggregate demand to aggregate supply.
    When this increase in leverage for some reason stops, we have financial crises and underconsumption crises.

    1. On point 3, everything you say seems right to me.

      Also right on point 4 — Steve Randy Waldman on twitter also pointed out that it is a bad metaphor.

      The one point I would push back on is a secular shortfall in consumption demand. First of all, it is logically possible for an economy to maintain an arbitrarily high investment share indefinitely. China today is a good example of an economy where the main source of demand validating past investment, is new investment. Second, I'm not convinced that there is any evidence of a decline in consumption demand in the US or most other rich countries. It is true that, all else equal, upward redistribution will tend to reduce desired consumption. But that can be, and evidently has been, offset by increased luxury demand by the rich and by increased third-party consumption in the form of public health benefits, etc. It seems to me that to the extent there is a chronic shortfall of demand, a better candidate is reduced investment due to stronger liquidity preference among capitalists.

  2. A few small points.

    It looks like you chopped off the end of a sentance in point 4. ".. Same flexibility that allows an entrepreneur Monetary reform"

    The Fed has shied away from the implication about their setting bond yields for some time. I agree that it is politics, but I would not think it is squeamishness about central planning. The Fed does not protests from widows and orphans every time bond holders suffer capital losses. Additionally, they targeted bond yields in the 1950s, and the objective was financial repression. Targeting yields looks like they are subordinate to fiscal policy.

    Finally, the _effective_ fed funds rate is a market rate. There used to be a single target for that rate, but now I believe the express it as an upper and lower bound. The effective rate used to have a reasonable spread versus the target, at least until he mid-2000s.

    Until some point in the early 1990s, the FOMC did not announce their target, and market participants had to guess what it was. (That was before my time in finance.) There was concern about making the target rate public, as it was unclear that Fed could hit it with open market operations.

    Other central banks used a corridor system with interest rates on deposits or loans from the central bank. These banks had less worries about guiding interbank rates.

    1. I understood that the Fed has moved a corridor system as well with interest on excess reserves (IOER).

      I'm not sure what your point is about central planning and widow and orphans.

      I would look to Kalecki. If supposedly the market sets long term interest rates, then it's Tom Friedman's "golden straightjacket" enforcing market discipline on governments that spend too much.

      But if the Fed can set it where it wants, then the government can spend what it wants up to a point. There is no market discipline.

    2. Brian-

      Thanks, fixed that sentence.

      On the central planning point, I don't think we disagree. Another way of saying the same thing would be that the Fed is reluctant to announce prices because that makes monetary policy more obviously a political choice. That's the point of the quote from Governor Morris.

      You are right there is a (tiny) range for the Fed Funds target. But I think there is a reason that that trivial technical point gets such prominent emphasis in the official definition.

      You are right also that the official target was not announced prior to the late 1980s. Krippner's book (which is excellent, by the way) gives a lot of evidence from Fed policymakers at the time that the main reason for this was to avoid the perception that the Fed was "setting prices" in an important market. Concerns about the ability to hit the target are an argument FOR announcing it, not against, since it is easier to hit a public target.

  3. Josh – can you reference research you find interesting on the liquidity point, liquidity point, liquidity point?

    What does it mean, exactly, to abstract from liquidity, or assume its non-scarcity?

    I'm sure you have in mind Perry Mehrling/the money view for this post. Are there more concerted efforts to describe what liquidity is and what it does with from this perspective?

    And what about our most recent nobel prize winner's work? This from the abstract of a 1998 paper of his ("Private and Public Supply of Liquidity" Bengt Holmström and Jean Tirole):

    When there is only aggregate uncertainty, the private sector cannot satisfy its own liquidity needs. The government can improve welfare by issuing bonds that commit future consumer income. Government bonds command a liquidity premium over private claims. The government should manage debt so that liquidity is loosened (the value of bonds is high) when the aggregate liquidity shock is high and is tightened when the liquidity shock is low. The paper thus suggests a rationale both for government‐supplied liquidity and for its active management.


  4. "In reality, consumption loans — while they do exist — are a very small fraction of total debt. The vast majority of private loans are taken to finance assets, which are expected to be income-positive."

    I'm not so sure about this. Home equity loans pre-crisis looked like a situation where a significant chunk of the credit markets was being used to finance current consumption. Of course the collateral was not your future income but the value of your house.

  5. You may be right about this — tho some significant fraction of home equity loans financed home improvements, no? But I think it's a mistake to take bubble-era home equity lending as representative of household credit in general. I had a long debate about this with Steve Randy Waldman on twitter — he had the same criticism of the post as you do.

  6. At the risk of typing to a dead thread I'd like to address #2 and #4. Would it be fair to characterize monetary policy as a tool that can boost aggregate output but at the expense of exacerbating financial imbalances? I'm thinking of Steve Randy Waldman's post mocking the Great Moderation, not the typical "too low, too long" framing that you've criticized in past blog posts.

    Key points as I understand it:
    1. Business investment may not be very sensitive to interest rates, but real estate investment is
    2. Interest rate sensitivity is pro-cyclical. People won't borrow unless they A) can expect to make sales or capital gains B) meet lenders underwriting standards (which deteriorate as memory of the past crisis fades, a la Minsky).
    3. Borrowing does boost GDP through investment and wealth effects, but mostly is used to bid up the price of existing assets (which means that debt tends to increase faster than income)
    4. The burden of debt servicing can be a drag on the economy and/or a contributor to financial fragility. This could be framed in terms of sectoral balances (retained earnings by financial corporations would correspond to a portion of household or business dissaving) or income distribution (even if retained earnings by financial corporations are small, dividends and deposit interest accrue mostly to the wealthy).
    5. Randy Wray talks about the perverse impact of higher interest rates through interest income from govt. bonds, but it is based on the silly assumption that most govt. debt is held by widows and orphans with a high propensity to consume out of interest income as opposed to central banks and pension funds.

Comments are closed.