Monday, July 13, 2015

Did Monetary Policy Really Offset Fiscal Austerity in Canada?

The blogosphere is once again talking about Canada's successful fiscal austerity in the mid-to-late 1990s. Paul Krugman rekindled the conversation with this statement:
[L]ook at everyone's favorite example of successful austerity, Canada in the 1990s. Canada came in with gross debt of roughly 100 percent of GDP, roughly comparable to Greece on the eve of the financial crisis. It then proceeded to do a pretty big fiscal adjustment -- 6 percent of GDP according to the IMF's measure of the structural balance, which is about a third of what Greece has done but comparable to other European debtors. But unemployment fell steadily. What was Canada's secret?
Ramesh Ponnuru and I have argued numerous times that Canada's secret was a monetary policy offset. That is, monetary policy eased to offset the drag of fiscal tightening. Paul Krugman agrees in the above post. The evidence that we and others have pointed to in support of this view is the Bank of Canada cutting its target interest rate more than 500 basis points between 1995 and 1997.  

Some of our conservative and libertarian friends, however, are not convinced by this evidence. David Henderson and Robert Murphy, in particular, have pushed back against this view. They contend there was no monetary offset. Henderson questions how much influence the Bank of Canada actually has over interests rates. Murphy goes further and provides a list of data points that he claims show the Canadian success story did not rely on loose money. So are Henderson and Murphy's skepticism of the monetary offset warranted?

The answer is no. Let us start with the Henderson's claim, echoed by Murphy, that the Bank of Canada has little control over interest rates. This point is generally true for long-term interest rates, but not for short-term interest rates. Central banks intervene in money markets and peg short-term interest rates all the time. It is true that if a central bank cares about price stability its short-run interest rate adjustments will conform over time to an interest rate path determined by the fundamentals. For example, Canada being a small open economy has its interest rates determined in part by capital flows from large economies like the United States. But this is a long-run tendency that still leaves a lot of wiggle room in the short run for central banks to tinker with interest rates. 

But do not take my word for it. See the figure below. It plots the target interest rates for both the Bank of Canada and the Federal Reserve over the period in question. The 500 basis point cut by the Bank of Canada is evident and occurs against a relatively stable federal funds rate. If the Bank of Canada has no control over its short-term interest rates then why was it able to create such large deviations around the federal funds rate? If the Henderson-Murphy view were correct this should not be possible.

Again, over the long-run the fundamentals will kick in and cause these two interest rates to follow a similar path. Henderson and Murphy assume this long-run relationship will also hold in the short-run. But it does not as shown above. The evidence, then, points to the Bank of Canada exogenously lowering short-term interest rates during the period of fiscal tightening.

I must say I was surprised to see an Austrian like Murphy makes this argument. Any Austrian worth his salt believes central banks can and do manipulate short-term interest rates. To go from this traditional Austrian position to one above is hard to reconcile. Put it this way: Murphy's reasoning, if consistently applied, would lead one to accept Bernanke's saving glut theory for the low interest rates during the housing boom. But Murphy does not accept this view. So it is hard to understand why he would suddenly embrace this emasculated view of central banks.

Murphy does attempt to provide other evidence to support his view on the Canadian austerity experience. It is impossible, though, to draw conclusions from his evidence because he does not provide the proper context for evaluating it. For example, one cannot simply look at the growth rates over a few years of the monetary base and nominal GDP as Murphy does and conclude with certainty whether monetary policy was tight or loose. Instead, one has to evaluate them against what was expected by the public and or desired by the central bank. 

So let us do that for the monetary base and nominal GDP. Consider first the monetary base (excluding required reserves) as seen below. This figure shows both the monetary base and its pre-1995 trend. Note the one-time permanent increase in it that occurs in the mid-to-late 1990s. A permanent increase in the monetary base is a sure way to raise aggregate demand and offset fiscal austerity. The above trend growth strongly suggests explicit monetary easing during this time.

Next, let us look at nominal GDP in the figure below. It shows the nominal GDP relative to its trend path. 

Note that nominal GDP follows its trend path rather closely during the period of fiscal austerity. The Bank of Canada, in other words, did what was necessary to keep aggregate demand on a stable growth path during this time. Given the evidence shown above, the Bank of Canada offset the fiscal tightening via lower interest rates and a permanently higher monetary base path. This story is completely missed by Murphy's cursory look at nominal GDP growth rates over a few years. So yes, monetary policy did offset fiscal austerity in Canada in the mid-to-late 1990s. 

The policy implications from this experience are clear. Economies undertaking fiscal austerity are best served by expansionary monetary policy. It provides a viable path to obtaining a more sustainable debt level. The ECB, however, tightened monetary policy twice during the Eurozone crisis. Given the one-size-fits-all approach problems, this tightening proved excessive for the periphery countries and helped spawn the soveriegn debt crisis. Just imagine how different the Eurozone would be today had the ECB began its QE program back in 2008. 


  1. Yep. Good post.

    The one thing missing is the exchange rate, which is an important part of the story. Both nominal and real exchange rates depreciated a lot, which is exactly what you would expect, with monetary offset in an open economy.

  2. I'm trying to reconcile the numbers used by Murphy with the charts used in this post. Using his data his points about the growth in monetary base and NGDP slowing down in the years in question seem factually correct and support his overall points, but that appears not to show up in the charts used here by David.

    - Murphy uses "growth in CB balance sheet" and shows a slowdown in the years in question. In this post "growth in monetary base" shows a clear spike in 1995 or so. How does one explain this difference ? (are "the growth in its balance sheet" and " "growth in monetary base" actually different things ?)

    - Murphy has a table of annual NGDP growth which shows below trend growth in 1996-8. This is not shown (or at least not easy to see) in the NGDP chart in this post. Is Murphy's NGDP data wrong, or is the slow-down in NGDP growth just hidden in the chart here by the long timeline (1990-2010) used ?

  3. Kevin, Not Trampis, and Nick: Thanks. Yes, I should have added the exchange rate which only further confirms the monetary easing story.

    Market Federalist: I am not sure what Murphy's "CB balance sheet" is, but I am clear as to what is the monetary base. You can find it here . Look for the monetary base excluding required reserves option. You have to scroll the down the page a bit to get to it.

    Regardaing his NGDP measures, my whole point is that one cannot say anything about trend NGDP growth without seeing the actual trend. And that is something one cannot tell by looking at the few years of growth rates in Murphy's post. For example, one year of slow NGDP growth may actually be bringing NGDP to trend if it had previously been growing too fast. That is why a figure like the one I provide is more informative.

  4. David, I never said the Bank of Canada couldn't influence interest rates. You're right, that would be impossible to come out of the mouth/keyboard of an Austrian.

    What I argued was that the fall in interest rates wasn't due to a decision to loosen monetary policy. Never reason from a--well, you know.

  5. David, I don't understand your NGDP point. Are you saying the Bank of Canada had super loose money from 1993-1995, which it scaled back just to loose money from 1996-1998, but thank goodness in absolute terms that loose money (not super loose, mind you) was enough to offset the fiscal austerity?

  6. Try it this way, David. Doesn't Sumner quote Friedman ad nauseum on saying that the way to judge monetary policy is NOT to look at movements in interest rates, but instead to look at NGDP growth and inflation rates?

    So didn't I show that during the period in question:
    (1) Base growth slowed.
    (2) Price inflation didn't increase.
    (3) NGDP growth (generally) slowed.

    So wouldn't Sumner have to agree the Bank of Canada was tighter during the 3 years after fiscal austerity began, than the 3 prior years?

    Are you OK with that, and you're just saying it was still looser than it had been, say, over the prior 10 years?

  7. And anyway, what the #)$# are you talking about, David? (I'm not yelling, I'm exasperated.) Your own chart shows NGDP growth falling below trend, precisely when austerity kicked in.

    So we're not disagreeing on anything. Your own metrics show that the BoC tightened when austerity kicked in.

    Just give in to the Good Side of the Force: Government spending money doesn't make an economy richer. We don't need to "offset" a great policy decision with the printing press. Capitalism works.

  8. David, last one I promise (for a 24-hour cycle): Please plot the monetary base and the monetary base (excluding required reserves) series. You'll see they are identical after that jump point.

    So that's why my Total Assets didn't pick up anything. It doesn't look like the BoC actually did anything with its purchases, but instead did it change reserve requirements in 1995?

  9. (Oops the jump occurs in mid-1994, not in 1995.) <==Doesn't count against my pledge to not comment again for 24 hours.

  10. Bob, glad you dropped by. But wow, what a flurry of responses. Let me attempt to respond to all of them in one comment.

    1. Your response on interest rates is still baffling. The Bank of Canada explicitly changed its target interest rate. It didn’t wake up one day and find that the short-term interest rates had magically fallen on its own. Rather it made a conscious choice to lower its interest rate target over 500 basis points. So as best I can tell, you seem to be arguing that this independent action by the Bank of Canada was made easier by other developments. If so, why do you invoke this argument here and not in 2003-2005 for the Fed? The same argument could be made then for the saving glut.

    Consistency questions asides, though, the biggest problem with your interest rate claim is the awkward fact that the Bank of Canada significantly deviated its target interest rate away from the federal funds rate. This is huge and there is no way get around the implications of this fact.

    2. Regarding the monetary base, you want to use the one that adjust for required reserves. That is the monetary base that actually is used as money and the one that matters. This is why, for example, the St. Louis Fed’s Monetary Trends publication only reports the monetary base adjusted for required reserves. And this is the series I graph above. So yes, the Bank of Canada did increase the monetary base that is used as money.

    Regarding the exact timing of its increase I don't see a problem here. Yes, the monetary base does take off in mid-1994 and continues to grow above trend through much of 1995. But so what? Do you really think the part that increased in 1994 had no bearing on aggregate demand growth in 1995 and beyond? There are lags to monetary policy and this was a permanent increase!

    3. Finally, the deviations of NGDP around its trend during the austerity period are trivial. I could have easily drawn a trend line where they were exactly on path. But I didn’t because the point is NGDP was roughly kept on track. The Bank of Canada was doing Milton Friedman’s “Thermostat” approach to monetary policy that Nick Rowe has talked about so many times. If it helps, compare the austerity period deviations to the the post-2007 deviations in my figure. The latter ones are large and matter. There is no redrawing of trends that will get one away from the spending gaps there. But this is not the case for the austerity period.

    4. Finally, none of us Market Monetarists have ever claimed this experience shows “government spending makes an economy richer.” Rather, we use it to show how to reduce government spending in the lest disruptive manner. Most of us want freer markets and more capitalism. But to get there one has to take a viable path. Almost every study that has systematically looked at successful cases of fiscal austerity has found that it required an easing of monetary policy. Canada is just another data point that supports this view.

    1. @ Prof. Beckworth.

      "Almost every study that has systematically looked at successful cases of fiscal austerity has found that it required an easing of monetary policy. "

      I find your overall case very convincing, however "fiscal austerity" can mean both higher taxes and less goverment spending, or a combination of it. For an Austrian, these things are so different that the term "austerity" does not make any sense to them. I guess Bob Murphy would probably reply that, if you do less government spending and LOWER taxes, then you would not need monetary easing. I am curious if there are studies that consider this latter combination?

  11. @Bob Murphy,

    You say
    "So didn't I show that during the period in question:
    (1) Base growth slowed.
    (2) Price inflation didn't increase.
    (3) NGDP growth (generally) slowed.

    So wouldn't Sumner have to agree the Bank of Canada was tighter during the 3 years after fiscal austerity began, than the 3 prior years?"

    If you imagine a situation where the govt reduced its budget deficit, and the CB attempted to use monetary offset but everything else (demand to hold money etc) remained equal then if the CB was successful you would see inflation and NGDP growth stay on trend , and an increase in the base to offset the fiscal reduction.

    If the CB did monetary offset but not quite enough to fully adjust for the fiscal reduction then you would see a slight drop in inflation and NGDP growth and an increase in monetary base, which is what we see (at least using David definition of "monetary base").

    I think you would have to say that Sumner might conclude that monetary policy was (slightly) tight compared with what was needed to keep NGDP fully on target, but still loose enough to mostly offset the fiscal changes.

  12. Jose Romeu RobazziJuly 14, 2015 at 11:45 AM

    Prof, Beckworth and others
    Maybe the (somewhat) slower NGDP em 1994-1996 is exactly what one should expect after a huge fiscal shock and monetary offset policy by the central bank. Should the monetary offset not occur, NGDP growth could have been much slower ....

  13. The real lesson is central banks should print more money. Unabashedly so.

  14. Stephen Williamson doesn't sound convinced about monetary offset (to me anyway):

    "Thus, if the job of the central bank is controlling inflation, I think we would say that a lower inflation rate is an indicator of tighter monetary policy. In other words, whatever was causing the exchange rate to depreciate in Canada during the 1990s, it wasn't monetary policy."

    He concludes thusly:

    "Are there lessons that we can translate to Greece in the current era? Only the obvious I think. Political stability and good government are important."


    1. A lower inflation rate due to higher productivity growth is not tighter monetary policy. Williamson is reasoning from a (basket) price change.

    2. A brief, but direct response comment from Williamson. (to the article, not your comment).

  15. Jose Romeu RobazziJuly 16, 2015 at 3:55 PM

    @Bob Murphy
    I tried to comment on your posts in Mises Canada, but somehow I couldn't. I saw your data. And although your contention with the keynesian story seems convincing to me, your analysis of monetary offset is correct, but your conclusion, at least to me, does not support your claim that there was no monetary offset. Yes, you are right, forget interest rates, focus on NGDP growth rates, and looking at that,the amazingly stable NGDP growth between 1993-1998 (fell from 5% average in the first 3 years do 4.3% average in the last three, prety minor change). But please don't forget that at that time, short rates were reduced sharply. Yes, given the stable NGDP growth, that policy was very successfull (in the context of a significant fiscal shock). Using market monetarists jargon, yes, policy stance actually was neutral at the time (stable NGDP growth), but that is actually successful monetary offset, given the abruptness and magnitute of the fiscal shock. Again, your contention that monetary policy stance was not neutral (correct) does not mean there was no monetary offset, actually, it says the central bank, by lowering rates aggresively, was able to keep NGDP on a stable páth, wich is actually the goal of a successful central bank. Huge success, no failure here. Think what could have happened should the central bank kept rates unchanged at the time (monetary policy stance would have been very tight, and perhaps NGDP would have slowed significantly).

  16. Hi everyone,

    For those who still care about this debate, I (finally) wrote up my position in this new Mises Canada post. David's graph above might be misleading you, because the shift is entirely due to changes in reserve requirements, not in asset purchases by the B of C. I think that makes the Market Monetarist narrative difficult here, but I agree it definitely helps their case tremendously compared to my approach of looking just at asset purchases.

    Also, David, I apologize for a poor objection above. You're right (and others pointed it out too), in your framework even a slight dip in NGDP growth would be consistent with the Bank of Canada offsetting the drag from a huge dose of fiscal austerity.

    However, just to make sure you get where I'm coming from, I deny that government spending cuts lead to problems which must be "offset."