Friday’s US labour market report rightly got a lot of media attention globally. The spike in US unemployment to 15% surely is historical and tells us quite a bit about just how big a shock has hit the US and the global economy. However, where most commentators are wrong is assuming that this has to be seen as a normal recession. I on the other hand would argue that this has little to do with a normal recession. In fact I am increasingly thinking that the use of the term ‘recession’ is a misnomer in relation to this crisis. Back in April I argued in my blog post ‘All set for a fast recovery after the ‘Great Lockdown’ argued that this crisis primarily should be seen as an unplanned and very unpleasant ‘vacation’. The IMF has called it the ‘Great Lockdown’ and I find this term very
Lars Christensen considers the following as important: Uncategorized
This could be interesting, too:
L. C. writes The Corona Crisis – a Scandinavian perspective.
Reinhilde Veugelers and Georg Zachmann writes Racing against COVID-19: a vaccines strategy for Europe
Friday’s US labour market report rightly got a lot of media attention globally. The spike in US unemployment to 15% surely is historical and tells us quite a bit about just how big a shock has hit the US and the global economy.
However, where most commentators are wrong is assuming that this has to be seen as a normal recession. I on the other hand would argue that this has little to do with a normal recession. In fact I am increasingly thinking that the use of the term ‘recession’ is a misnomer in relation to this crisis.
Back in April I argued in my blog post ‘All set for a fast recovery after the ‘Great Lockdown’ argued that this crisis primarily should be seen as an unplanned and very unpleasant ‘vacation’.
The IMF has called it the ‘Great Lockdown’ and I find this term very telling. Economies around the world have been locked down – either by governments or by voluntary behaviour by people who want to protect themselves against the coronavirus.
Most people don’t really think about it, but most industrialised economies in the world every year goes through large “recessions” in the form of a major drop in economic activity. This happens both on the supply side as for example Southern Europeans go on Summer vacation typically in August or with private consumption as it fluctuates wildly before and after Christmas.
We don’t see articles about this in the financial media and the reason is of course that we know this is happening. It happens every year. So nobody cares – we plan for it so it isn’t an economic problem.
The lockdowns – boths by government intervention and voluntary social distancing – wasn’t preplanned and that’s why it has created a major economic disruption.
But that said, if we instead of sensationalising it, take a look in our economic textbooks and look at economic history then we will realise that economies return to ‘normal’ very fast after ‘vacations’ (also very unpleasant vacations).
The reason demand shocks take long to ‘disappear’ is that prices and wages are sticky and that economic policy reacts too slowly or insufficiently. But that is not the problem with supply shocks – they very rarely have longer-term effects.
In fact that was the problem with the entire idea in the so-called Real Business Cycle (RBC) models that became popular (to discuss) in the early 1990s – it simply was impossible to show empirically that supply shocks would have very long-lasting effects on economic activity and certainly no long-lasting impact on labour markets.
So when I in the headline argues that US unemployment will be back below 6% in November then it is simply because that is what the economic textbook tells us – market economies adjust fast to supply shocks.
This is essentially the point I also was making in my blog post in April, but I am happy to repeat it and I haven’t become more negative since then. The labour report, while it was horrible wasn’t the least surprising.
We have known for weeks that US unemployment would spike to these levels so I really haven’t become more worried about my forecast for a sharp recovery in economic activity in the US towards the end of the year.
In this blog post I will present three more arguments – other than the purely theoretical arguments I have just repeated – why I believe that unemployment will soon be down to a level close to before the lockdown-shock hit.
The markets is telling us so
When I back in 2011 coined the term ‘market monetarism’ is was because (now self-declared) market monetarists like Scott Sumner and David Beckworth and myself believe that financial markets tend to be efficient and hence reflect all available information about the outlook for the economy and that markets therefore will be the best available ‘forecast’ for the outlook for the economy and that policy makers should utilize this information when they conduct policy.
This also means that market monetarist economists to a much larger extent than more traditional macroeconomists tend to look at financial markets when they try to forecast what will happen in the economy going forward.
And if we look at what markets have been telling us since the second half of March it is that this is not a demand shock. Market inflation expectations have first rebounded and then stabilised after the initial first shock and due to the actions of the Federal Reserve.
And the stock market is telling us the same story. The US stock market is not back at the levels we saw in late February but we have made a significant recovery as markets got better visibility about the outlook for the spread of the coronavirus and the Fed demonstrated that it would not allow a new debt-deflation spiral to set-in.
To me the US stock market, while sometimes wrong, is a fairly reliable indicator of future growth in US nominal income (NGDP) and normally fluctuations in NGDP growth cause fluctuations in US unemployment so we should expect the stock market to be a fairly good indicator of the US labour market as well.
We can illustrate this with a simple xy-graph with changes (%y/y) in US stock prices (Wilshire 500) versus percentage-point changes (y/y) in the US unemployment rate lagged six months.
As we see there has historically been a fairly strong inverse relationship between the development in the US stock market and in US unemployment (over the coming 6 months). In the graph I use data going back to 1982.
It is not hard to spot the outlier – the US stock market “failed” to predict the sharp rise in US unemployment we saw in April.
One conclusion of course could be that the market is just plain wrong and the situation is much worse than the collective wisdom of the market says it is.
This essentially what a lot of commentators are saying at the moment – the market has become way too optimistic and it is all driven by “Fed liquidity”.
Well, I trust the market where investors have money on the line rather than pundits.
So what is that market telling us?
Presently the broad US stock market is down around 2% over the past year and given the historical statistical relationship that would (based on a simple linear regression) imply that US unemployment should be up around 0.5%-point in November (compared to November 2019). This would imply US unemployment at 4% in November – somewhat lower than 6%.
Obviously this is meant as an illustration rather than an actual forecast, but the overall story is nonetheless that judging from the US stock market the sharp increase in unemployment we saw in March and April is going to be temporary and unemployment will soon be back to normal levels.
Hence, the markets are presently pricing that this will not be a long-last economic downturn and hence the increase in unemployment will be temporary and this naturally brings us to the next topic – most of the increase in unemployment is driven by people who have been laid off temporarily.
Most layoffs are temporary
Historically around 10% of US unemployment has been made up by workers temporarily laid off from their jobs.
However, if we look at the US unemployment in April 78% of all unemployed had the status of being temporarily laid off.
This means that essentially the entire increase in US unemployment in April was due to temporary lay-offs. We see that in the graph below.
What I here call ‘core unemployment’ is unemployment minus the unemployed who has been temporarily laid-off.
We see here that ‘core’ unemployment actually has been more or less unchanged over the last couple of months (also in April) around 3.2%
This contrast sharply with the recessions of 2001 and 2008-9 where the number of temporarily laid-off workers didn’t increase at all, but core unemployment rose sharply. Both the recessions of 2001 and 2008-9 were recessions caused by demand shocks.
If we, however, go back to the recessions of 1973-74, 1979-80 and 1990 we see that the share of temporary laid-off workers increased initially during all of these three recessions.
What did all of these three recessions have in common? They initially all were triggered by a sharp increase in oil prices – hence, a negative supply shock.
While an oil price shock is something very different from a lockdown both are nonetheless negative shocks to the production (supply) side of the economy.
This basically means that workers are being laid-off not because they have become too “expensive” relative to output prices (that is what is happening with a demand shock), but because the cost of OTHER inputs have increased. In the case of lockdowns simply because production basically is outlawed in certain sectors.
Hence, a supply shock is not about prices and wage rigidity as I discussed above and this means that workers have not been ‘priced out of the market’ and should therefore be expected to return to work once production gets up and running again.
During the first oil crisis the spike in ‘temporary unemployed’ lasted a bit more than a year while it took only around 9-10 months to get back to ‘normal’ during the second oil crisis.
However, it should be noted that both during the first and the second oil crisis monetary conditions were tightened in response to rising headline (supply-side) inflation, which caused nominal demand growth to slow. Said in another way, on top of the supply shocks we got a negative demand shock.
This isn’t the case this time around – the Fed has responded to the crisis by moving to ease monetary conditions and even though market inflation expectations are too low (lower the Fed’s 2% inflation target) the Fed nonetheless has stabilized inflation expectations
This means that once the lockdowns come to an end people will be able to return to work – not necessarily to their old jobs and not necessarily in the sectors they used to work in, but the reason they haven’t been working is not that their reservation wage were higher than their productivity so there is little reason why we shouldn’t see the share of temporarily unemployed come down very fast in the coming few months.
Another illustration of this is to look at initial jobless claims in different US states.
I have looked at some of the hardest hit states where there also have been fairly strict lockdowns and States which have been less hard hit and also a State where there hasn’t been a lockdown (Utah).
For a comparison I have also included the state of Louisiana, but with the data from 2005 when the state was hard hit by Hurricane Katrina.
When we look at the data we see a very different pattern than what we would see normally during a recession where initially jobless claims keep rising months. In fact during the recession in 2008-9 initial jobless claims rose for more than a year.
This time it is different and as we see the initial jobless claims numbers now behave much more like a shock like Hurricane Katrina in 2005.
In fact so far the pattern has been very similar for most States – an initial sharp (very sharp) increase in jobless numbers for a couple of weeks followed by a relative sharp drop in initial jobless numbers thereafter.
So far the numbers more or less have tracked the ‘Katrina pattern’ and if that continues we should expect ‘claims’ to be back to normal levels by the end of June. At that time we should already have seen US unemployment numbers having started to come down significantly.
Consumption will rebound sharply – the money is there
I have for some time when talking to the media or with clients been making the argument that this crisis primarily is a supply shock and as long as the Fed (and other central banks) are doing their job of ensuring nominal stability when the economy should rebound very fast once we around the world move out of the ‘lockdown’.
However, most people (even many economists) tend to have a rather rudimentary perspective on economics and what they observe is that since private consumption is down and therefore (they believe) should GDP be – whether nominal or real doesn’t seem to matter.
So even though I don’t really think it is important how the GDP‘cake’ is sliced in terms of aggregate nominal demand I will nonetheless try to address the issues of private consumption.
Basically I believe that it makes most sense to think of private consumption on a macro level within Milton Friedman’s permanent income hypothesis.
Over time private consumption is determined by permanent income expectations. So if our expectations about further permanent income decline we will tend to spend less.
Obviously on a micro-level one can have all kinds of reservations about Friedman’s permanent income hypothesis, but I still think it is the best we got.
That essentially also means that we can think of how much each of us spends in a given week or month as reflecting the ‘targeted private consumption’ we have, which in turn reflects our expectations about permanent income.
However, we do also change our spending depending on a lot of other things – we spend more when we are on vacations and during weekends. The same goes for holidays like Christmas. And then we spend less during other periods.
From time to time we don’t hit our ‘targeted’ consumption. For example if you have planned to buy a new laptop, but it turned out that the model you wanted was sold out and you now have to wait another month. Or the opposite happens – you find something you have been looking for on eBay and simply has to buy it – even though that increases your spending above your targeted spending.
With this ‘model’ in mind I think we can understand what has been happening with US private consumption, but also what will happen going forward.
The first question we need to ask is whether there has been a change in the outlook for US permanent income.
That is essentially whether there has been a drop in the US long-term growth potential. One can of course say that this pandemic will cause all kinds of frictions both as a result of changed behavior and changes in regulation. However, I think it is very hard to say now, but I would also stress that I think commentators in general make far too wild predictions about just how much this really is going to change things.
Right now it is much easier to say “everything will change – nothing will be the same” in the global ‘media contest’, while the most likely scenario that we will gradually adapt as humanity always does and that this is likely to have an insignificant long-term effect on global growth will not get you either on CNN or Fox TV these days.
Furthermore, judging from what the financial markets are telling us there is no reason that we should have become significantly more worried about long-term growth in the US or globally because of this pandemic.
This means that it is reasonable to assume that there really hasn’t been a change to permanent income expectations in the US.
But what we have seen is a shock to private consumption, but not because people have become overly worried about their future income in general (some clearly have), but because people simply haven’t been able to spend. Your favourite restaurant has been closed and so has your hairdresser.
So what has been happening is that your actual private consumption has been lower than your ‘targeted consumption’.
A way of illustrating this is by looking at consumption and bank deposits. If private consumption develops as planned then we should expect deposits to grow steadily (with income) over time. However, if there is a shock to private consumption then this should be reflected by a similar shock to bank deposits.
The graph below shows that the lockdowns and the behavioral reaction to the perception of the risks of the coronavirus in the US have had exactly this effect.
We see that prior to the shock bank deposits and private consumption expenditure was growing much in line with a strong positive correlation. However, as the ‘lockdown shock’ hit consumption dropped like a stone while deposits increased sharply.
It is particularly noteworthy that in dollar terms the increase from February to April bank deposits nearly is exactly the same amount as the decline in private consumption in March.
We don’t have the April numbers for private consumption yet but given the continued increase in bank deposits we should expect private consumption to have declined a further in April.
This to me is a pretty clear indication that the drop in US private consumption does not primarily reflect worries about the future permanent income of US consumers or a negative shock to income itself (then deposits would have dropped and not increased). But rather this is simply a reflection of the fact that US consumers have been taking temporary ‘vacation’ from spending.
The question of course is when will consumers start to spend again?
Here history might help us. There are not a lot of examples in US modern economic history of such ‘unplanned spending vacations’, but one example is very similar and that is 911.
When terror hit the US on September 11 2001 the US economy in many ways also came under a ‘lockdown’ and Americans stopped spending from one day to another.
It was a shock many at the time said American consumers would never recover from – in the same way many today say that it will take a very long time to return to ‘normal’ consumption patterns.
The stories at the time were the same as today – people will never fly again, they will not go to restaurants, they will never go to a basketball game again etc.
We today know that the shock didn’t have a very long-lasting impact on US consumers.
The graph below shows the consumption-deposit-shock of 2001.
We see the 911-consumption-deposit-shock is quite similar to what we are seeing now. 911 caused a ‘spending lockdown’ in September-October 2011, which in turn caused bank deposits to increase in parallel.
However, the cut in spending had not been planned and consumers had not changed their permanent income expectations and consequently consumers quickly got consumption and deposits back on their ‘targeted’ levels – indicated by the dotted lines.
In fact, nearly to the dollar the amount US consumers ‘under-spend’ in September 2001 they ‘over-spend’ in October 2001. And they had the money in their accounts to do it.
After having made up for ‘lost’ consumption consumers got back on track in November and continued on the pre-911 spending path.
Those of us who still in horror remember the terror attacks on that horrible day in September 2001 also remember the fear of flying and the fear of just going out and about. However, life returned and so did consumption – in less than a month.
I believe this is an important lesson for those who think that ‘we will never be back to how it was before’ in terms of consumption.
If economic theory (the permanent income hypothesis) and economic history (911) teaches us anything it is that we should expect US consumption to make a very swift recovery.
In fact there is no reason not to believe that private consumption expenditure will be back on track in July after likely overshooting in June as US consumers catch up on what they have ‘lost’ in terms of spending since February.
Another reason to be optimistic is what we are now seeing in terms of private consumption in the Nordic countries.
Private consumption has followed a very similar pattern as US consumption in March and April in all of the Nordic countries. We are, however, now beginning to get out of lockdown.
In Denmark schools and kindergartens have been (partly) open since Easter and over the last two weeks certain shops that were closed (by government regulation) during the lockdown have reopened – for example hairdressers.
My former colleagues at Danske Bank publishes a weekly “Spending Monitor”, which is based on among other things the bank’s clients’ credit card and cellphone payments. The Spending Monitor gives a near-real time update on Danish private consumption.
It will be very interesting to follow in the coming weeks as Denmark opens up more and more for business and as we return to a more normal life. The ‘re-emergence’ has been under way over the past month.
A very notable graph in the latest edition of Danske’s “Spending Monitor” is this graph with the turnover at hairdressers.
I particularly note that not only have Danes returned to the hairdressers – they are also making up for lost ground (and long hair) and spending around 20-30% more at the hairdresser than a year ago.
I believe we will see something very similar overall in the US as the US economy also re-emerges from lockdown in the coming weeks and I see no reason why private consumption shouldn’t recover very fast.
Yes, US unemployment will drop below 6% by November
So in conclusion, I think that despite the tragedy of the Covid-19 epidemic there is no reason to believe that the US economy – and the global economy for that matter – shouldn’t recover quite fast from this crisis. Much faster than after the 2008-9 crisis.
Numerous policy mistakes have been made around the world both in combating and containing the pandemic and in terms of the monetary and fiscal response to the crisis and more mistakes are likely to be made, but we should nonetheless remember that market economies emerge much faster from negative supply shocks than from demand shocks.
That is what I have tried to argue is this blog post and finally let me repeat my forecast – I strongly believe that US unemployment will drop very fast in the coming months and will likely have dropped below 6% when US voters vote at the US presidential elections in November.
Lars Christensen, [email protected], +45 52 50 25 06.