Recession or not? Answering a question the Fed should not care about

A rather bizarre discussion about the lastest US real GDP numbers has been playing out since the 2nd quarter GDP numbers came out on Wednesday.

The official real GDP numbers showed that US GDP dropped 0.2 percent (0.9 percent annualised) in Q2 making it the second quarter in a row with negative real GDP.

This got pundits to scream “Recession!” and yes this is surely a heated political topic (of less real economic relevance).

And yes, one definition (on US data but not data around the world) says that the US economy is in a (technical) recession when we have two consecutive quarters of negative (Q/Q) real GDP growth.

The “official” – rather than the technical – definition of a recession (again only in the US and nowhere else) is made by National Bureau of Economic Research (NBER) who defines a recession in the following way:

“… a recession is that it is a significant decline in economic activity that is spread across the economy and that lasts more than a few months. The committee’s view is that while each of the three criteria—depth, diffusion, and duration—needs to be met individually to some degree, extreme conditions revealed by one criterion may partially offset weaker indications from another. “

This is two very different definitions, but historically there has nonetheless been a very close relationship between the two – and since the World War II all “NBER recessions” have coincided (more or less) with periods where we have had two quarters of negative growth in a row.

However, this is not the case this time around – or we really won’t know for some time as the NBER hasn’t announced anything in this regard and normally only do after the fact, but it is also clear that this doesn’t “feel” like a recession.

The reason it doesn’t feel like a recession is that there certainly is no general glut – there are no signs of demand (in nominal domestic spending continue to growth strongly) declining unexpectedly and unemployment is close to historically low levels and have continued to drop in the Q1 and Q2.

What feels bad is the very high and increasing inflation. And that certainly feels bad, but economic “uneasiness” and uncertainty is not the same thing as a recession.

A way to illustrate this paradox is to look at the relationship between real GDP growth and (quarterly) changes in US unemployment – what is sometimes called Okun’s Law.

We see a quite clear (negative) correlation – slower real GDP is normally associated with higher unemployment.

We have two extreme observations – one negative and one positive. Both are from 2020 and is not really relevant to what we are trying to analyse here.

What is important is what is happening in the ‘red box’ – what happens to unemployment when real GDP growth is negative.

The graph below zooms in on the red box.

What we see is that the ‘normal’ relationship is that in quarters where we have negative GDP growth (all of the dots) we have higher unemployment (32 out of 41 quarters).

We also see that in the quarters where unemployment declined (the 9 red dots) real GDP growth dropped by less than 1 percent (annualised).

But ONE quarter of negative real GDP growth is not a (technical) recession. A technical recession is defined by two consecutive quarters of negative real GDP growth.

So let’s look at that – in the graph below I have looked all of the technical recessions since 1948 (the period where we have both real GDP and unemployment numbers).

In that period there has been 11 technical recessions – including the present one.

Some have been longer than two quarters (e.g. the 2008-9 recession) so I have looked at the average quarterly change in GDP and the same for unemployment during the ‘recession period’.

The graph below shows relationship.

Again the extreme observation is 2020 and again that really isn’t important. The important thing is that for all of the 11 technical recessions with the exception of one we see that not only is there a close correlation between the change in real GDP and the change in unemployment, but it is also so that negative real GDP is associated with an increase in the unemployment rate.

There is only ONE exception – and that is the present technical recession.

What we also see is that this in fact is the mildest ‘recession’ on record (at least so far). So mild that unemployment hasn’t even increased.

Finally if we run a regression on the data we see that there is more or less a one-to-one relationship between real GDP growth and changes in unemployment.

Overall, it seems that the present ‘technical’ recession is somewhat of an outlier as it is clearly very unusual to have a recession where unemployment actually has been declining.

Don’t ever forget NGDP

My friend Scott Sumner famously (at least among market monetarists) in 2009 wrote an influential article with the title The Real problem was Nominal about the causes of the Great Recession making the argument that policy makers (particularly central bankers) should focus on nominal GDP rather than real GDP.

Scott’s argument – as is mine – that central banks control nominal variable and not real variables, but since prices and wages are sticky nominal shocks will in the short-run have an impact on the real variables – such as real GDP growth.

In this regard it is interesting (depressing in fact) that when the Q2 GDP numbers was published there essentially was nobody reporting on what happened to nominal GDP growth.

But not reporting and not discussion the NGDP numbers mean that we are missing a very important point about the US economy – and that the real GDP numbers are be misinterpreted.

If we look at the NGDP we see a complete different picture that what we get from the RGDP numbers – there was no contraction in Q2 at least not a nominal contraction.

As the graph above shows nominal GDP growth in fact accelerated in Q2 to an annualised rate of growth of 7.8 percent – up from 6.6 percent in Q1.

Hence, measured on NGDP growth alone US monetary policy remains excessively easy – in fact very easy.

Assuming that potential real GDP growth in the US is around 2 percent then nominal GDP shouldn’t be growing much faster than 4 percent to ensure the Federal Reserve 2 percent target.

This is contrary to the general interpretation where many commentators have noted that the negative real GDP growth was caused by Fed’s monetary tightening. However, this can’t be correct – if that had been the case then we would have seen a slowdown in both real and nominal GDP growth. Or to be precise – the impact of monetary tightening actually happened from Q4 last year and until Q1, but not from Q1 to Q2. This is by the way fully consistent with the fact that monetary policy often works with long and variable leads as markets react to expectations of changes in monetary policy instruments ahead of it actually happening.

It is also correct that US recessions historically has been monetary in nature – in the sense that historically downturns in real GDP has be preceded by or coincided with a downturn in nominal GDP.

The below shows the de-trend quarterly growth in both real and nominal GDP. I have de-trend the quarterly growth rate with the average quarterly growth rate over the previous 10 years.

What we see is that there close relationship between the two – then NGDP moves up so does RGDP.

However, since early 2021 we have seen an unusual development – NGDP growth have remained high (very high), while RGDP growth have slowed back towards and now below the 10 year trend growth rate.

What the graph essentially shows is that normally (and we could go back much longer and show the same pattern) ups and downs in the US economy is essentially driven by ups and downs in nominal domestic spending or what we in the normal macroeconomic textbook call aggregate demand (AD).

When aggregate demand increases we know that both prices (P) and production (Y) increases. P*Y is exactly nominal GDP – so the graph above shows that we have had very, very strong growth in aggregate demand (obviously caused by the massive monetary policy easing undertaken in 2020-21).

If prices and wages are sticky then an increase in aggregate demand will cause an increase in production (Y). This is of course what we see when real GDP increases in parallel with an increase in nominal GDP.

In this is regard it is not surprising that real GDP also grew strongly in 2021. However, we also know from the textbook AS/AD model that we cannot just expand aggregate demand forever – soon or later we will run into supply side constrains.

The short-term supply curve (SAS) is upward-slopping, but as (input) prices and wages adjust to the increase in aggregate demand the SAS curve shifts leftwards . Hence, in the long run there is a hard constraint in the economy in the form of a vertical Long Run Aggregate Supply curve (LRAS).

I think this is part of the story of the slowdown in real GDP growth – we have simply hit the LRAS. So when commentators are talking about “supply chain distruptions” then it is actually this – it is suppliers (and employees) saying that they will not supply more at present (input) prices and wages.

The graph below illustrates this.

If we think of the 2010-19 trend (the yellow line) in real GDP (Y) as the the long run aggregate supply (LRAS) curve then we see that after the 2020 the gap between actual real GDP and the LRAS was more or less closed in the Autumn of 2021.

The US economy (and the global economy for that matter) hit the roof.

However, we also see there was no roof for NGDP (the blue line) – as NGDP rose well-above the the 2010-19 trend line.

We of course know from the textbook that there is no limits on how much NGDP can increase – there is a limit on how much Y (RGDP) can increase, but prices (P) can increaes without any limits and hence NGDP (P*Y) can increase without any limits.

Normally of course there is a limit in the form of the Fed’s inflation target – and one should expect the Fed to slow NGDP growth once inflation starts to outpace Fed’s 2% target.

However, we all know that the Fed during 2021 was sending a very clear signal that it considered the increase in inflation as “transitory” (it clear was not) and as a consequence NGDP was allowed to run wild. Or rather the Fed caused NGDP and therefore inflation to run wild by first massively increasing the money supply and then later by signalling it had no plans to reverse the massive easing.

The graph below shows what happened to the price level (P).

We see that initially in the crisis prices dropped below the 2010-19 trend (which more or less is a reflection of the 2 percent inflation target).

However, as the real GDP level closed in on LRAS, but NGDP kept on rising strongly towards the end of 2020 prices (the blue) started to accelerate strongly. Exactly as the textbook model would tells us.

There really isn’t anything surprising about this and it was avoidable that inflation would increase sharply with Fed ignoring the strong growth in NGDP. And this is why I in April 2021 forecasted double-digit US inflation – I didn’t need odd explanations such as “supply disruptions” or Putin’s invasion of Ukraine (which had not happened).

All I needed was the textbook AS/AD model and the equation of exchange, which tells us that NGDP is determined by the money supply (M) times the velocity of money (V).

So far so good – the story in 2021 very much was one of an increase in NGDP combined with the RGDP hitting the roof. And this obviously should start to slow real GDP. There shouldn’t be any surprise about that.

But there is more to the story – because what we are seeing now is NGDP continuing to grow robustly (at least compared to the historical trend) while RGDP growth has turned negative. At the same time we see inflation accelerating further.

The textbook AS/AD model tells us that the price level can increase both due to high aggregate demand (essentially higher NGDP growth) or a decline in aggregate supply caused by for example an increase in input prices (such as energy prices).

So just look at the prices level we cannot on its own say that that is caused by easy monetary policy (higher NGDP growth). Therefore, we need to look at real GDP growth (Y) as well.

So what we are seeing now is that in Q1 and Q2 real GDP has not only hit the 2010-19 trend, but it is actually decline relative to this trend and as we cannot observe a similar decline in nominal demand this has to be caused by a negative supply shock (the short run AS curve has shifted left).

This at least is the story if we can trust the data, but let’s start out by assuming that the published GDP numbers actually are correct.

But what have caused a negative supply shock? Well, it is hard to avoid the war in Ukraine and the impact this have had on particularly global energy prices.

The problem of course is that higher energy prices can also be caused by increased demand as well as a negative supply shock. For example energy prices also rose strongly in 2021, but that was clearly not caused by the war in Ukraine (as it hasn’t started) and real GDP grew strongly in 2021.

But let’s nonetheless try to figure out how much increased energy prices might have impacted real GDP growth in Q1 and Q2.

A simple model for RGDP

To answer this question I have estimated a simple model for quarterly real GDP growth with two explanatory variables – nominal GDP growth and the growth rate of US household expenditures on energy as a share of total expenditure.

I should stress this is purely illustrative model, but nonetheless it works very well in explaining the historical development in US real GDP growth (since 1960 and until the end of 2019).

The regression output is below.

We see that the signs on NGDP and “energy prices” (think of it as AD and AS respectively) have the expected signs – higher NGDP growth increases real GDP growth, while higher energy prices reduces real GDP growth.

We can now use this model to “simulated” how we think real GDP should have developed from early 2021 until today given the develpment in NGDP and energy prices.

The graph below shows that.

We see that model works quite well during 2021, but in 2022 a ‘gap’ between the model forecast and the actual real GDP growth opens up.

Hence, even taking the ‘energy shock’ into account we really can’t explain why real GDP have been negative in the US in Q1 and Q2.

However, this does not mean that that shock to energy prices is not large. It is and it certainly have had a significant negative impact on real GDP growth.

Based on the estimation results we can estimate the impact of higher energy prices on real GDP in the US.

The graph below shows that.

We see that already in 2021 higher energy prices lower real GDP growth, but both in Q1 and Q2 the energy price shock lowered US quarterly real GDP growth (annualised) by more than 1 percentage point in each quarter. That is a substantial negative shock.

In fact if we compare this shock to the first (1974-75) and the second (1979-80) energy shocks then this shock is of a similar size as the largest quarterly shocks during those shocks. The difference is, however, that those shock lasted longer than this shock has lasted so far.

We should, however, also note that the US today actually is much less dependent on energy imports than in the 1970s, which might tend to over-estimate the impact of higher energy prices on real GDP, but the impact nonetheless should not be ignored.

Conclusion: Both monetary policy and energy prices are slowing growth

We started out discussing the paradox that US real GDP has dropped two quarters in a row, but unemployment has continued to decline. This is highly unusual and it could indicate that something is ‘wrong’ with the data.

However, our analysis also shows that tighter monetary policy – lower NGDP growth historically will lower RGDP growth and NGDP growth even though it has remained high clearly has slow in 2022 compared to 2021.

Furthermore, high energy prices clearly is very significant negative supply shock to the US economy, but it is much less important than the tightening of monetary conditions (particularly from Q4-2021 to Q1-2022).

Overall, however, we are left with the conclusion that it shouldn’t be a surprise that US real GDP growth have slowed in 2022. However, it nonetheless seems very hard to explain the extent of the slowdown which could be an indication that there simply is something wrong with the data.

This is less controversial than it might sound like – national account numbers are often revised – and if we compare the US real GDP numbers with the similar numbers for the euro zone it is somewhat puzzling that there is no technical recession in euro zone where there certainly has been less easy monetary conditions since 2020 than the US and which arguably should expected to be lot harder hit by the war in Ukraine than the US.

I the graph below I have compared the US and the euro zone numbers.

Interestingly we see that our model for the US actually fits the European data better than the US data in 2022, which is a further indication that there is something wrong with the US data.

So again it is clear that US growth is slowing and this slowdown is caused by a combination of tighter monetary conditions and a negative supply shock, but the US economy is not yet in recession.

That doesn’t mean that the US will not fall into a recession. I believe it likely will has the Fed have to continue to tighten monetary conditions to get inflation back under control and in that process the Fed might have to accept a recession.

However, as the analysis above shows it becoming challenging for the Fed right now. On the one hand NGDP growth remains far too strong and even though NGDP growth is set to slow further in the coming quarters (I discuss that here) there is a risk that the Fed becomes preoccupied with real GDP growth, which as we have shown also is influenced by supply side factors, which it cannot control.

If the Fed where to conclude from the real GDP numbers alone then it would likely abandon monetary tightening right now and ease monetary policy instead and the calls for doing that are likely become louder in the near future.

This would be a repetition of the monetary policy mistakes of the 1970s where the Fed interpreted real shocks as being nominal shocks (or rather it was driven by political pressures and economic thinking that excluded real shocks and ignored monetary policy as a reason for inflation).

Therefore, even though whether or not the US is in fact is in a (technical?) recession or not apparently is hugely important for the political debate in the US the question actually should be of very little importance for the Fed.

What the Fed needs to focus on is the growth rate of nominal domestic spending (eg measured as NGDP growth).

The Fed should be happy to see that NGDP growth has been slowing (in fact significantly), but it should also note that NGDP should be brought down to around 4 percent to ensure that inflation stabilises around 2 percent. We are not there yet, but we are getting there – if the Fed stay focused on domestic spending growth rather than getting distracted by discussions about whether or not the US is in a recession.

Leave a comment

2 Comments

  1. Tom Clifford

     /  July 30, 2022

    Excellent analysis. Don’t you think the lack of growth of the US labor force is one of the things constraining AS?

    Reply
  2. Market Fiscalist

     /  July 31, 2022

    Great article that I am slowly trying to digest.

    Probably a simplistic question but I was wondering that as RGDP is NGDP minus inflation and that as current inflation is a bit of a new thing – is it possible that inflation is a little lower than used in the calculations and we actually do have GDP growth but are just not measuring inflation correctly ?

    Reply

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

%d bloggers like this: