Hetzel on “Learning from the Pandemic Monetary Policy Experiment”

Robert Hetzel undoubtebly is one of the most important monetary thinkers of our time and for more than five decades he has been involved in US monetary policy making and has furthermore greatly contributed to monetary theory and monetary history.

And he is one of my absolut biggest idols in the world of monetary thinking and I am proud to call Bob my friend.

Luckily Bob and I rarely disagree, but I continue to learn from Bob and anybody interested in monetary matters should read Bob’s papers.

Bob now has a new paper out – published by the Mercatus Center where he now is Senior Affiliated Scholar.


In response to the COVID-19 pandemic, which unfurled starting in March 2020 and raised unemployment dramatically, the Federal Open Market Committee (FOMC) adopted a highly expansionary monetary policy. The policy restored the activist policy of aggregate demand management that had characterized the 1970s. It did so in two respects. First, the FOMC rejected the prior Volcker-Greenspan policy of raising
the funds rate preemptively to preserve price stability. Second, through quantitative easing, it created an enormous amount of money by monetizing government debt. Inthe 1970s, the activist policy was destabilizing. Reflecting the “long and variable lags”phenomenon highlighted by Milton Friedman, a temporary reduction in unemployment from monetary stimulus gave way in time to a sustained increase in inflation. In response, the succeeding Volcker-Greenspan FOMCs rejected an activist monetary
policy in favor of a neutral policy. That neutral policy concentrated on achieving low trend inflation and abandoned any attempt to lower unemployment by exploiting the inflation-unemployment tradeoffs promised by the Phillips curve. The success or failure of the FOMC’s activist monetary policy offers yet another opportunity to understand what types of monetary policies stabilize or destabilize the economy.

Read Bob’s paper here.


The Fed is still way behind the curve – what an “implied NGDP level target” is telling us

I started this blog more than a decade ago in 2011 primarily because I was frustrated with the way monetary policy was conducted around the world.

At that time it was my clear view that both the Federal Reserve and the ECB had far too TIGHT monetary policy and that the reason that the 2008 shock to the global financial system had developed into a deep recession because the Fed has failed to react appropriately to the sharp rise in dollar demand during the Autumn of 2008.

My analysis of the situation at that time was based on a fundamental monetarist analysis combined with what financial market indicators were telling me about the outlook/expectations for nominal demand in the economy.

Other econ bloggers like Scott Sumner, David Beckworth and Marcus Nunes were using – and still are – a similar approach. This approach – or school of thought – later became know as Market Monetarism. A term I coined in a paper back in 2011.

At the core of market monetarist thinking is also advocacy of the use of a NGDP level target. Hence, market monetarists like Scott Sumner and myself have argued that at the core of what central banks should do is to keep nominal GDP (NGDP) on a “straight line”. Or rather, the for example the Fed should ensure that over time NGDP growth at a fixed rate – for example 4% – and that it should be level targeting meaning if the target is undershot one year – NGDP growth is below 4% – then Fed needs to ensure that it will be above the target in the following period and there by ensure that NGDP returns to the targeted level.

There are a number advances with NGDP level targeting over traditional inflation targeting.

First of all, it is the general consensus that central banks should not respond (try to increase nominal demand) to supply shocks. A narrow focus on inflation – even when corrected for energy and food price fluctuation – risk causing central banks to nonetheless to respond such supply shocks. A good example of this if the ECB catastrophic interest rate hikes in 2011, which essentially was a response to an increase in energy prices driven by a supply shock (the 2011 Japanese tsunami).

Contrary to this an NGDP target would allow for inflation to rise temporarily in the case of a negative supply shock while keeping nominal demand growth on track.

Second, by targeting the level rather than the growth of NGDP (or prices) the central bank will signal that it will make up for past mistakes. This ensures that the market will do a lot of the lifting in terms of conduct of monetary policy and thereby ensure a faster return to the target.

The story now is the opposite of in 2008

Neither the Fed nor the ECB implemented a NGDP target following the shock in 2008. However, I think it is fair to say that market monetarist thinking have had a substantial impact on monetary policy discussion and both in Europe and North America over the past decade.

But we didn’t quite make it all the way. The ECB has maintained its inflation target with a few adjustments, while the Fed essentially has been through a process of nearly continuous adjustments to its monetary target.

It should also be noted that the Fed actually did not officially have an inflation target prior to 2008. However, from 2012 the Fed has officially had a 2% inflation which have been adjusted numerous times.

That being said I have earlier argued that de facto the Fed had introduced an 4% NGDP level target in the summer of 2009 – basically with out announcing it.

The graph below illustrates that quite clearly.

The graph shows us that starting at the end of 2015 NGDP started to undershoot the “target” but later returned to the target by the end of 2017.

The undershooting was casused by then Fed chair Janet Yellen’s premature monetary tigthening that was initiated (announced) in October 2015. A decision that I strongly criticised at the time. See for example this blog post from 2016.

However, over all from early 2010 to early 2020 the US NGDP level was kept close to 4% path.

The Covid/Lockdown shock of 2020, however, caused NGDP to drop substaintially below the unannounced NGDP level target and even though the lockdowns likely significantly has distorted the US economic data it seemed pretty clear that aggressive monetary easing was warranted in the US to ensure that we would not have a repeat of the 2008-9 deflationary shock.

And the Fed reacted – fast and aggressively – and consequently we saw a swift recovery in NGDP to the previous 4% path.

So HAD the Fed officially been targeting a 4% path like the one we see in the graph then one would have to argue that the Fed’s policy reaction had been appropriate and that the Fed had done it’s job.

And I have certainly said so – the Fed did the right thing initially and the operation worked.

However, during the second half of 2020 I became increasingly concerned that the Fed was overdoing it in terms of monetary easing and in April 2021 I warned that the US might be heading for double-digit inflation. This of course quite closely coincided with the actually NGDP level starting to move above the 4% path “target” level.

But the Fed do not have a NGDP level target

While we clearly can use NGDP (and NGDP expectations) as a very useful indicator of the monetary policy stance the Fed do not in fact target the level of NGDP.

Rather officially the Fed now has an “Average Inflation Target” (AIT). The AIT was announced by Fed chief Jay Powell in August 2020.

The word “average” is important as it means that the Fed should  “achieve inflation that averages 2 percent over time” as stated by Powell.

This essentially means that the Fed is moving towards “level targeting” in the sense that if inflation has been below 2% for some time then the Fed needs to make up for this by ensure inflation above 2% in the following period.

And as the Fed also is fairly clear that it should not respond to supply shocks we are moving closer to a NGDP level target.

The problem of course is that the Fed has failed to announce what is the starting point of this regime. It is therefore unclear whether we should look at “average” inflation going back one, five or ten years and it is unclear for how long the Fed can take to bring the average back to 2%.

Furthermore, we cannot just look at for example a five-year moving-average of inflation as monetary policy clearly should be forward-looking.

So is the Fed for example targeting the average inflation over the past three-year plus the expected average inflation in the coming two years?

All this is unclear.

Introducing an “Implied NGDP level target”

So the AIT is actually not giving us a very clear indication of whether or not monetary policy is too easy or too tight presently in the US.

We can of course from the communication from the Fed conclude that the Fed presently see a need for monetary tightening, but again it is really unclear what makes the Fed come to this conclusion.

So maybe the Fed needs a bit of help from market monetarist thinking.

I have therefore tried to construct a measure that we (and the Fed) can use to assess for monetary tightening in the US.

I call this measure an “Implied NGDP level target”.

The idea is that the target over time will ensure that “average inflation” is at 2%.

NGDP (N) by definition is real GDP (Y) times the price level (P):


Based on this we can calculate an implied level of NGDP – or what we could call N-star (inspired by the P-star model)

We define N-star in the following way:

N-star = (P-target level) * (Potential Y).

The “P-target level” is simply a 2% path for the US GDP deflator (hence the inflation target as defined in level), while “Potential Y” is potential real GDP as calculated by the US Congressional Budget Office (CBO).

If the Fed conduct monetary policy in such a way as to keep actual NGDP close to N-star over time then it will at the same time ensure that inflation (measured by the GDP deflator) more or less will average 2% over time. Hence, this is a NGDP target that is consistent with the Fed’s 2% Average Inflation Target.

We, however, have one challenge and that is to determine what should be our “starting date” for this target. Should be go back to 2010? Should we start in 2012 when the Fed first offcially introduced the 2% inflation target or should we start in August 2020 when the AIT was introduced?

The best starting point in my view is a starting where there are no imbalances in the REAL economy.

Hence, basically at a point where monetary policy is neutral. This would mean that we should find a recent date where real GDP is close to potential real GDP and/or unemployment is close to the structural level of unemployment (NAIRU).

The Fed officially has a “dual mandate” meaning that is officially should ensure “price stability” (that is the AIT) and “maximum sustainable employment”.

We already got “price stability” covered with our implied NGDP target and if we base the “starting point” on “maximum sustainable employment” then we have a NGDP target that is consistent with Fed’s dual mandate.

The graph above shows the actual US unemployment rate (the blue line) and the CBO’s estimate for the noncyclical rate of unemployment (NAIRU).

We see that just prior to the shock of 2020 unemployment was in fact below CBO’s estimate of NAIRU. Therefore, we need to go further back to find our starting point for our implied NGDP level target.

Last time prior to 2020 that the unemployment rate was equal to NAIRU was in the first quarter of 2017 with unemployment of just above 4.5%.

So now we have both our starting point and our implied growth rate of our implied NGDP level target.

The graph below show the actual NGDP level and our implied NGDP level target as well as the “NGDP gap” with it the percentage difference between the two. A positive NGDP gap implies that monetary policy is too easy.

The first thing to note is that US monetary policy became excessively easy during the Trump presidency. Whether this reflects president Trump’s very public pressure on the Fed to ease monetary conditions or not is a matter that is open for discussion, but at least based on our implied NGDP level target US monetary policy was indeed too easy during this period.

That being said this “easiness” was within a reasonable “uncertainty band” and in general it is hard to argue that monetary policy became unanchored in this period.

It is also clear that PRIOR to the Covid-lockdown-shock in 2020 NGDP returned to the level target in late 2019 and by the beginning of 2020 the monetary stance was more or less perfectly calibrated.

We, however, also see that the shock of 2020 was very substaintial, but also that the Fed’s appropriate (initial) monetary easing fast brought the NGDP level back towards the target level as also discussed above.

From early 2021 we see that NGDP started to overshoot the target level and hence at that time monetary conditions clearly had become too easy.

In fact as monetary policy works with long and variable leads as Scott Sumner likes to say the Fed obviously should have initiated monetary tightening somewhat earlier.

Hence, in the Autumn of 2020 it was becoming increasingly clear from watching the financial markets that NGDP growth would be very robust and the Fed could easingly have forecasted that actual NGDP would overshoot the target level in early 2021.

However, during that period the Fed rather downplayed this risks and continued to argue that inflationary pressures were temporary and was due to supply side factors.

Looking at the graph above we have to conclude that that was a major policy mistake and the Fed’s reluctance to initiate monetary tightening has caused NGDP to very significantly overshoot the implied NGDP level target.

Based on this it is hardly surprising that US inflation has spiked and market inflation expectations have increased significantly.

Therefore, it is also blatantly wrong when for example President Biden blame the increase in US inflation on geopolitical factors. Vladimir Putin is to blame for a lot of bad things, but not higher US inflation.

As Milton Friedman used to say “Inflation is always and everywhere a monetary phenomenon and that is certainly also the case this time around.

The Federal Reserve has allowed nominal GDP growth to grow far too fast and consequently we have over the past 12-18 seen a substaintial acceleration in inflation.

Positive and negative supply shocks obviously can influence the inflation data from month to month or even quarter to quarter, but supply side factors should not be used to ignore the fact that monetary conditions remain far to easy in the US.

Bring back NGDP to the target over the next five years

The way forward for the US monetary policy right now is for the Fed to clearly announce that to maintain price stability it necessitates that the Fed ensures that NGDP grows at a rate over the medium-term, which is consistent with 2% inflation.

Furthermore, as the Fed is targeting AVERAGE inflation then it needs to “undo” previous mistakes. Hence, NDGP needs to be brought back to the implied target level for example within the next five years.

Below I have simulated such a scenario.

I have assumed that potential real GDP growth in the US will be around 2% in the coming five years. This is also more or less CBO’s forecast and as we target 2% inflation our implied yearly NGDP target growth rate is 4%.

To close the NGDP gap over a five year period actual NGDP growth should hence be slower than 4%.

In fact my simulation shows that NGDP growth need to be slowed to less than 3% (to 2.8%) on average over the next five years.

The graph below shows that simulation.

In a graph this looks like an easy policy to implement, but in reality it would be a lot harder as it would necessitate a very significant slowdown in NGDP growth and a “sudden stop” to NGDP could easily cause the US economy to fall into a recession and potentially also trigger financial distress.

The alternative to slowing NGDP growth, however, is that NGDP growth expectations and therefore inflation expectations permanently shifts up.

This would make it a lot harder (more costly in terms of an increase in unemployment) to re-anchor expectations to ensure 2% in the medium-term.

To me there really isn’t any way around this – the Fed needs to slow NGDP growth and announce a clear target for NGDP for the coming five years. That might hurt in the near-term, but by not doing it the costs will increase sharply going forward.

What will the Fed actually do?

Broadly speaking I believe that Jay Powell and the majority of member of the FOMC understand the logic of my discussion above and understand the risk that if the Fed does not commit to slowing NGDP growth (the Fed will use another language) then it comes with the serious risk of repeating the mistakes of the 1970s.

Therefore, we certainly should expect the Fed to turn more hawkish going forward. However, the problem for the Fed is that it has not defined and announced a clear and transperant target – and the focus of attention for the Fed might change over time.

I therefore think (fear) that the Fed will continue to be behind the curve – tightening monetary conditions too slowly – but at some point the Fed will slam the brakes. That is likely to happen within the next year.

At that time it is too late to avoid a recession or as Rudiger Dornbusch once said economic expansions do not die of natural causes rather “they were all murdered by the Fed over the issue of inflation.” 

A recession – in 2023 or 2024 – then will cause the Fed to do a u-turn even if inflation expectations remain well above 2%.

Hence, I certainly fear that we are entering a period of stop-go or rather go-stop-go monetary policy in the US. And the ECB will likely follow the same path.

With that of course comes more economic and financial uncertainty and it is the badly needed that the Fed (and the ECB) get to work on a proper monetary policy framework.

In this blog post I have outlined such framework. Unfortunately I am not too optimistic that the Fed will listen.

The ruble has appreciated exactly BECAUSE the sanctions are working

The question I have been asked most over the last week is why the Russian ruble is been appreciating despite the fact Russia has been hit by extensive economic sanctions.

I must honestly admit that since the sanctions were introduced I have not spent much time following the development of the ruble – or rather the ruble exchange rate on our screens does not really tell us much as the ruble today cannot really be said to be a convertible currency in the traditional sense.

If, for example. showed up at an European bank with rubles and wanted to exchange them for euros then in practice you would hardly be able to do so.

This is because the Russian central bank (CBR) has been sanctioned and what happens in practice is that if you have to exchange rubles for euros the money is basically deposited in the Russian central bank, which then gives back euros – directly from the Russian foreign exchange reserve.

This can now not be done as the CBR simply does not have access to trade in e.g. euros or dollars. The foreign currency remains in the foreign exchange reserve, but the CBR just cannot use it. This is equivalent to having money in your bank account, but the online banking and credit cards do not work.

At the same time, Russia is largely shut out of the so-called SWIFT system used to conduct international currency transactions. This means e.g. that in practice you cannot transfer money between e.g. Denmark and Russia. So if e.g. If a Danish company sells a product to Russia, it will now in practice not be able to receive payment for the product.

On the contrary, Russia’s oil and gas exports, which are basically the only thing that Russia exports, are not covered by the sanctions, and Russia can continue to receive payment for e.g. gas exports to EU countries.

Slightly simplified, Russia can now not import very much, but the country can continue to export. In other words Russia’s imports have collapsed while exports are less severely affected by sanctions.

Consequently, Russia’s trade balance surplus has increased sharply in the past month.

And that is basically the reason why the ruble has appreciated in recent weeks.

Does that mean that the sanctions do not work?

If you have a mercantilist view of the world – that is if you think that it is always good to export and that imports are bad then the sanctions are great for Russia.

This view, however, corresponds to saying that it is good to work and bad to consume, but there is basically only one reason to work – namely to consume. We do not work for fun – we work to be able to consume – whether it is food or luxury goods.

Right now, Russia is now basically forced to “work”, that is, to sell oil and gas, but at the same time Russia cannot use the revenue.

At the same time, it is part of the stpry that the Russian government has severely restricted the right and opportunity of Russian citizens and businesses to own foreign currency, and citizens and businesses are forced to convert the bulk of their foreign exchange earnings into rubles. This may also to some extent also have supported the value of the ruble.

In addition, it should be noted that Russian interest rates have skyrocketed and that is basically the market compensation for the fact that the ruble has lost value – and is expected to lose additional value in the future.

Finally, the Russian foreign exchange reserve has fallen quite sharply in the past month. Thus, the weekly foreign exchange reserve data from the Russian central bank show that the foreign exchange reserve on February 25 amounted to USD 629.4 billion. By March 25, that number had dropped to USD 604.4 billion.

This is remarkable given that we know that the trade surplus has risen sharply and that Russia cannot trade Western currencies.

Do the sanctions work?

If one is to judge whether the sanctions “work” then one must relate that to what the purpose of the sanctions is.

The purpose of sanctions is not just to “punish” Russia and revenge is certainly not the main purpose. The main purpose must be to sharply reduce Putin’s ability to continue waging war.

Russia needs imports when it comes to waging war. For example Russia must use technology in military production or tires for military vehicles etc. All indications are that these imports have now been severely hampered.

At the same time Putin is dependent on economic growth both to finance the war itself (and possible future “adventures”) and to bribe those he need to support him both in the wider Russian population and among oligarchs, and more importantly in defence and the security apparatus (the so-called Silovik).

Hence, the sanctions hit the Russian economy and Putin regime very hard, but as Russian oil and gas exports are not covered by sanctions the West will be affected to a much lesser extent.

If on the other hand the EU and the West in general closed off imports of Russian oil and gas it would send oil and gas prices skyrocketing and the European economy would be hit quite hard and the European economy would most certainly end up in recession.  

A recession that could potentially threaten popular support for the hard line against Putin and which could create divisions among the EU countries.

On the contrary, Russia is already in a situation where oil and gas revenues, which are typically in euros or dollars, are not worth much, as Russia basically do not have access to spend euros and dollars.

The only thing these foreign exchange earnings today can be used for is basically to service the Russian debt in these foreign currencies. Again, it is primarily the Western investors who own the Russian government bonds that are benefiting rather than the Russian government.

So yes the ruble has appreciated in recent week, but it actually reflects that the sanctions against Russia are quite well calibrated – it maximizes the negative effects on Putin’s ability to wage war and at the same time it minimizes the negative effects on the global economy.

The “black market value” of the ruble has fallen sharply

Finally, it must be said that the ruble exchange rate that we can observe on various financial news sites is not necessarily the actual exchange rate.

In Soviet times, there was an official ruble exchange rate, but it did not have much to do with the real exchange rate. The actual exchange rate was the one you could read if you asked the black market currency traders on the streets of Moscow when exchanging physical dollars for rubles or the other way around.

In currency terms, we are now to a large extent back to the Soviet era, but today we have an easier way of observing the “black market rate” of the ruble. Hence, you can still trade bitcoin and other cryptocurrencies in Russia.

If I want my money out of Russia, then I can exchange my rubles in Russia for bitcoin, and then fly to for example Turkey and get my bitcoin paid out for dollars or Turkish lira on a Turkish crypto exchange.

Thus, by using the “local” Russian bitcoin price, we can calculate what the real value of the ruble is if one were to exchange it for dollars on the “street” in Moscow.

This “premium” which you have to pay for for example Bitcoin in Russia relative to Bitcoin elsewhere has fluctuated between 10-30 percent over the past month. It is smaller now than in the days immediately after the war started, but it is still significant.

Finally, Russian exporters of goods must also accept a discount when they export. Thus, the price of oil from Russia today is about 30 percent lower than what oil otherwise costs on the world market.

We can thus conclude that, paradoxically, it is the sanctions – because it works – that help keep the ruble exchange rate up, but the goods that Russia exports are sold at a significant discount, and at the same time the “street price” of the ruble is significantly below the “official “Exchange rate.

There is therefore no reason to be blinded by the development of the ruble exchange rate, which we can observe on various finance sites, as it does not really say much about the financial flows in and out and Russia and the state of the Russian economy.

%d bloggers like this: