Recent Posts

Wednesday, September 5, 2018

Japan And The Costs Of Bond Yield Control

Chart: 10-year JGB Yield

The dangers of distorting free market interest rates is one of the bits of market folklore that keeps getting passed around. There is actually not a whole lot of data to defend this view; it is best viewed as faith-based reasoning. This topic is particularly interesting in the case of Japan. I am somewhat agnostic on this issue; I do not see particular risks from manipulating the yield curve in the current environment, yet I can see some plausible dangers.

This article was triggered by the article "Bank of Japan once again shows who calls the shots," by Bill Mitchell, one of the leading Modern Monetary Theory (MMT) economists. In addition, I had a discussion about this topic with someone doing some research awhile ago. Rather than re-hash Professor Mitchell's points from the MMT perspective, I will put on my "generic market analyst" hat and give a description of the issue from a more theory-agnostic perspective. This article probably covers topics I have already covered, but I am still recovering from the Banjo Bowl disaster on the weekend (plus I am now doing more home renovations).

Mainstream Macro Theory

I do not want to get involved in the great mainstream-heterodox mud-slinging match right now. However, it would be crazy to ignore the distinction in views when discussing this topic. The standard view is that interest rates are critical for determining economic outcomes, and so any manipulation of the yield curve is extremely important.

Standard mainstream macro -- and even offshoots, like Austrian theory -- assume that interest rates are critical for all economic decisions. The reason being is that everyone buys all products in spot and forward markets extending over all time horizons. Interest rates are assumed to be important for the relative prices between spot and forward.

For fixed income markets, interest rates and forward purchases obviously matter. Furthermore, there are some flexprice commodities (oil, grains) that are bought/sold forward. However, these markets are a small subset of all market transactions in the economy. So it is very much unclear how applicable the assumption that interest rates are paramount really is.

From what I have seen of the mainstream empirical literature (which is admittedly a small portion), the analysis bakes the assumption that interest rates matter into the cake; there is literally no way of falsifying the thesis that interest rates matter or not. From my perspective, the interesting observation is that I cannot think of any empirical observations that confirms the conventional view of the effectiveness of interest rates, beyond the Volcker episode. Being able to point to one data point -- at a period of history when there was a lot of economic policy shifts -- is not the most impressive defence of a theory.

I am not going to resolve that debate herein. All I can say is that if one is not willing to assume that small changes to interest rates are of critical economic importance, we need to dig further into what the costs to yield curve manipulation are (which is what will be discussed in the remainder of this article).

Costs to Yield Curve Manipulation

The Mitchell post explains why the government can set the entire yield curve. I will instead focus on the potential risks to such a posture.
  • Losing the ability to influence the economy via setting interest rates.
  • The political cost of changing bond prices.
  • The loss of market information.
  • Political cost of interest expense.
I will cover these in turn.

Influencing the Economy

From a real world political perspective, the dominance of conventional thinking about the effects of interest rates on the economy makes this the primary practical concern. However, since I am skeptical about the ability of the central bank to control the economy with interest rates, I will dig further into the topic from this vantage point.

Even if "control" of the economy with interest rates might be far less feasible than mainstream thinking suggests, it seems reasonable to argue that interest rates can be useful to influence it under certain circumstances. For example, higher interest rates will eventually curtail real estate speculation. Admittedly, this is not a big worry in Japan right now.

An alternative use of high interest rates is of more interest: attempting to defend the value of a currency in foreign exchange markets. I am unconvinced that a high policy rate will necessarily help defend the value of the currency, but I am almost certainly in the minority with that view. If enough market participants believe that high interest rates boost a currency's value, that is what we should expect to happen. (This is perhaps a better topic for anthropology than economics.)

For Japan, defending the yen in a reasonable concern (if we put aside Japan's rather sizeable foreign exchange reserves). Japan is an island nation in a rather awkward geopolitical neighbourhood, and is dependent upon various imported raw materials. The Japanese government can certainly always buy domestically produced goods and services with yen (as per MMT arguments), but import requirements need to take into account the external value of the currency.*

Another practical problem is the design of pension systems. Pension systems were designed on the assumption that it would be possible to eventually meet actuarial cash flows with assets with a positive real rate of return. Locking bond yields at a negative real rate represents a serious incoherence in policy design. This is a problem for Japan, as well as elsewhere.

If everyone switched over to a MMT-ish world view, concerns about the loss of interest rate control might disappear. However, this has not happened yet.

The remaining sections of this article are based on the assumption that interest rates can change in the future; if they are locked at 0% permanently, they are moot.

Political Cost of Changing Bond Prices

If the central bank changes the policy rate, it is changing the pricing of an overnight instrument -- with a duration remarkably close to zero. If it changes its target for bond yields, it is handing capital gains/losses on long duration instruments.

This will cause annoyance among bond holders, and creates a huge potential for shenanigans. We live in a world where there is a revolving door between governmental posts and the private sector, and I am unsure whether there is a widespread belief that this revolving door should be closed. (I am a prairie populist, and not a fan of these revolving door arrangements. However, I recognise that my view is in the minority.) Even if there are no shenanigans, the suspicion that they exist will always be there. If I were a central banker, I certainly would not want a system that induces people to assume that central bankers are corrupt.

Loss of Market Information

Anyone from the Chicago School, or of an Austrian bent, will be quite adamant about the importance of market information. By pegging bond yields, policymakers have destroyed the information content of the yield curve, which is one of the most reliable recession indicators (in the United States; ZIRP destroyed the information content of the JGB curve).

I am highly skeptical with regards to the importance of the loss of market information for the private sector. In the real world, people do not plan all consumption decisions in hypothetical forward markets that use the risk-free rate as a discounting instrument. Meanwhile, changing the risk-free rate will not greatly influence lending decisions: the private sector lends on a spread basis. Government interference in private lending only matters if they are distorting credit decisions -- exactly like the CMHC in Canada. Otherwise, we are just back to the debate around the importance of the level of the risk-free curve for the economy, as discussed earlier.

Where the loss of information might matter is for policymakers. If we believe that policymakers can fine tune the economy with interest rates, the signal provided by the yield curve presumably gives them some information from private sector market participants. The exact value of the information can be debated. If one believes that the term premium tears around in a random, unpredictable fashion, the curve is not going to contain a lot of information. Furthermore, the market consensus can be quite wrong, particularly in the early parts of expansions.

Policymakers could turn to other markets for information, such as the inflation-linked market. (My book on which is supposed to be nearly done...) All they need to do is avoid destroying the information available in the market by passing it through an affine term structure model, or by pinning those prices as well. (In the same fashion that Market Monetarists insist that central bank can set the level of nominal GDP by buying or selling hypothetical GDP futures, one could imagine the New Keynesian brain trust arguing that inflation rates can be set by pegging inflation breakeven rates.)

Political Costs of Interest Expense

One of the side effects of central bank yield curve control is that it cannot force people to buy long maturity bonds at the price it sets. Eventually, it will run into a situation where it owns most of the long end of the yield curve.

If we consolidate the central bank with the Treasury (which we should), this creates a situation that is economically equivalent to replacing long maturity debt with short maturity paper. As a result, changes to interest rates will immediately change fiscal interest expense, whereas a long-duration debt structure will insulate overall interest expense for years.

Financial commentators, mainstream economists, and politicians have a well-recorded tendency to scream about rising interest costs. Since there is no chance of a competent currency sovereign going bankrupt, these concerns are meaningless. However, in the real world, we have to deal with people who indulge in magical thinking (as I believe most anthropologists would attest).

Although it would be great if everyone were willing to think sensibly about government finance. I am not going to hold my breath waiting for that to happen. From my perspective (which I believe would not be shared by Bill Mitchell and other MMTers), I would not want to waste political capital walking into an obvious trap.

Returning to Japan, I see very little risk of rising inflation on any reasonable horizon, outside the possibility of some form of energy price shock. However, I see the risk of the Bank of Japan robotically raising rates in response to some inflationary shock. This will create a feedback loop to fiscal policy, with rising interest spending creating the spending power to sustain higher prices. (This was a common view in the inflationary 1970s, which disappeared after the consensus decided that high interest rates suppressed inflation.) A short maturity debt structure allows for such a feedback loop; if interest costs are largely fixed, this effect cannot kick into gear. In other words, a long maturity debt structure helps prevent the economy from being blown up by pro-cyclical policies by mainstream central bankers (again).

The relatively high debt-to-GDP ratio of Japan makes this more than a theoretical corner case. It would not be hard to craft a "Japan is doomed!" narrative based on post-Keynesian theory, and assuming that Japanese policymakers follow the mainstream script to the letter.

Concluding Remarks

Austrian economics is far more influential in market commentary than it is academia. As a result, we should expect complaints about the dangers posed by distorting the yield curve to continue. That said, there is not a great deal of evidence that it actually matters.


* Some Post-Keynesians go on about the "external constraint" when discussing MMT. For floating currency sovereigns, the "external constraint" is really just the external value of the currency. A falling currency value should probably be lumped in with "inflation," which is already discussed in Functional Finance. In other words, people who are concerned about "the external constraint" are more worried about semantics than the operational effects of policies.

(c) Brian Romanchuk 2018


  1. The basic reason for disapproval of "distorting free market interest rates" is surely the one set out in introductory economics text books, namely that GDP is maximised where prices (including the price of borrowed money) are at free market rates. GDP is maximised where the price of apples, steel and everything else are at free market prices, unless there is a clear social reason for taxing or subsidising something.

    1. Well, that would require using introductory econ textbooks as anything other than punchlines to jokes...

      There is no “free market rate” for the risk free short rate in a free-floating fiat currency. (In a currency peg system, your domestic rate will typically be at a markup over the policy rate of thd senior central bank in the system.) It is always set as an administrative rate. The only question is determining the rules for setting that rate.

    2. The fact that central banks the world over try to "administer" a rate does not mean they can't abstain from that activity: sitting around doing nothing isn't all that difficult, is it? To illustrate with a simple economy which uses say cowrie shells as money (with the number of cowrie shells issued being enough to induce the population to spend at a rate that brings full employment), why would some sort of genuine free market rate of interest not establish itself, assuming the central bank or committee which issued cowrie shells made no attempt to influence interest rates?

    3. The cowrie snails have a committee that decides how many shells are issued?

    4. In order for a banking system to function properly, it needs a lender-of-last-resort. The discount rate on those operations is an administrative decision. Private interest rates end up as being a markup over that interest rate.

      I understand that you have an ideological distaste for lender-of-last-resort operations, but that is how the real world functions.

    5. Jerry, Yes: clearly there has to be some sort of committee or central bank that issues enough cowrie shells to induce the population to spend at a rate that brings full employment, without excess inflation.

      Brian, A bank system where there is no lender of last resort is perfectly feasible. Indeed, the existing system does not offer “loans of last resort” where the central bank thinks a commercial bank is in a sufficiently bad state: e.g. Lehmans and Northern Rock.

      As for a system where there are no last resort loans under any circumstances, that’s easily arranged. All one needs to do is raise bank capital ratios by enough to ensure that bank failures are near impossible. Martin Wolf and Anat Admati claim that 25% would do the job. As for any banks which fail, despite having a 25% capital ratio, they can be given the “Lehman / Northern Rock” treatment.

  2. I didn't see any exploration of the possible role low-interest-rates played in the 40 year income shift illustrated here:

    Having lived through that period, I noticed the steady decline of interest rates over the period. That decline has been accompanied by a progressive shift away from individual savings into institutional investment (such as through pension funds). The government-big business nexus has tightened at a cost to smaller entities. Maybe we could apply a label of socialized capitalism.

    CB controlled interest rates would theoretically have a role here in suppressing the supply of financial capital by individuals.

    I think I would be driving toward a political-economic theory here.


Note: Posts may be moderated, and there may be a considerable delay before they appear.

Although I welcome people who disagree with me, please be civil.

Please note that my spam comment filter appears to dislike long "anonymous" posts. I get no warning about this, and only go through my "spambox" infrequently. The best bet it to keep comments short, and if you think the spam filter struck, let me know with a short comment.