Webinar: Too much money chasing too few goods

My former colleagues at Danske Bank has asked me to explain why I believe that the US might be heading for double-digit inflation in 2021-22.

Watch the webinar here.

Heading for double-digit US inflation

I have spend a lot of my time since 2008 arguing that US monetary policy was much less expansionary than most people thought and has been arguing for a more aggressive response from the Federal Reserve to combat deflationary pressures.

Furthermore, I have last year welcomed the Fed’s policy respond to the lockdown crisis – see for example here – as I feared a repeat of the deflationary shock of 2008-9.

Furthermore, even though I have been somewhat worried about the sharp pick up in US broad money supply growth I for while was of the view that breakeven inflation rates were quite low and as a consequence we should not worry too much about inflation.

However, over the last couple of months I have become more and more convinced that particularly elevated stock prices, property prices and commodity prices reflected sharply increased inflation pressures.

I have therefore, gradually changed my view on inflation and am now quite convinced that inflation will pick up very strongly in the US. In fact, I now seriously fear that we are heading for double-digit inflation in the US before the end of this year.

I has taken me a lot of time to spell out this view publicly because I full well-know that this certainly is not the consensus view and it is certainly not (fully) priced – at least not by fixed income markets. As somebody who tend to believe markets are close to efficient I don’t lightly second-guess the markets, but I have also convinced myself that this ‘mis-pricing’ in particularly in the fixed income markets at least to some degree reflects a massive liquidity effect that ‘overshadows’ rising inflation expectations.

But today I call it – the US is heading for double-digit inflation in 2021 and it will happen very fast. What I expect is not necessarily permanently higher inflation, but rather a sharp one-off jump in the US price level.

What happens after this starts to unfollowed I believe is a lot harder to forecast and it will strongly dependent on the Fed’s response to this jump in the price level. One possibility is that we will see a serious erosion of Fed’s credibility and longer-term inflation expectations will jump. Alternatively the Fed moves aggressively to curb rising inflationary pressures and is able to convince the markets that this is indeed a post-pandemic one-off jump in the price level, but not permanently higher inflation.

No matter what this is likely to be THE main topic for global financial markets in 2021 and I have a hard time seeing this playing out without causing some volatility in global financial markets.

I am very hesitant even calling this a “forecast” for US inflation. Rather it is a simulation of what we should expect to happen to the US price level if Fed allows the ‘liquidity overhang’ to feed fully through to prices. I doubt that will happen but on the other hand the Fed seems to be ‘deliberately’ behind the curve so at least for the next 3-5 month we are likely to see a very sharp increase in the price level.

Below is my ‘simulation’ for US inflation.

This simulation is based on the so-called P-star model.

In a Twitter thread earlier today I discussed the model and the implications for US inflation and partly the implications for asset prices.

See the discussion below.

Again, I don’t make this forecast lightly. There is a lot of things that can change to change the outcome, but again and again over the last couple of months I have postpone making this forecast because I know it is a rather wild prediction, but I can no longer find excuses not to make this forecast because I fundamentally believe the analysis is correct and I have to follow the logic of the analysis and the numbers and the only conclusion I can reach is that we are in for a very sharp increase in US inflation – very soon.

I am as always happy to discuss the my analysis with clients and potential clients. Contact: lacsen@gmail.com

Post-Covid Recovery – fast, but inflationary

I have become substantially more worried about inflation than I was last summer and I now believe we will see a very significant increase in inflation in the US in the coming months.

I will write more on that in the coming week, but until then have a look at this presentation that I did recently at the Danish Chamber of Commerce in Lithuania on April 15.

Watch the presentation here.

R* strikes back: The Fed will hike sooner rather than later

The Federal Reserve’s mandate is clear – ensuring the maximum level of employment while at the same time ensuring price stability.

Over time the Fed’s interpretation of this mandate has changed, but we can maybe get a bit closer by saying that the Fed has an ordering of the dual mandate.

First the Fed wants to insure that inflation (inflation expectations) over time is close to 2%.

Second, once this is the case the Fed will try to “push” economic activity in the “right” direction – meaning that the Fed will ease (tighten) monetary conditions if US unemployment is above (below) the structural level of unemployment (NAIRU).

This of course also is what is reflected in for example the traditional Taylor rule, where the Fed ‘sets’ its policy rate – the Fed Funds rate – to reflect both inflation relative to the inflation target and the output gap.

However, what is not often discussed is that the Fed does not actually ‘set’ interest rates. Or rather the Fed cannot just set a level of interest rates, which it likes.

Rather the Fed will have to set the interest rate relative to the structural level of interest rates – or what the Swedish economist Knut Wicksell termed the natural interest rate – or which is often today referred to as r*.

Hence, the task of the Fed is to set monetary conditions so to ensure that actual short-term interest rates are close to r*.

If the Fed tries to push interest rates significantly below r* for a sustained perioded by injecting large of amounts of liquidity into the economy then sooner or later we will get a spike in inflation.

Therefore, the Fed really isn’t in control of interest rates – or at least the fed cannot independent of its inflation target set interest rates.

All the Fed can do is to shadow r* and thereby ensuring than inflation will be close to its inflation target.

The challenge for the Fed is that r* isn’t constant over time.

Hence, we know both theoretically and empirically that the real interest rate has been moving up and down and is different from country to country.

Hence, r* reflects structural factors – such as demographics and productivity.

In this blog post the purpose is to try to help understand the development in r* over time and to assess what implications this has for present day monetary policy in the US.

To shed light on this matter I have estimated a rather simple, but nonetheless robust model for the development in the actual Fed Funds rate since the early 1960s. This is part of a bigger research project I am doing at Copenhagen Business School.

Normally models for the Fed funds rate is seen as ‘policy rules’.

The purpose here is not to estimate a ‘policy rule’ but rather to think about the Fed’s challenge to shadow r* by estimating a model for what the Fed historically has done with interest rates.

Estimating Fed’s r*

Basically the starting point for the analysis is that the Fed is both ‘targeting’ real and nominal economic activity, which is that the dual mandate is about. Real economic activity in terms of the ‘output gap’ and nominal activity in terms of inflation.

However, we also know that this has been changing over time and the Fed’s 2% inflation target is a fairly “new” target.

Furthermore, we also know that r* is not constant over time so even when inflation is at the inflation target and the output is closed (zero) then the Fed might have to change its policy rate to reflect changes in r* that are a result of for example higher saving in the US economy (for example due to consolidation of public finances) or to reflect higher or lower productivity growth.

To reflects these factors I have estimated a model for the Fed funds rate since 1962.

To explain the development in the Fed funds rate I use the following variables:

First of all I look at potential nominal GDP growth (relative to population) growth.

This of course partly reflects the Fed’s inflation target and that we know from the so-called Fisher Equation that high inflation should cause higher nominal interest rates. Furthermore, this also captures the fact that higher productivity growth should be expected to lead to higher interest rates – as we for example recognize it from the so-called Solow growth model.

Second, I use population growth to capture both growth trends as well as for example the impact of changes in savings and risk appetite as a consequence of changes in demographic trends.

Third, even though the Fed’s task is to ensure ‘price stability’ it has not always succeed in that. The opposite of price stability is price volatility and we therefore use a 10-year moving average of yearly changes in inflation to capture the degree of price (in)stability.

Finally, we also want to capture changes in investors’ and consumers’ optimism and pessimism and the general ‘business cycle’ and the general level of economic activity. I could have used unemployment or the output gap, but I also wanted to capture structural demographic trends so I have used the so-called employment-population ratio as an explanatory variable.

Estimation results

I don’t want to bother the readers too much with the specific statistical tests, but what I have done is simply to run a simple linear regression explaining the (nominal) fed funds rate with the four variables discussed above. The estimation period is from 1962 and until today. I use quarterly data.   

The graph below shows the actual Fed Funds rate and the predicted Fed Funds rate and we see the fit over nearly six decades is quite good (R2=0.8).

Furthermore, all four explanatory variables are strongly statistically significant and the sign on all four variables are as expected positive – meaning growth nominal GDP and population growth push up the Fed Funds rate as do higher inflation volatility and a high employment-population ratio.

Furthermore, it is notable that the model well captures the major trends in interest rates over the past six decades – higher interest rates in the 1970s and the decline in rates from the mid-80s and the ‘flat lining’ of rates over the 12-15 years.

I will not go in to too much detail in terms of the decomposition of what drives interest rates during the different periods and I will return to that in later work, but it is sufficient to say that the very low level of interest rates we have seen in the US since 2008 is more or less fully explained by the four explanatory factors – lower potential NGDP/capita growth, low population growth, low inflation volatility and a moderately low level of employment-population ratio.

Hence, the Fed funds rate is not low because the Fed’s has “manipulated” interest rates and is not low because monetary policy is has been overly easy (if anything it has been too tight – at least until recently).

Rather, low interest rates in the past decade reflect economic conditions and mostly structural economic conditions – for example weak population growth and a declining labour force due to the aging of the US population (the baby boomers have been leaving the labour market).

Looking ahead – Fed will need to hike sooner rather than later

In response to the ‘lockdown’ crisis of 2020 the Fed has cut the Fed funds rate (and implemented quantitative easing). This is completely in line with what the prediction of the model have been.

In fact the model predicted that the Fed funds rate should have been cut to -3% (!) in the second quarter of 2020.

This of course didn’t happen due to Zero Lower Bound on interest rates, but the Fed instead has opted to undertake significant quantitative easing of monetary policy. Hence, the money base has been expanded significantly. This in mind my undoubtably has been the right policy – at least in terms of the magnitude of monetary easing.

The predicted decline in the Fed funds rate primarily reflected a sharp drop in the employment-population ratio.

However, both of these variables are now clearly improving, but despite of that the Fed has so far insisted that we are far away from any rate hikes.

However, this week we have nonetheless seen quite a bit of a repricing of the markets’ expectations for rate hikes.

To answer the question about the outlook for rates we can try to make a simulation based on the model we have estimated above.

What we will try to estimate is where the Fed Funds rate should be in ‘steady state’ – hence, once the US economy has moved out of the crisis and returned to the underlining trends that we saw prior to the ‘lockdown shock’ hit in 2020.

Lets have a look at the four explanatory variables.

First we see that potential nominal GDP growth (per capita) as calculated by the Congressional Budget Office (CBO) over the past decade has been stable around 3.0%  – considerably lower than earlier decades. Hence, we use 3.0% as our measure of the long-term growth rate of potential NGDP growth (per capita).

Second, in terms of population growth we use a forecast of 0.7% annual population growth in the US – reflecting more or less most forecasts for US population growth over the coming decade.

Third, we assume that inflation volatility will remain low around the same level as we have seen over the past decade (0.25).

Finally, we assume the decline in the employment-population ratio will continue to reverse – reflecting the closing of the output gap.

Judging from recent economic data this recovery will be quite fast. For simplicity we assume that the employment-population ratio will return 60% by mid-2022. This is not necessarily a forecast of what will happen, but rather an assumption we use when making a simulation for the Fed Funds rate for the medium-term.

Fed Funds rate is heading back to 1.5% sooner than you think

If we assume that these for variables over the coming 4-6 quarters return to their steady-state levels then we can simulate what we should expect to see in terms of the Fed Funds rate.

The graph below illustrates that.

According to the simulation the Fed Funds rate should return to positive territory in the middle of this year. This can be interpreted as the Fed will end QE and change its forward guidance in a more hawkish direction and in the second half of the year we should actually see the Fed start hiking the Fed Funds rate.

During 2022 the Fed Funds rate should approach 1.5%.

This is of course a much more aggressive rate hike path than presently signalled by the Fed and expected by the market (around 100bp above market pricing in by the end of 2022).

However, it does not reflect an extreme assumption regarding the economy.

Rather the only thing we here assume is a return to ‘normal’ over the next 6-8 quarters and given what we are seeing in terms of the macro data at the moment and the planned major fiscal easing this certainly does not seem to be an overly optimistic take on the economic outlook. In fact it might be too pessimistic.

Furthermore, this does not reflect the change in the Fed’s inflation target from a pure 2% inflation target to an average inflation target and it does not reflect any change in the risk premium on US debt despite a very significant expansion of US public debt.

So if anything there might be reason to speculate that the Fed could be forced to hike even faster than this simulation indicates.

What if the Fed fails to hike?

Fed officials continue to argue that there is no reason to hike rates any time soon. In fact Fed chairman Jerome Powell has argued that the Fed is not even thinking about thinking about hiking rates.

But what happens if the simulated path for the Fed Funds rate is right and the Fed despite of that sticks to the promise of not hiking rates?

Then the Fed effectively will be in a situation where it will try to keep r (significantly) below r*.

The Fed has tried such policies before in the later part of 1960s and going into the 1970s and the result was rather catastrophic – inflation rose sharply. As did economic volatility.  

I certainly don’t think that the Fed will repeat the mistakes of the 1970s – at least not to the same extent – but while the Fed mostly have erred on the downside (too tight monetary policy) over the past 12 years the opposite now seems to be the case.

Hence, during the next 2-3 quarters we will likely have a test of the Fed’s commitment to it’s 2% (average) inflation target. Will the Fed for example allow inflation expectation to increase above 3%?

Or – more likely – will the Fed reverse course once it realizes that the output gap will be closed already during 2021 and inflation risks will increase further if it refuse to hike rates.

Unfortunately, the Fed has continued the mistakes of the past and communicated its monetary policy stance in terms of its policy instruments rather than in terms of communicating in terms of its policy goals.

This means that we in 2021 are likely to see the Fed being forced to change its message quite a bit. I am tempted to say that this smells of the stop-go policies of the 1970s and even though the outcome is likely to be less catastrophic than during the 1970s we could very likely see an increase in financial market volatility as a result of this.

PS there are also some elements of what we saw during 2006-7 in what is happening in financial markets, but I do not expect the same outcome as in 2008-9, but I might return to that in a later post.

PPS I have done similar modelling on the US 10 year Treasury bond yield. See here.

A model for US bond yields – the Twitter version

I have been tweeting…once again.

Tweeting on Central Bank Digital Cash

As part of my on-going research at Copenhagen Business School on the future of the Nordic banking sector and monetary policy have been researching the idea of Central Bank Digital Currency or rather as I prefer to Central Bank Digital CASH (CBDC).

Tonight I have been tweeting on some ideas and views regarding CBDCs and as with private cryptocurrencies it is not really the technology, which interests me, but rather the monetary aspects of CBDCs and what the importance for the conduct of monetary policy and the monetary transmission mechanism would be. And of course the impact on the wider economy and on the banking sector in particular.

You see my Twitter thread on the topic here (including a lot of typos…).

The clash of two Chicago school ideas

The Economist magazine quotes me on the important topic of whether we are going to see a spike in inflation or not, but at the same time also illustrates what market montarism really is about:

In the aggregate, though, investors seem unconvinced. The inflation expectations which can be derived from prices in financial markets have recently picked up a little thanks to the good news on vaccines and the prospects for a rebound in the world economy. But they still suggest that investors think next year’s inflation is more likely to be below the 2% central banks target than above it (see chart 3).

Lars Christensen, a Danish economist, points out that this means there is a “clash” between the two best-known economic theories associated with the Chicago school. Milton Friedman said sustained growth in the money supply leads to inflation; Eugene Fama argued that market prices fully reflect all available information. “If you believe that we are going to have inflation now…the efficient-markets hypothesis would have to be wrong,” Mr Christensen argues.

Source: The Economist

I discussed the same topic at a recent presentation at Buckingham University. You can watch that presentation here.

Covid-policies should focus on ‘health fundamentals’

The number of Covid-deaths per capita is converging towards a level which essentially is determined by what I have called ‘health fundamentals’ – or simply X*.

Europe and the US have different health fundamentals – Europe is ‘older’ and the US is more obese. Over all I would actually expect more deaths in the US mostly because of obesity, but it is clear that Europe is caching up fast now – X (actually mortality) is moving fast towards X*. Despite new restrictions being put in place everywhere.

Source: https://91-divoc.com/pages/covid-visualization/

The effect of these restrictions (as well as voluntary behviourial changes) might be to ‘postpone’ the ‘convergence’ towards what health fundamentals ‘dictate’, but not for long.

It is comparable to what economists call the vertical Phillips curve. We can use monetary and fiscal policy and other policy measures to push unemployment below the ‘natural’ rate of unemployment which is given by the structures in the labour market – for example – unemployment benefits, unionisation rates, taxes, minimum wages etc. but we can not maintain unemployment below this level for long and sooner or later unemployment will return to its natural level. Furthermore, we know there are strong negative side effects (accelerating inflation) from doing that.

It is the same with covid-deaths. We can artificially maintain mortality rates lower than what is given by health fundamentals for a short period, but at a high cost socially and economically and sooner or later mortality will return to what is given by health fundamentals.

The policy recommendation therefore should be to forget about these restrictions all together and instead focus on health fundamentals. Improving health fundamentals of course includes treatment including vaccines and for example vitamin D supplement, but also protecting the groups most at risk – e.g. the elderly and the obese. Furthermore, we also have to remember that X* is strongly seasonal in nature.

Finally, in the same way we in economic policy should focus on improving labour market structures and to the extent we want to ‘lean-against-the-wind’ with certain restrictions it should be rule-based and transparent restrictions with a proven evidence based track record (there are very few examples of that). Ideally the easing and tightening of restrictions should be decided in a transparent way by an independent ‘health council’ of experts rather than by politicians, but again most importantly we should focus on X* rather than on X.

Recovery and inflation scenarios in the USA for 2021

Swedish mortality in 2020 – It’s a lot better than you might think

Sweden has been the ‘outlier’ in terms of handling the global Covid-19 pandemic as the country’s health authorities have relied on a more laissez faire approach which have relied on the common sense of the Swedish population rather than on draconian government measures such as lockdowns and mask mandates.

In that sense Sweden has been different than basically every other country in Europe and Northern America.

Consequently, Sweden has also become the benchmark case to compare other countries to.

Unfortunately from day one of this pandemic it has all been about counting the number of people who have died from or with Covid-19. The countries with the least Covid-deaths are the “winners” – at least according to the media, commentators and politicians.

I must say I have long ago come to the conclusion that it makes very little sense making such comparisons without taking what I have termed ‘health fundamentals’ into account.

We for example now well know that the elderly and the obese are much more at risk from dying of Covid-19 than the young and healthy. Therefore, it makes little sense for example to compare a country with a lot of old people like Italy or the UK with a country with a young population like Pakistan. Or an obese nation like the US with a much more fit nation like Japan.

Hence, in my view the outcome in terms of Covid-deaths is mostly explained by these health fundamentals rather than by policies, behaviour or culture. These later factors may have an impact in the very short-term (weeks or months) but over months they are much less important than health fundamentals.

But one thing is to look at Covid-deaths and compare them across countries, but what about total mortality? After all the important thing really isn’t Covid-19. The important thing is total mortality (and age-adjusted mortality).

Thinking about this got me to look at Swedish mortality over the last couple of years. In this blog post I will share my main conclusions from looking at the numbers.

I have looked at daily deaths in Sweden from 2015 and until today.

Daily deaths spiked in April-May 2020

Lets start with the clean numbers for daily deaths for each of the years since 2015.

The years from 2015-19 are different shades of grey while 2020 is is red.

The first thing we notice obviously is that starting from around mid-March 2020 daily death numbers started to rise fairly sharply. In the same periode in the previous years the ‘normal’ seasonal pattern had been a gradual decline in daily deaths.

So there is no doubt that the Covid-19 pandemic clearly is visible in the numbers. Furthermore, this period of “excess deaths” lasted until around the beginning of June.

However, what we also see if we take a closer look is that prior to the pandemic had hit in mid-March the number of deaths had actually been rather low compared the previous years and similarly from around mid-May deaths have again dropped below the average number of deaths in prior years.

This is an indication that there was a significant amount of fragile elders who had survived longer than normally would had expected.

Therefore, some of the excess deaths during the March-May period might therefore be explained by “too few” deaths during January and February compared to what we would have expected.

In fact if we look at 2019 we see that year had somewhat lower general mortality in Sweden normally – further supporting what my fellow Danish economist friends Christian Bjørnskov and Jonas Herby have called the ‘dry tinder’-effect.

In the graph below we see this fairly clearly. Looking at the last decade we see a pretty strong ‘dry tinder’-effect – if the number of deaths is high one year (eg 2012) then change in the number of deaths will likely be negative the following year (eg 2013).

Based on this simple historical relationship we should expect thee dry tinder-effect to have had around 2,300 deaths to the total number of deaths in 2020 due to the ‘low base’ in 2019.

However, even if we ignore the ‘dry tinder’-effect the mortality rate in Sweden during 2020 has been a lot less dramatic than some pundits would have you believe.

We can see that by simply looking at the total number of people who has died in Sweden so far this year.

We see that during March-May there was a considerable excess mortality so by the end of May total deaths was more than 10% above normal levels.

However, since then we have see a gradual return towards ‘normal’ and presently the total number of deaths if around 4% above normal.

In total numbers this means around 3,000 more Swedes have died in 2020 than ‘normally’, which is less than half of the number of people who has officially died with/from Covid-19.

This to me is a strong indication that the Covid-19 pandemic more than anything has moved forward deaths by weeks or months rather than by years and it is an indication that Covid-19 to a considerable extent has ‘replaced’ other ‘normal’ causes of dead among the old such the flu or

In fact if the trend from recent months continue during the next couple of months we might we the entire excess mortality for the ‘pandemic year March 2020-March 2021’ disappear.

So yes, a deadly pandemic hit Sweden in 2020, but if one steps back a bit and look at the total number of deaths for the entire year it to see the pandemic.

A way to illustrate that is to forecast how many deaths in total we will see in 2020.

I have done that by assuming that we will see the numbers of deaths in the reminder of the year as normal then we are likely to see around 94,000 deaths in total for all of 2020.

That will mostly likely make it the most deadly year over the last decade but it will none the less be fairly close to the number of deaths in 2012, 2017 and 2018. Sweden was as fairly hard hit like other European countries by influenza pandemics during these years.

Mortality is higher in Denmark than in Sweden

The number of deaths with/from Covid-19 in Sweden has been considerably higher than the other Nordic countries, but what about total mortality.

As a Dane (of Swedish descent) I of course can’t help comparing Swedish and Danish mortality.

The graph below shows the number of daily deaths in Denmark and Sweden adjusted for population size.

It is clear that Sweden was much harder hit by the pandemic in March-May than Denmark.

However, it is also clear that Swedish mortality was considerably lower both period to the pandemic and from May and onwards.

In fact if we look at the total number of deaths in Denmark and Sweden more Danes have died than Swedes adjusted for population size.

The reason for this simply is that the ‘normal’ mortality rate is higher in Denmark than in Sweden. Or said, in another way – a lot more Danes die from cancer than Swedes die from Covid-19.

Mortality has been high, but certainly not catastrophic

Every death is tragic and Sweden has certainly been hard hit by Covid-19. However, when we take a closer look at Swedish overall mortality in 2020 it is also clear that 2020 hardly is the kind of disastor that some pundits would like it into being.

Covid-19 is not “just a flu”. It is certainly more deathly for particularly the elderly. However, in terms of the impact on total Swedish mortality it has more or less been on a comparable level to the ‘bad flu-years’ 2012, 2017 and 2018.

Furthermore, while Sweden has been hard hit by Covid-19 the overall mortality rate is still lower than in neighboring Denmark, where Covid-19 hardly is visible in the mortality rate.

All data in this blog post is from the Swedish statistics office (SCB) and from Statistics Denmark.


Mail: lacsen@gmail.com

Phone: +45 52 50 25 06.

%d bloggers like this: