It is a great honour and pleasure to be here. I, like many in this room, owe Stan Fischer a great debt. I was the last graduate student that he supervised. You could argue that I drove him out of academia into policymaking!

When Stan was teaching his macro course at MIT, in the very first lecture he used to say: you learn [the] ISLM and Mundell-Fleming [models] in your undergraduate course (and most of us learnt it from [Rudiger] Dornbusch and [Stanley] Fischer), then you come to graduate school and we cover that in half a lecture before moving onto teaching more advanced technologies. But then you go out into the real world of policymaking and your first and best intuition still comes from Mundell-Fleming.

So I must say I was struck by Emmanuel Farhi and Ivan Werning's claim upfront in their paper^[1] that, to quote, ‘because our model has explicit microfoundations and a formal treatment of general equilibrium, it is better suited for normative analysis than the traditional Mundell-Fleming models’.

That's a pretty big claim. Call me old-fashioned but I'm not sure that I buy it.

Before I get to that, let's think a bit more about some issues that come with a floating exchange rate. Emmanuel and Ivan talk in their paper about emerging markets, but it's not just an emerging market issue. I will also draw on the Australian experience.

When the exchange rate is moving in line with the terms of trade, the real fundamentals, it plays the appropriate macro stabilisation role. Farhi and Werning have that explicitly in their model. In Australia's case, over the 30 years of the float, the exchange rate has done just that most of the time.

The problem is when financial factors cause the exchange rate to deviate from the real fundamentals.

How can that come about?

Well let's think of a country, which I'll call the Usual Suspect or US for short. That country pursues an extremely expansionary monetary policy, which is completely justifiable from its perspective. This spills over to the rest of the world which experiences an exchange rate appreciation as capital flows to earn a higher yield.

Even if the global portfolio reallocation that comes about from this is small from the US' perspective, it can be large from the point of view of a small open economy that is the recipient of these flows.

In the emerging market world, the concern is often that the capital inflows will become capital outflows. In Australia's case, an exchange rate appreciation that is not in line with the fundamentals, if persistent enough, can lead to Dutch Disease.

This is the fundamental problem, be it a Trilemma, or Dilemma, as Hélène Rey labels it.

But, you might say, isn't a stronger US economy good for the rest of the world? Or in other words, ‘suck it up sunshine’.

That may well be true for the rest of the world as a whole, but it is not necessarily true for every other country. There is an income and a substitution effect at work here. The income effect is the boost from a stronger US in terms of greater global demand. The substitution effect is the negative effect of the exchange rate appreciation. The substitution effect can outweigh the income effect, particularly if you don't trade that much with the US and/or if the capital flows are large relative to the size of your economy leading to a relatively large appreciation of the exchange rate.

That is the fundamental problem that the authors are getting at in their paper. Though, so far, Mundell-Fleming has done just as well in getting to this point as their more advanced technology.

Now let's think about what happens when the capital flow reverses. (I would note that the term ‘sudden stop’ maybe isn't the right term to use, being a hold-over from a fixed exchange rate world.)

In Australia's case, the exchange rate depreciates and the main concern is inflation. This is manageable if the appropriate macro framework is in place, such as an inflation targeting framework. This is one of the points of Fischer's 2001 lecture at the first incarnation of this conference.

In the emerging markets world there are two problems. The first are also inflation concerns, particularly if inflation expectations are less well anchored. The second are balance sheet effects in the form of unhedged foreign exchange denominated liabilities. This is the genesis of the fear of floating and was critical in the Asian crisis. Paul Krugman's second generation models capture this well.

Does the model in the paper here capture this? Mostly. It captures the issue of the exchange rate not moving in line with the fundamentals. It doesn't capture this balance sheet channel though, where the exchange rate depreciation hurts so much.

Why doesn't the standard macro response work? That is, when the exchange rate appreciates, lower interest rates to ameliorate its impact. Mundell-Fleming would say to do that. It would also say use fiscal policy, which is not discussed in the paper.

Extreme reductions in interest rates can cause problems of excessive price rises in domestic asset markets. So could one turn to ‘macro-prudential’ measures? These are effectively domestic capital controls on the locals, rather than the capital controls on the foreigners that the paper proposes.

Basically one is trying to make yourself less attractive to foreign capital inflow. That's the fundamental issue here in this paper.

That has posed an interesting conundrum in Australia in recent years. We have been experiencing a ‘boom with gloom’. We have had the difficult balancing act of trying to tell foreigners that the economy is not as good as they think it is, so stop sending us so much capital, while at the same time trying to convince the locals that the economy is not as bad as they think it is. Now that's a real dilemma!

In the end I see the model in this paper as proposing capital controls to address the distortion of the exchange rate moving out of line with the fundamentals. Basically it boils down to the Tinbergen principle. With only monetary policy, there is one instrument and two objectives. Give me another policy instrument and it is easier to hit the two objectives.

But other potential instruments are not considered in the paper. For example, intervention in the foreign exchange market in the form of reserve accumulation, which has been used in many emerging markets. This is helpful on the way in, in terms of dampening the inflows, but it also helpful in dealing with the outflows when they come.

Moreover, the paper is silent on how to go about applying the capital controls in practice. Indeed, it is ironic to be discussing the appropriate implementation of capital controls at the IMF, which for a long time has regarded them as a complete anathema. Capital flows are a repeated game, an issue I know the authors have addressed in another paper. If you are a capital importing country, you need capital inflows to sustain growth in the medium term.

In the end, the paper does not convince me of their claim to abandon that long-time stalwart, the Mundell-Fleming model. The intuition of Mundell-Fleming gets me to the same point as this more fancy technology. So I'm not sure I'll invest in the technology upgrade. So the first thing that Fischer said in his macro lecture stands the test of time. In the real world, in the end you always come back to ISLM and Mundell-Fleming.

Endnote

Farhi E and I Werning (2013), ‘Dilemma not Trilemma? Capital Controls and Exchange Rates with Volatile Capital Flows’, Paper presented at the 14th Jacques Polak Annual Research Conference, Washington, DC, 7–8 November. [1]