Recent Posts

Tuesday, February 18, 2014

MMT And Constraints

The economist Thomas Palley launched a broadside against Modern Monetary Theory (MMT), raising a variety of comments across the blogosphere. JW Mason covered a lot of objections that I had to the article in this response. However, I will comment on a few points, mainly related to the government budget constraint, and external constraints.

I am not a historian of economic thought, and so I cannot really comment on the extent of the novelty of MMT. From my readings of MMT authors (Wray, Mosler, Mitchell in particular), they make pains to acknowledge the roots of MMT. As an ex-academic, I understand the concern, but I do not see any particular problem from what I have read. And having been in involved with applied mathematics, I developed the sensibility of a mathematician with regards to "novelty". Within mathematics, re-deriving an existing result in a more elegant fashion is considered to be a good thing.

Thomas Palleys's complaint about the zero rate policy being unsustainable is empirically wrong, as JW Mason notes. Japan has run a zero rate policy for almost two decades without any problems of "sustainability". The yen did not go to zero; if anything, it is still too strong. Portfolio rebalancing effects and trade flows are more important than carry. As for speculation created by zero rates, there would be at most a one-time upward shock to asset prices as the risk-free curve converges to zero, but the prices will still be bounded by the risk premium.  (I will point out however, that I think there are negative side effects associated with a zero rate policy. In particular, it increases the need to hoard financial assets to fund retirement.)

I will cover two other issues he raised, both of which are related to "constraints" on policy.

Governmental Budget Constraint

Palley's argued that the MMT formulation of the governmental budget constraint is a special case of the already known general budget constraint for a single time period (e.g., the transition from time t to t+1).

Well, that is exactly what the MMT authors say: they accept the one period transition "constraint", as it is just an accounting relationship. What is rejected by MMT is the condition based on looking at time as it goes to infinity - which gives rise to Ricardian Equivalence. Palley completely ignored that part of the constraint, which is the interesting part of the MMT discussion of the subject. And the MMT authors are undoubtedly correct - I demonstrate here with a counter-example that Ricardian Equivalence does not hold.

External Constraint

Finally, it is also noteworthy that MMT appears more plausible to US audiences than to other country audiences. All countries face inflation and financial sector stability constraints, but the US is essentially free of a foreign exchange market constraint.However, that constraint is very visible in many other countries, which explains their greater intuitive skepticism about MMT.
Yes, the foreign exchange constraint is visible in a lot of countries. However, those countries are mis-managed, or else they are very small countries have that inherently difficult economic policy decisions. MMT underlines the freedom of action that is created by having a free-floating exchange rate. But it is necessary that countries guard this freedom by doing things like stopping locals from borrowing in a foreign currency (which is what competent developed country regulators already do).

There appear to be three potential mechanisms for a "constraint" on policy to appear from foreign exchange considerations:
  1. inflation passthrough from a weak currency;
  2. government debt holdings by foreigners;
  3. trade balance considerations.
On the inflation side, it does not appear to be that big a deal. If you are a small country like Iceland, you have to deal with inflation passthrough from exchange rate moves. But this is likely to only be a one-off shift, as a currency cannot fall rapidly forever. For a larger country like Canada, passthrough is a complete non-issue.

Can the fact that foreigners hold their debt really constrain governments? Not really. Foreigners can dump domestic bonds and drive down the value of the currency, but all that would accomplishes is that they have destroyed a lot of their own capital. Other investors will be able to snap up a lot of assets that were driven below fair value. In particular, the central bank could engage in some massive open market operations and profit at the expense of the foreigners.

As for trade flows, a country could have its currency driven down so that it moves from a persistent current account deficit to a persistent current account surplus position. This means that the economy would tend to grow faster than it would otherwise, and so if no changes were made to policy, it would be inflationary. But does that really matter? Any structural change to the economy will force a change in policy settings. Policymakers need only react when the changes occur, and so there is no constraint on current behaviour. And even if the change is made, it is only made because domestic conditions have changed (domestic inflation has risen), not because of what some foreigners think.

(c) Brian Romanchuk 2014


  1. It’s not just MMTers who have advocated a permanent zero interest rate. Milton Friedman advocated the same. See paragraph starting “Under the proposal…” (p.250) here:

    That is, he advocated having the state issue currency, but no debt. Plus he opposed open market operations by the Fed. I.e. he is saying “no interest rate adjustments” far as I can see.

    Moreover, there is a fundamentally illogical aspect to interest rate adjustments, namely that they involve imparting stimulus (or its opposite) via just one form of economic activity: borrowing, lending and investment. You might as well adjust demand by boosting just the auto industry and restaurants.

    1. Right - I forgot about the Friedman proposal. Good catch.

      I am agnostic as to how effective interest rates are for stimulating the economy. But the attraction is obvious, when we consider that central banks are bureaucrats who do not want to get involved in politics. (If the winds change, they get replaced by the other party.) Operations in the government bond market creates negligible direct income transfers, so they appear neutral. Purchases of goods and services are a fiscal operation, and are inherently political.

  2. Needless to say, I agree with most of what you write here. However, i think you are a bit too quick to dismiss the external constraint. It is far from clear that the Marshall-Lerner condition is satisfied in general; if it's not, a depreciation will move the current account balance further toward deficit rather than toward surplus. Even if it is satisfied in the long run, the disruptions associated with adjustment can be substantial and long-lived, as many economic plans are made in the expectation of certain exchange rates and are slow to adjust. In particular, export industries tend to be among the most important users of imported inputs and capital goods. Finally, even if a depreciation would be macroeconomically desirable, there is no assurance that foreign exchange markets will lead to the "correct" change. if governments feel they have to actively manage the exchange rate, this really does constrain their ability to set the interest rate on the basis of domestic considerations. This was Keynes' number one concern about post WWII economic arrangements, and I think it would be a mistake to dismiss it.

    Note however that this is a problem for orthodox views on macroeconomic management just as much as for MMT. I think it is better to take this constraint seriously, but to be very clear that it is different from a financial constraint, which governments of countries with their own currencies do not generally face.

    1. Thanks for the response.

      I will immediately put a disclaimer that I am coming into economics as an applied mathematician with a background of working in fixed income. I note that relatively often, and I hate to announce the limitations of my knowledge in every single post. But I mainly look at what I refer to as "developed markets", which is a bond market practitioner definition, and possibly a much more limited set of rich countries that economists might consider "developed". This means that my assertions are not applicable to a lot of countries.

      Part of the reason why my response was short is that Palley wrote very little to respond to. He just asserted that a foreign exchange constraint exists, without any justification. I have no idea what this constraint is supposed to be. Maybe this is well known in the literature, but I am still coming to grips with the horror show that is closed economy macro theory. (An entire school of thought based on mathematical models that they are unable to solve?)

      However, my experience was that for the post-1990 experience for the select handful of countries (e.g., Canada, U.S., Japan, Australia) , forex moves have almost no explanatory power for the domestic economy. We have had highly correlated business cycles, and things like inflation outcomes have been very similar. This is despite the fact that currency levels went up and down like yo-yo's. The Bank of Canada used to follow the "Monetary Conditions Index", which is a combination of the policy rate and currency moves. They were forced to abandon it in the 1990's as they discovered it had no explanatory power. Given the indifference of the domestic economy to forex that I have observed, I see no reason that policy making is constrained.

      Although I believe that the level of the currency should be ignored, I do not think it will necessarily be at the "correct" level. However, rich countries in which domestics have large balance sheets full of financial assets, and which do not face capital controls, will eventually lean against stupid currency levels. (I have stated this often, so I did not feel like repeating.) I believe that this mechanism is the "secret sauce" why a country like Canada has been able to ride out forex volatility. There may be turmoil in the real economy, but the turmoil is a lot less than following ill-advised domestic policies on the theory that the "foreign exchange vigilantes" will somehow punish you.

    2. Coming at this with a very UK-orientated perspective, I have a lot of time for what JW is saying here.

      UK import and export volumes appear to be relatively price inelastic, so most of the response to significant exchange rate movements is through price effects (see, for example, my own analysis - Import penetration is significant enough for these to have a marked impact on consumer prices. Indeed, much of the UK's persistent inflation in the face of weak wage growth since 2008 can be put down to the depreciation at that time.

      Furthermore, the UK has historically shown a fair amount of real wage stickiness other than at quite high levels of unemployment. This means that unless the economy is operating significantly below capacity, imported inflation can quickly fuel domestic price inflation. This then has the potential to feed further depreciation.

      The harder question is far the UK can go in persuading foreigners to finance its external deficits. It's possible that there is a lot of scope here, but the external constraint is not one the UK can ignore.

    3. I realise I made an error in making the dichotomy between small and large countries; I sent it out before I noted that was probably not the best phrasing. I was aware of the UK case, where there does seem to be more of a forex passthrough. (However, I did note the Iceland case; so I probably should have said there are countries (like the UK) that are somewhat closer to the Iceland case then Canada.)

      But even for the UK, I believe that the government hiked the VAT two years in a row, and this had a very large mechanical effect on consumer price inflation. From data that I saw one time, a lot of the behaviour of UK inflation post-2008 becomes a lot easier to explain. (Most of the UK research I have seen did not strip out the tax impact on inflation.) I have not had time to look at your article, sorry. Once I get around to getting my hands on UK data, I will be able to comment further.

      Even so, there was not effect on monetary policy - the BoE "looked through" the one-time shock, and did not hike rates. Since rates were at the effective lower bound, there was no measurable effect on interest rate policy. It may be that I am taking advantage of a special case, but the fact is that I do not have to create a counter-factual model to "prove" that interest rates would have been lower is sterling had not fallen. I can comfortably assert that interest rates could not have been lowered even if the currency had not weakened.

      If one takes mainstream inflation targeting as given, I think it it is debatable that forex acts as a true behavioural constraint. Interest rates go up and down in response to a wide range of factors, based on a reaction function that is based on domestic variables. If you believe that forex is a big indirect input to that reaction function, I cannot say you are wrong (although the experience of the Bank of Canada indicates that the passthrough is a lot weaker than was conventionally supposed). But my point is that the forex market cannot force the central bank to change its reaction function, which is what I would consider to be a behavioural constraint.

      A country with a free-floating currency has no sensible policy mechanism to strengthen the external value of its currency. This is unavoidable, and it creates the potential for a big one-off inflation shock. And for a country like Iceland, this is a big dilemma. However, as I argued above, for the larger countries, this does not seem like a big enough issue to adapt current policy settings to this potential problem.

      Even if foreigners decide to collude against buying sterling in the forex market, U.K. equities will become extremely attractive, and so at some point, the equity flows would crush the people who are attempting to boycott sterling. So, there is no reason to expect that the currency can fall forever. Since this crisis is most likely correlated to weak growth, domestic weakness will probably counteract the surge in imported prices.

      Although imported input costs for exporters go up, by definition, those are world prices. Domestic costs fall, and so the competitive position of exporters improves, as long as they use the same mix of imported goods as foreign competitors. The trade account may move against the country in the short term (as Japan is seeing), but forward-looking profitability improves. If I am buying some equities, I am supposed to be buying that forward stream of profits, not the current quarter's.

      If you are arguing that the external constraint can force a tightening of fiscal policy as fiscal policy is too loose and the import channel is causing inflation, I guess I have to conditionally agree. However, I would note that the problem is domestic inflation; rising import prices would be just one channel of rising inflation. As a result, I would lump this in with the general "too loose fiscal policy leads to inflation" behavioural constraint that is part of Functional Finance.

    4. If we are assuming that the central bank is sticking to a reaction function based on inflation targeting, doesn't this conflict slightly with MMT ideas about its relationship with the fiscal authority? If, for example, the central bank's actions involve restricting growth of the monetary base (even if this is achieved indirectly through interest rate setting), doesn't the fiscal authority find itself in a position similar to governments which do not have a sovereign currency? Unless, of course, we allow for the possibility of the central bank abandoning its policy if it puts too much strain on public funding. But then we have the credibility issue.

      BTW, I agree that the VAT hike was also an important factor in UK inflation in this period.

    5. If we take an initial assumption of consolidation, government finance can be thought of as a two-step process: (1) the government spends whatever it wants, increasing the monetary base, and then (2) draining excess reserves to keep the overall yield curve at a desired level. We can readily see that there is no "behavioural" constraint on government finance (deficits can always be financed and default is impossible), even though the "mathematical/accounting constraint" Palley discusses does hold.

      OK, consolidation is not a perfect legal description of real world issuance. (Although in corporate finance, people use consolidated accounts when the justification consolidation is a lot shakier.) The issue is whether those legal constraints can show up in an economic model. For all I know, debt managers need to wear funny hats and give a secret handshake in order to issue bonds. But does this legal constraint have an observable effect on an economic model?

      If we assume that we cannot consolidate, then one could argue that there is little legal reason to distinguish the Government of Canada (or the UK Treasury...) versus the Province of Ontario, or Greece (!). However, there is a big practical difference - the central bank and private banks treat short-dated bills as an extension of the interbank market. As long as your 3-month TBill rate is not allowed to deviate "significantly" from the target rate, it is always possible to finance yourself. (Warren Mosler often makes this point.) There are very considerable regulatory reasons for this to occur, as well as "moral suasion" by regulators (who would stop getting paid if the government went bankrupt) - banks only exist courtesy of a banking license, which can be removed if the bank is acting in an "improper" manner (hint, hint). And in practice, since expectations that the government will not default are self-reinforcing, relative value investors would snarf up TBills if they got too cheap relative to the target rate.

      This obviously does not apply to Greece, which is why they blew up, and the UK didn't.

      There is an added wrinkle. Debt Management offices make their jobs difficult by running fixed quantity auctions. Those auctions could fail. For example, if the UK decided to do all of its issuance by issuing 50-year gilts, I doubt that enough buyers would exist. Well, that is an extreme example. But you could have more subtle examples, and yes, the government may be forced to drop the duration of issuance if there are not enough buyers. This is an inherent weakness of their quantity-based issuance methodologies; a monopoly issuer can set quantity or price, but not both.

      I think that "What Is A Government Bond?" article, even though it is supposed to be a simple primer, covered some of these subtle issues.

    6. You're right.

      I'd been thinking that it was an MMT principal that, regardless of the deficit, the government can still manage down the interest rate, but of course that's not necessary in order to be able to always pay for stuff.

  3. JW,

    The Marshall-Lerner condition is a bit misleading. It just says it is a necessary condition under some assumption for xyz. It is perfectly possible for a nation's trade deficit to widen in historic time even if the currency is depreciating, whether or not ML holds. This is because income elasticity is important.

    Example: Currency starts at 100 and depreciates to 110 over the year. There is deflation in the rest of the world during the year and/or domestic producers are out-competed by foreign firms because of non-price reasons. Exports reduce and widen the trade deficit.

    1. Ramanan,

      You are right, of course. In fact there is quite a lot of material on my blog about income effects, and I am generally of the view that they are usually more important in explaining real world trade flows than price effects.

      But, income effects only operate to constrain macroeconomic policy if (1) they cannot be offset by appropriate change rate changes AND (2) there is a limit to a country's ability to finance current account deficits. So I think the M-L condition is relevant, since if it is not met then condition (1) will not generally hold.


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.