A 121 x sigma event

What happened on the oil market this week was completely crazy.

I believe everybody noticed the headlines, but it is worthwhile to pause and reiterate, as this was another sign of the very unusual and severe recession we are going through: For the first time in history, oil prices were negative at minus USD 38 for a barrel of oil. I think this graph nicely illustrates how unusual this was.

Percentage daily changes in the WTI oil price.
Data source: Thomson Reuter Datastream via Eikon.

The graph shows daily changes in oil prices since 1983. Until April 16, the average daily price change is 0.03%. The standard deviation of daily price changes is 2.5%. On April 20, Monday, the price of oil fell by 306% compared to its price Friday. This is a 121-sigma fall. Almost surreal, but true.

The price of oil fell because demand for oil has tanked as a result of the recession. When economic activity falls, as it does right now, demand for oil falls. When we do not travel, when we do not drive our cars to work, when production facilities are closed down, and so on, we do not use oil. When supply of oil is not reduced, or at least not reduced as much as demand for oil falls, prices adjust.

If oil is produced but not used, it needs to be stored somewhere. Traders panicked on Monday, as they feared that there is simply not enough physical space to store oil. If you possess thousands of barrels of oil and have nowhere to store them, you are willing to pay somebody to take care of it. FT explains it nicely: “Analysts believe a lack of available storage capacity at the WTI contract’s delivery point of Cushing, Oklahoma — known as the Pipeline Crossroads of the World — set off panic among traders holding derivative contracts, who found themselves with nowhere to put the oil.”

The negative price of oil relates to West Texas Intermediate, the US benchmark oil contract. The price of Brent oil, the international benchmark, has been positive, even on Monday at USD 17 per barrel. This indicates that Monday’s event was more of an issue with storage of oil in the US, and the expiration of futures contracts, than a general issue with storage of oil. E.g., Brent oil can more easily be stored at sea, at least as long as tankers are available. One should also notice that the price of WTI was back in positive territory already Tuesday and stayed positive during the rest of the week. This was, even if very dramatic, a one-day shock.

Prices of both WTI and Brent have been falling since the outbreak of the virus in China in early 2020, however, particularly in early March when Europe and the US started shutting down. Except from the astonishing event on Monday, the price of both Brent and WTI have followed the same path. The agreement from April 10 limiting oil production has thus failed to slow the fall in oil prices.

Prices of WTI and Brent oil since January 1, 2020. Daily prices.
Data source: Thomson Reuter Datastream via Eikon.

The behavior of oil prices reflects the severity of this crisis. The drop in oil prices since the start of the year tells a different story about the recession than does the stock market. The stock market has sprinted ahead since mid-March, indicating that stock market investors believe the recession will soon be over (the stock market did fall on Monday and Tuesday, though, due to the turmoil on oil markets). Oil markets, in contrast, indicate that the recession is very much still ongoing. What signal do you believe – the one from the oil market or the one from the stock market?

V, U, W, or L? The stock market votes “V”.

How will this recession play out and how will the recovery look? Will we get a short recession with a fast rebound (V), will we get a longer recession before the recovery (U), will we get a new lockdown in autumn resulting from a second wave of the Corona virus (W), or will economies remain depressed for a long time (L)?

Nobody knows, particularly this time around when the shape of the recovery will be determined by a virus and how this develops.

Supporting the V-shaped recovery, this recession is not caused by economic imbalances, such as an overvalued housing sector, too much debt, a fragile banking sector, or the like. Also, we know from history that recessions last longer and are deeper if they are caused by financial crises. This is also not the case this time around, in contrast to 2008, for instance. This means there is hope for a reasonably fast rebound, if we manage to get the virus under control.

On the other hand, if we get a second outbreak, e.g., in autumn, and economies are shut down again, the path of economic activity will probably take the shape of something like a W.

What we do know is that the depths of the recession is unprecedented. The fall in economic activity and the increase in unemployment is mind-blowing, and very scary (link).

This makes one fear that the recession could drag out. When economies open up, it will take time to find jobs for those who get unemployed during the recession. Firms will probably be reluctant making new investments until we have a vaccine. Consumers want to go to restaurants, cinemas, on vacation etc., but will probably hold back until they feel on safe ground, too. Uncertainty abounds.

In this environment, one would imagine that the stock market suffers tremendously. With a recession impeding, and with so much uncertain surrounding the future path of economic activity, one would imagined that stock markets, like economic activity, would be in freefall. This is not the case.

In the beginning of the lockdown, the stock market tanked. It was the fastest bear market ever (link). It reminded us very much about the dark days of autumn 2008. This graph shows the SP500 on a daily basis during the 2008 financial crisis and this crisis. “0” in the figure is September 19, 2008, respectively February 21, 2020.

S&P 500, daily closing prices. Normalized to “1” 25 days before September 19, 2008, respectively February 21, 2020.
Data source: Fed St. Louis Database.

Financial markets were in stress in mid-March, when even yields on safe assets increased (link). Fearing a replay of 2008, central banks and governments came to the rescue, and the stock market started its recovery. Since the bottom on March 23, the S&P 500 has gained an astonishing 28%. This is remarkable, given all the uncertainties and the depth of the recession. The contrast to the autumn of 2008 is stark. The stock market kept on falling throughout 2008 and early 2009, only to start its recovery in March 2009.

The stock market seemingly believes this recession will result in a V-shaped recovery. Let’s hope it is right. One might fear that it is not.

Horrifying figures

The corona crisis is unique in many ways, not least with respect to the speed with which economies have started freefalling. Literally, from one day to the next, economies have been shut down, with grave consequences for economic activity.

Aggregate economic figures, such as GDP, consumption, etc., are collected and published with a lag. Normally, this is not a big problem, as changes in business cycles are typically not too abrupt. But this time, it has been a problem. In the early phase of the crisis, we had to rely on data from selected sectors that provided real-time insights. Restaurant bookings completely collapsed, footfall (number of people entering retail shops) fell, flight traffic dropped, etc. This was dramatic and foretold that economic activity more generally contracted significantly. But we did not know exactly how much.

We start getting the aggregate numbers now. They reveal with horrifying clarity that this recession is unprecedented.

Initial claims for unemployment insurance in the US has been published weekly since 1967. The average from 1967 through March is 350,000, i.e. on average 350,000 people enter unemployment per week in the US, with a peak at 700,000 in October 1982 and another high of 665,000 during March 2009. In mid-March 2020, the number was a staggering 3,3 million, only to reach more than 6 million getting unemployed during the last couple of weeks, with the number last week being 525,000. In just four week, 22 million people have lost their job in the US. It is mind-blowing. And very sad.

Initial jobless claims in the US per week.
Data source: St. Louis Fed Database.

We have today received the latest GDP figures for China, for the first quarter of 2020. China has published quarterly GDP data since 1992. Annual growth has always been positive, even during the Great Recession of 2008-2009. In the first quarter of 2020, Chinese GDP was 6.8% lower than in the first quarter of 2019. An unprecedented fall.

Year-on-year percentage change in Chinese GDP. Quarterly figures.
Datasource: Thomson Reuters Datastream via Eikon.

We have also just received the latest figures for US industrial production. The drop in industrial production in March 2020 was the largest monthly drop since 1946. Industrial firms in the US cut production with 5.4% relative to February 2020. We also notice, though, that monthly drops in industrial production were larger during the Great Depression in the early 1930s and at the end of the second world war.

Monthly percentage change in US industrial production.
Datasource: St. Louis Fed Database.

These pictures tell a horrifying story about the freefall in economic activity. Expect more sad numbers to come. Hopefully, when economies start opening up again, we will see a rebound in economic activity. But the recession will leave lasting scars.

Me on TV

Yesterday evening, I played the role of the expert in a primetime airing on the main Danish national broadcaster (link). The topic was the corona crisis and how it affects the economy, and, particularly, the Danish real estate market and real estate prices, as well as how it compares to the aftermath of the financial crisis of 2008.

I argued that I would be very surprised if we do not see a fall in Danish real estate prices. I also argued, though, that I would also be surprised if we will see Danish real estate prices fall as much as they did during the aftermath of the financial crisis.

As you might know, I chaired the government-appointed committee that investigated the causes and consequences of the financial crisis in Denmark (the so-called Rangvid-committee). The real estate market played a major role back then, i.e. the committee analyzed in detail the Danish real estate market before, during, and after the financial crisis. Danish real estate prices fell by something akin to 30% following the financial crisis.

Depending on a lot of uncertainties (the duration of the crisis, what happens to interest rates, etc., etc.), I believe it is a fair assessment that we will not see as large a drop this time. This has not least to do with the fact that there clearly was a Danish real estate bubble when we entered the financial crisis. There is no bubble now.

I talk a number of times during the airing, at: 2:20, 14:05, 18:40, 21:20, 24:45. Unfortunately, for non-Danish readers, the airing is the Danish.

(You also get a chance to see a corner of my living room :-). The hosts were in the studio, but due to the corona situation, I was recorded live at home).

The normal economy is never coming by Adam Tooze

It is by now painfully clear that the corona crisis will be very severe and very different from any other economic and/or financial crises most of us have ever seen. The Spanish flu in 1918-1920 is the closest example. And even this we cannot really rely upon as the economy and the role of the state and central bank was obviously much different back then. So, how bad is it and how bad will it be?

Adam Tooze (https://adamtooze.com/) from Columbia University has taken a step back and written this interesting account on where he sees the economy right now and where it is heading (link).

He writes: “The economic fallout from these immense human dramas defies calculation. We are left with the humdrum but no less remarkable statistic that this year, for the first time since reasonably reliable records of GDP began to be computed after World War II, the emerging market economies will contract. An entire model of global economic development has been brought skidding to a halt.”

I think Adam Tooze is right on many accounts. At the same time, we should not forget that the economy enters this crisis in a better condition than the one we faced when we entered the financial crisis of 2008. I believe there is hope that the economy will be able to bounce back when economies are opened up again, as is also the main scenario in the latest WEO from IMF (link). Economic activity will suffer tremendously, but we will come back.

Of course, everything depends on how the virus develops, and, in particular, whether there will be a second wave in autumn where we need to close down again. Should this happen, this will be the worst crisis ever. Should it not happen, it will still be very bad, but probably not the worst crisis ever. I.e., we should not be naïve, this is an unprecedented crisis, but it need not be doomsday.  

Eurozone in trouble again, but this time with a new twist

Fear of a new Eurozone debt crisis, similar to the one in 2010-2012, resurfaced in March. Within a week or two, Italian yields more than doubled. Given Italy’s large stock of sovereign debt, nervousness increased. Since then, yields have come down and markets stabilized somewhat. A key difference to the European debt crisis of 2010-2012 is that yields on “safe” assets, like German and US yields, rose, too. What happened and will calm remain?

First, the facts. This figure shows Italian yields (yield on a ten-year sovereign bond), as well as its spread to the German ten-year government bond yield, from the beginning of the year through April 14. The spread to German bonds is larger than the Italian yield itself because German yields are negative (see below).

Yields on Italian ten-year government bond and spread to German.
Data source: Datastream via Thomas Reuters Eikon.

Bond markets were calm in the beginning of the year, only to explode from March 3 through March 18. On March 3, the Italian benchmark yield was one percent. On March 18, it was 2.4 percent. Similarly, the spread to German yields rose. Given the speed of the adjustment, and the size of Italian government debt, this is worrisome. It reminds us of the situation in 2010 where yields on debt from Italy and other southern European countries rose sharply. There are, however, additional ingredients to the story this time.

The worry with Italy is that it will face difficulties remaining solvent if its yields rise. And, given the size of the Italian debt mountain, it will be difficult for the Eurozone to bail out Italy. It is also a story of disastrous communication by the new ECB chief Christine Lagarde, who at the ECB Press conference on March 12 said the by-now famous words “We are not here to close spreads”, meaning ECB will not come to the rescue of Italy. Many of us wonder why Italy has not tried to stabilize its finances since the European debt crisis, but this was not the right time to raise this point. Italy was in the middle of the terrible corona crisis and markets were nervous. The remark of Lagarde should have been saved for another day. Markets tail-spun.

This time around, however, there is an additional element to the story. Something unusual happened in other bond markets. Look at this graph.

Yields on ten-year government bonds from different countries.
Data source: Datastream via Thomas Reuters Eikon.

It shows yields on ten-year sovereign bonds from a number of countries. Italian rates spiked, as mentioned. However, and this is the curios fact, rates of what is typically viewed as safe-haven bonds also rose dramatically.

German yields rose from their low of minus 84 basis points on March 9 to minus 17 basis points on March 19, an increase of more than sixty basis points. The same goes for Danish and Dutch yields. Denmark, the Netherlands, and Germany are Triple-A rated. Even US yields more than doubled, from fifty basis points to 126. A rise of eighty basis points in US yields within a week or so is very dramatic. It is also very different from what we typically see during times of crises. Typically, in crises, investors sell risky assets and buy bonds of safe-haven countries, causing their yields to fall. This did not happen in March. If safe assets lose value during crises, where should investors seek shelter?

So, this is a story about Italy and a reassessment of the credit risk of Italy. But it is more than this because otherwise safe havens saw yield rose, too, exerting an additional upward pressure on Italian yields.

From early March to mid-March, Italian yields rose by almost 150 basis point. The yield spread to Germany “only” rose by 100 basis points. The difference is the rise in German yields.

This BIS Bulleting provides an interesting account of what caused the rise in US yields during early-mid March. They only look at the US, but probably some of the same factors caused the effects in Germany. BIS argues that hedge funds and other levered investors were forced to sell because of margin calls. “Leverage” means that you borrow to invest and “margin calls” means that you have to come up with extra money when the assets you have bought for borrowed money fall in value. So, when you have borrowed a lot to invest (you are highly leveraged), you will face large margin calls. In this case, you have to sell many assets to generate a lot of cash you can pledge as security. If markets function “perfectly”, dealers will be able to absorb the sales and build up inventories. Prices should be unaffected. But if dealers cannot absorb the sales, for instance because dealers are subject to capital requirements and cannot raise capital immediately, then there is nobody out there to buy and prices of bonds fall and yields rise. BIS argues that these events had spillover effects on other types of investors causing them to sell bonds, too. I.e. a lot of investors had to sell, causing such a big effect. In their FSR, IMF is aligned.

More broadly, BIS argues, these events imply that central banks might need to employ different tools from what they have been used to during previous crises (buy bonds from dealers instead of providing liquidity). The events also imply that a different composition of sovereign bond investors (leveraged private investors instead of sovereign wealth funds) could imply different market dynamics during periods of stress.

Where does this all take us? First, I am still worried about Italy. The Italian sovereign debt is huge, and the situation is shaky. Right now, it might look OK. Italian yields are higher than in January and February, i.e. spreads to e.g. Germany have widened, but not by too much, even if the tendencies of the last couple of days might be worrying. Unrest could easily resurface, however. In addition, events in March told us an important lesson about the implications of changes in market structures. If owners of sovereign bonds of safe countries are mainly leveraged private investors, and if dealers cannot absorb large sell-offs, even yields of safe-haven assets might rise during turmoil. This is worrying for investors, as they will then have fewer places to seek shelter during times of stress.

So, Italian yields rose because investors repriced credit risk in Italy and ECB made a blunder. Spreads to Germany rose. But, Italian yields also rose because yields on safe assets (German and US) rose, pushing up all yields. The latter effect is new, and somewhat scary.