Monetary policy, interest rates and exchange rates

Nick Rowe has a nice new post that is loosely related to the “never reason from a price change” theme.  Nick suggests that it doesn’t make sense to talk about the effect of changes in exchange rates.  Instead one should either talk about the effects of factor “X” that caused the exchange rate to change, or if you are interested in public policy issues you might want to talk about the effect of a central bank policy aimed at preventing the exchange rate from changing.

On a related note, Tyler Cowen links to a new paper by Aart Kraay that discusses the effectiveness of raising interest rates when a currency is under speculative attack:

According to conventional wisdom, currencies that come under speculative attack can be defended with high interest rates. By raising interest rates high enough, the monetary authority can make it prohibitively costly for speculators to take short positions in the currency under attack. High interest rates may also convey a positive signal regarding the commitment of the monetary authority to maintaining a fixed exchange rate. According to the contrarian view, neither of these mechanisms is persuasive. Interest rates have to be increased to very high annualized rates in order to entice investors to hold local currency-denominated assets in the face of a small expected devaluation over a short horizon, and such extremely high interest rates are rarely observed in practice. The signaling value of high interest rates is also unclear. Although signals must be costly in order to be credible, often they impose costs that are too high for the monetary authority to take in stride. Moreover, as the costs of high interest rates mount, the monetary authority’s signal can become less credible over time, raising devaluation expectations. A vicious spiral can result, as expectations of a devaluation force higher interest rates, which in turn impose greater costs on the economy.  In the end, high interest rates can have the perverse effect of increasing the probability that a speculative attack ends in the devaluation of the currency.

I think that’s basically right.  But it’s also an odd way of stating the thesis.  Notice that changes in interest rates are implicitly viewed as a sort of monetary policy stance.  That’s especially dangerous during a speculative crisis, when rapid changes in the expected future exchange rate cause enormous and rapid changes in the Wicksellian equilibrium nominal interest rate.  The real question is not whether high interest rates help; it’s whether or not tight money helps.

Again, I’m not quibbling with the Kraay’s paper, which is sophisticated and seems accurate.  But using the term ‘monetary policy’ rather than ‘interest rates’ would help clarify the analysis:

1.  High interest rates might reflect expectations of devaluation, and hence policy failure.  In that case they will not reflect tight money and will be associated with policy failure, on average.

2.  High interest rates might reflect the liquidity effect from a tight money policy, which makes devaluation less likely.

3.  Even if high interest rates reflect tight money, the policy may fail in the long run of the tight money leads to politically unacceptable output losses, which forces an eventual change in policy. (see Argentina 1998-2002, the USA 1931-33 and Britain 1992.)

Off topic, Matt Yglesias has a nice post on the issue “everyone” is overlooking, monetary policy. Among the “everyone” is Jonathan Chait:

The basic underlying fact of the situation is that the economy is growing very slowly because Congressional Republicans have done everything in their power to apply the fiscal brakes to the recovery. Among macroeconomic forecasters, this is not a remotely controversial assertion but rather an obvious fact that they wearily plug into their models.

File that away in your “obvious facts” folder along with Prescott’s claim that it is a proven scientific fact that monetary shocks have virtually no effect on the business cycle.

HT:  Ramesh Ponnuru



39 Responses to “Monetary policy, interest rates and exchange rates”

  1. Gravatar of Kevin Erdmann Kevin Erdmann
    29. January 2014 at 07:43

    Slightly off topic, but I thought you might be interested in this graph I did of the expected date of the first rate hike over time. Thank heavens for QE3.

  2. Gravatar of benjamin cole benjamin cole
    29. January 2014 at 07:56

    I never savied the defending the currency arguments. Is there a financial group that benefits from this? Or is it more from the same tight-money crowd?

  3. Gravatar of J Mann J Mann
    29. January 2014 at 08:39

    Ian Shepherdson, the economist Chait quotes, does routinely assume that any government spending increase reductions will cut growth.

    I don’t know what he thinks about Sumner’s fiscal offset arguments – every time he shows up in the press, it’s for a pull quote that austerity has hurt or will hurt growth.

  4. Gravatar of J Mann J Mann
    29. January 2014 at 08:42

    If you could divide the forecasters into those who believe in offset and those who don’t, it would be interesting to compare their accuracy over the last few years.

    I also wonder about this quote:

    “I regard the spending cuts of the last several years as just mad,” Ian Shepherdson, chief economist at Pantheon Macroeconomics, told the New York Times, “It delayed what would otherwise have been a quicker recovery.”

    Last SEVERAL years? Is Sheperdson saying we should have stayed at 2009 spending levels indefinitely? That certainly wasn’t how those spending programs were sold at the time.

  5. Gravatar of Karl Smith Karl Smith
    29. January 2014 at 08:43

    Perhaps a better statement of Kraay would be to say: A speculative attack is an expression of the market’s belief that “the central bank will not or cannot effectively tighten monetary policy” Hence, central banks rarely wind up effectively defending against a speculative attack by tightening monetary policy.

  6. Gravatar of John Becker John Becker
    29. January 2014 at 08:51


    I’d be extremely interested to hear you and Nick comment on an imaginary monetary system. What would the economic implications be of the governments of the United States, Canada, and the Eurozone, allowing all financial transactions including paying taxes in dollars, Euros, or Canadian dollars? I like this idea because it would introduce some consumer choice into monetary policy. It would be a way to move away from a government monopoly on money.

    Anyway, I think it is an interesting idea to play with. In this scenario, a central bank could target inflation but I’m not sure how they could target nominal GDP. If you had one central bank targeting a basket of commodities, another targeting 2% inflation, and another targeting 0% inflation, which currency do you think people would pick?

  7. Gravatar of TravisV TravisV
    29. January 2014 at 09:01

    Prof. Sumner,

    Do you guys view it as a good thing or bad thing that India (Raghuram Rajan), Turkey, etc. are tightening monetary policy?

  8. Gravatar of Nick Rowe Nick Rowe
    29. January 2014 at 09:20

    Scott: thanks. Yep. It’s very closely related. I thought of the title “Never reason from an exchange rate change”, but changed it to “Is the fall in the exchange rate good news or bad news” to try to attract more journalists to read it.

    John: Dunno. My guess is that it wouldn’t make much difference. If the government of Canada said Canadians could pay taxes in US dollars, at the spot exchange rate, I think very few Canadians would take advantage of the opportunity. Why bother? Which money we use is one of those network externality things. I don’t accept the chartalist theory of money. (Bitcoin has finally demolished that theory empirically). The government is just one of many players. We don’t always speak the same monetary “language” that the government speaks. We speak the same language that everyone around us speaks. The government normally follows along.

  9. Gravatar of TravisV TravisV
    29. January 2014 at 09:27

    David Tepper (a genius investor) clearly does not understand market monetarism. When U.S. stocks fall, bond prices should INCREASE, not decrease……..

    “Bluntly that hedge did not work this week and particularly on Friday.

    Equity markets had their worst day in yonks. The airlines (Tepper’s biggest position) were not exactly good either (as they were typically off over 4 percent – double the market).

    Bond markets however had a very good day. Long bond indices rose several percent in value.

    In other words Tepper was long the bad stuff, short the good stuff and the short was meant to be his hedge.

    His hedge did not work. Badly. I suspect he was down more than 4 percent on the day – maybe double the equity market.”

  10. Gravatar of TravisV TravisV
    29. January 2014 at 09:57

    Prof. Sumner,

    Why was the (initial) global market reaction to Turkey’s tightening so strongly positive?

  11. Gravatar of John Becker John Becker
    29. January 2014 at 09:59


    Thanks for your input. The Canadian dollar and the U.S. Dollar are so similar right now that the network effects overcome the advantage of using one that is expected to hold more value. Imagine that the government of Canada accepted Bitcoins at the spot exchange rate as well. If one currency is expected to gain value relative to another one, it would make sense to at least pay taxes in that currency. A once a year transaction like paying taxes is easier to plan for than daily purchases.

    The right to pay taxes in different currencies would be a much bigger deal in countries that are suffering from inflation. For instance, the Venezuelans have become masters at currency arbitrages using the dollar against the Bolivar; showing that when inflation is bad enough, the costs and inconveniences of escaping into alternative currencies become unimportant.

    If Venezuelans could buy dollars legally and pay their taxes in dollars, they would quickly stop using Bolivars at all which would bring their inflation problem to a rapid end. Because of their rapid inflation, the convenience of using Bolivars is not important. Under this system, the central bank of Venezuela would have to produce a money that held value almost as well or better than Dollars (they might have some leeway because of network externalities). This would have the added benefit of making it impossible for the central bank of Venezuela to bail out the failed policies of the government.

    My thoughts are that freedom to use competing currencies would curb the worst excesses of the international system of government money provision.

  12. Gravatar of Doug M Doug M
    29. January 2014 at 10:09

    Hong Kong 1997…The whole Asian region was under attack. Hong Kong jacked up their interest rates, Korea, Thailand, Indonesia devalued. Hong Kong successfully maintained their peg.

  13. Gravatar of TravisV TravisV
    29. January 2014 at 11:07

    Prof. Sumner,

    Another question: could China’s aggressive easing-then-tightening of monetary policy explain gold’s dramatic rise (2009 to 2011) then fall (2012 to 2013)?

    Or has gold fallen because we’ve found so much new oil (increasing real interest rates and lowering the attractiveness of gold)?

  14. Gravatar of CA CA
    29. January 2014 at 11:16

    Fed continues its tapering and as of this writing, Dow is down 177. Scott, are we witnessing yet another premature tightening of monetary policy?

  15. Gravatar of ssumner ssumner
    29. January 2014 at 11:28

    Kevin, Thanks. Thank God it leveled off.

    Ben, Better to ignore the exchange rate and focus on NGDP.

    J Mann, Notice he doesn’t call the monetary policy “mad.”

    Karl, Exactly.

    John, I don’t think that change would have any signficant effect. Hardly anyone would take advantage, and even if they did it wouldn’t effect the Fed’s ability to target NGDP.

    TravisV, Assuming they are tightening, I have no opinion on the wisdom of the action, as I haven’t followed those cases closely. I’d look at NGDP forecasts.

    Doug, Good example.

    Travis, Yes, China has a big impact on the gold market.

    CA, It looks that way.

  16. Gravatar of John Becker John Becker
    29. January 2014 at 11:54


    From reading this post, I can’t tell what monetary policy (monetary policy described as a path for nominal GDP) you would recommend for a country that wants to stop the currency from losing value. For a country like Turkey that has a high NGDP (11.8%), why would tightening monetary policy be a bad thing?

    If a country’s currency is under attack, why doesn’t the central bank simply make a credible commitment to stop open market operations or anything else they do that creates new money until the value of their currency rises to some ratio in relation to another currency; sort of like what Switzerland did but in reverse.

    From your perspective, I’d imagine you’d say that Turkey should reduce their expected path of NGDP. Turkey currently has about 7.4% inflation, 4.4% real GDP growth, and 9.7% unemployment compared to a natural rate of around 9% by most estimates: some excess capacity, high inflation, solid growth.* I imagine that if they brought NGDP down form 11.8% per year to 7% or 8% per year, the Lira would strengthen.

    I’ve never heard you talk about how a central bank should tighten yet Scott. Please enlighten us on your approach to that hornet’s nest.

    *Data based on Q4 data from

  17. Gravatar of John Becker John Becker
    29. January 2014 at 12:05


    You and Nick are right to point out that people probably wouldn’t switch currencies if they were given the opportunity to use the EUR, USD, or CAD because the price stability of those currencies is so similar. What type of difference in inflation rates do you think would be necessary for someone using a different currency would want to switch to using the USD for everything if given the option. The Venezuelans, Iranians, and Argentines certainly would like to transact in dollars.

  18. Gravatar of Brian Donohue Brian Donohue
    29. January 2014 at 12:11


    As an empirical matter, stocks and bonds are sometimes positively correlated, sometimes negatively correlated. I don’t think a blanket statement works here.

  19. Gravatar of TravisV TravisV
    29. January 2014 at 12:29

    Brian Donohue,

    Over the next year, if NGDP growth expectations increase to, say, 5.5%, stock prices should increase substantially and the yield on the 10-year treasury should increase substantially.

    Conversely, if NGDP growth expectations decrease to, say, 2.5%, stock prices should decrease substantially and the yield on the 10-year treasury should decrease substantially.

    I’m very confident that that’s what would happen. Are you?

  20. Gravatar of Vivian Darkbloom Vivian Darkbloom
    29. January 2014 at 13:18

    Brian and Travis,

    You may be interested in the following by Warren Buffett:

  21. Gravatar of TravisV TravisV
    29. January 2014 at 13:35

    Vivian Darkbloom,

    You raise an interesting question:

    For equity investors, what would be the optimal NGDP growth rate over the next 20 years? 3.0%? 6.0%?

  22. Gravatar of Vivian Darkbloom Vivian Darkbloom
    29. January 2014 at 13:41

    “For equity investors, what would be the optimal NGDP growth rate over the next 20 years? 3.0%? 6.0%?”

    I might venture a guess if you also tell me what the average inflation rate will be for the same period.

  23. Gravatar of James in London James in London
    29. January 2014 at 14:43

    The last of the Bernanke era FOMC’s and he really seems to have left the place with a policy course set to remain unchanged whatever the weather. $10bn less QE per month every month until it’s over. No wonder the markets are reacting poorly, both bonds and equities.

    The FOMC really seems to be undoing their policy “mistake” from last time when they “mistakenly” overcompensated with dov’ish guidance the start of the tapering.

    I joked with you then that they’d made a mistake, it now appears to be no joke. Great! And some of your fellow MM’ers appear to think it’s all the fault of the PBOC! I don’t think so.

  24. Gravatar of TravisV TravisV
    29. January 2014 at 14:56

    Vivian Darkbloom,

    I think that’s the wrong way to look at it.

    6.0% NGDP growth would result in higher inflation + higher real growth than 3.0% NGDP growth. For picking the optimal NGDP growth target, the relevant question is at what point the costs of higher inflation outweigh the benefits of higher real growth.

  25. Gravatar of Vivian Darkbloom Vivian Darkbloom
    29. January 2014 at 15:24

    “6.0% NGDP growth would result in higher inflation + higher real growth than 3.0% NGDP growth.”

    Well, if that’s the “right way to look at it”, surely you can tell me how much real growth would be under 6 percent NGDP growth and under 3 percent NGDP growth. I’ll then figure out what the inflation rate is on my own.

  26. Gravatar of ssumner ssumner
    29. January 2014 at 17:34

    John, You said:

    “I’ve never heard you talk about how a central bank should tighten yet Scott. Please enlighten us on your approach to that hornet’s nest.”

    Simple, Reduce the monetary base until NGDP growth is on target.

    James, You may be right, but there wasn’t much of a stock market reaction to the news, and indeed the stock market in the US hasn’t changed very much in recent weeks. So I don’t see any big change in policy.

  27. Gravatar of benjamin cole benjamin cole
    29. January 2014 at 17:36

    Vivian Darkbloom: What is important is what monetary policy leads to best long-term real growth. In any macroeconomic policy there are relative winners and losers.
    Free trade has crushed whole industries and even cities—should we be against free trade?

  28. Gravatar of TravisV TravisV
    29. January 2014 at 17:56

    Vivian Darkbloom and Ben Cole,

    There’s also the issue of sticky wages / “musical chairs” / unemployment.

    A 3.0% NGDP growth target might not be enough to overcome sticky wages and ensure full employment. But an NGDP growth target of 4.0% or 4.5% might do the trick.

  29. Gravatar of James in London James in London
    29. January 2014 at 23:00

    Scott. I think you are becoming a bit complacent. There was a sharp bond market reaction, and one in equities, that have also been having something of a slow realization over the last few days, including yesterday. The market expectations for tapering show no let up, come what may. Maybe the markets are just testing Yellen, but it seems the Fed has entered a rather blinkered phase.

  30. Gravatar of Vivian Darkbloom Vivian Darkbloom
    30. January 2014 at 02:01

    “Vivian Darkbloom: What is important is what monetary policy leads to best long-term real growth. In any macroeconomic policy there are relative winners and losers.
    Free trade has crushed whole industries and even cities””should we be against free trade?”

    What exactly is this responsive to? I thought the initial question was which rate of NGDP growth would be “best for equities”. I don’t know the answer to that. Do you?

  31. Gravatar of Benjamin Cole Benjamin Cole
    30. January 2014 at 05:02


    I’d say about 6 percent NGDP growth would do fine for equity investors, but I also think it is not an important question.

    The important question is what rate of NGDP growth leads to highest (sustainable) real GDP growth?

    Again I think about 6 percent, though for the next few years we could do 7 percent nicely.

    If Buffet thinks equity investors do poorly in a mildly inflationary environment (about 3 percent) he seems to have forgotten the 1982 to 2008 run. We had varying rates of inflation, and terrific economic growth and a booming stock market.

    The risk going forward if of a too-timid, feeble, dithering Fed that faints at the slightest whiff of inflation. The FOMC needs smelling salts now if inflation pokes its head above 1 percent.

    All others:

    I think we have been duped, and rather easily too. The Fed says it has a target 2 percent inflation rate, and we believe them.

    So, here we are with the PCE dipping below 1 percent and the Fed cutting QE. The Fed has been undershooting the putative 2 percent target for five years.

    A crummy scenario may unfold, and it goes like this: The economy slows down, and we get very sluggish growth, around 1 percent, as the Fed pulls out of QE. Inflation drops even lower, maybe 0.5 percent.

    But the Fed can’t go back to QE, as it would then be essentially admitting to a grievous policy error. So it does nothing. This may persist for years, or until we get another recession,and at that point the Fed will have to go back to QE, since we will be near ZLB-land anyway.

    If it sounds a lot like Japan….

  32. Gravatar of TravisV TravisV
    30. January 2014 at 06:34

    The editorial board of the Wall Street Journal rips Ben Bernanke:

  33. Gravatar of Brian Donohue Brian Donohue
    30. January 2014 at 06:34


    Re Buffett (and our conversation at Marginal Revolution), he made comments in his shareholder letter a couple years ago about bonds representing “return-free risk” and suggested “they should come with a warning label right now”. Why? Inflation/currency debasement.

    My point is that Buffett is not suggesting that equity investors are uniquely exposed to this risk.

  34. Gravatar of Mark A. Sadowski Mark A. Sadowski
    30. January 2014 at 06:37

    Off Topic.

    The advance estimate of 2013Q4 GDP is in and NGDP grew by 4.6% at an annual rate and RGDP by 3.2%. (For the record my nowcast was 4.2% and 2.5% for NGDP and RGDP growth respectively.)

    This means that year on year NGDP growth was 4.2% in 2013Q4 up from 3.8% in 2012Q4 and RGDP was 2.7% in 2013Q4 up from 2.0% in 2012Q4.

    A careful reading of the CBO’s estimates from November 2012 indicates that the fiscal consolidation (the 2% payroll tax increase, the high income tax increase and the sequester) should have subtracted 1.3% from year on year RGDP growth through 2013Q4. Prior to the beginning of 2013, the CBO’s baseline forecast (without fiscal consolidation) was for 2.4% year on year RGDP growth in 2013Q4. Thus we should have experienced only 1.1% growth instead of the 2.7% that we did.

    A similar thing applies to the major private forecasters. The effect of the fiscal consolidation (again, the 2% payroll tax increase, the high income tax increase and the sequester) according to Bank of America, IHS Global Insight, Moody’s, Goldman Sachs, Morgan Stanley, Macroadvisers and Credit Suisse ranged from 1.0% to 2.0% of GDP, with the average estimate being about 1.6%. The baseline forecast prior to the beginning of 2013 of these same seven private forecasters was for year on year RGDP growth of 2.0% to 3.5% in 2013Q4 with the average forecast being about 2.7%. Thus the average forecasted RGDP growth adjusted for fiscal consolidation was about 1.1%.

    In short, RGDP growth ended up as high, or higher, than what was forecasted without fiscal consolidation.

    It sure looks like monetary expansion offset fiscal contraction to me.

  35. Gravatar of Brian Donohue Brian Donohue
    30. January 2014 at 06:39


    In all probability, you are correct with your near-term prognistication, based on current circumstances (“risk on/ risk off” mentality.) But your comment:

    “When U.S. stocks fall, bond prices should INCREASE, not decrease” sounds more like a general rule of negative correlation, which ain’t always the case.

    The 1980s saw bull markets in both stocks in bonds, while 1973-74 was a bear market for both. If it starts to look like stagflation (which, admittedly, it doesn’t right now), we could easily see higher rates and lower stock prices.

  36. Gravatar of Vivian Darkbloom Vivian Darkbloom
    30. January 2014 at 06:44


    I recall that and I don’t disagree but the other discussion was about tax consequences. I don’t want to rehash our other discussion about the tax consequences; however, as I’ve said there, if Buffett is right and inflation depresses bond prices, then a sale of a bond that decreased in value would generate a capital loss, deductible to the extent of a capital loss on any other capital asset, including shares of stock. The cash in your wallet or under your mattress or in a demand account is another story.

  37. Gravatar of TravisV TravisV
    30. January 2014 at 07:06

    Brian Donohue,

    Good point. That relates to our discussion of what the “optimal” NGDP growth rate is for investors.

    In all probability, an NGDP growth rate of 2.0% is too low and a growth rate of 12% is too high. As you point out, if long-run NGDP growth expectations fall from, say, 12% to 8%, both equity investors and bond investors should benefit.

    And if long-run NGDP growth expectations increase from, say, 2% to 4%, equity investors should benefit and bond investors should lose.

    Bottom line: when estimating the optimal NGDP growth rate, I don’t think we should try to maximize real growth. We should just make NGDP growth high enough that it comfortably ensures full employment and no higher.

  38. Gravatar of John Becker John Becker
    30. January 2014 at 07:56


    A thoughts that’s what your approach would be, but what target should Turkey set for the Nominal GDP? Over the past 10 years, their NGDP has been all over the place so you can’t pick 5% the same way you did for America. They’ve always had pretty high inflation though but growth has very strong some years and weak others.

  39. Gravatar of ssumner ssumner
    31. January 2014 at 15:06

    James, I think the Fed is a bit too tight, but I don’t see market evidence that anything fundamental is changing. I see another year of 4% NGDP growth and gradually falling unemployment. I wish it was 5% and rapidly falling unemployment.

    We’ve had other stock “corrections” that had little impact on the economy.

    Mark, Good point. I think the experts missed this because monetary offset is so counterintuitive. It doesn’t have to be “concrete steppes” although QE3 was part of it. The forward guidance factor is an underrated offset.

    John, I haven’t studied it, but I’d guess if they’ve had 11% in recent years they should gradually lower the target, by perhaps 1/2% per year, until down to about 5%. It would take 12 years to get there.

    But honestly that’s an off the cuff answer, I’d need to study the Turkish situation in more detail.

Leave a Reply