Shortly following our May publication of this series ("U.S. Business Cycle Barometer: Digging Out Of A Deep Hole," published May 29, 2020), the National Bureau of Economic Research (NBER) announced the U.S. had entered recession following a peak in monthly economic activity that occurred in February 2020. The peak marked the end of the longest expansion in U.S. history that began in June 2009 (following the Great Recession) (1). The committee also determined that a peak in quarterly economic activity occurred in fourth-quarter 2019 (2). This announcement was the quickest determination of a recession in NBER history. We suspect that this recession will also be the shortest in U.S. history.

Although the NBER hasn't announced officially when the economy troughed (i.e., recession ended), data out since May on both coincident and leading indicators reveals that the U.S. economy bottomed when it started opening in varying capacity across states near the end of April. After contracting by 1.3% and 9% in the first and second quarters (-5% and -31.4% annualized), respectively, U.S. real GDP jumped by 7.4% in the third quarter (33.1% annualized).

Chart 1

It's quite clear now that unless the numbers get drastically revised, the recession could now be said to be over and we've begun the recovery phase following the 12th post World War II U.S. recession. Where does this remarkable swing leave the economy in the beginning of the fourth quarter? A recovery underway with the economy still about 3.5%, or $167.5 billion ($670 billion annualized), below its precrisis peak set in fourth-quarter 2019.

Probability Of A Recession

It is hard to miss the enormous swings to begin the new business cycle. On a quarter-over-quarter change basis, both the second-quarter downturn and the third-quarter recovery were well outside of the historical mean--11 standard deviations (11-sigma event) and eight standard deviations (8-sigma event), respectively (3). While the COVID-19 recession is a remarkable outlier in terms of the magnitude of the change in real GDP (and other supporting co-incident indicators) and the initial recovery, it is also different from other business cycles in history in terms of its drivers. In all previous historical recessions, consumer spending on services was quite stable during the downturn, but this time, consumer services have been the key driver of the swing (given the nature of the shock) and will remain a headwind to recovery until a vaccine for COVID-19 is widely available.

This means a normal distribution-based quantitative model to estimate recession probability will not adequately characterize the monthly swings spanning from the end of the first quarter to the third (4). That means qualitative judgment is required on top of any results from our quantitative recession model. Our regression model estimates the probability of a recession within the next 12 months (based on the bond market and equity market pricing) has receded steadily from its 50% peak in March to now at 24%.

With viral cases climbing higher and emergency stimulus fading, the risk of a relapse remains high, something not necessarily captured in our regression model. But until there are trusted and widely available treatments for COVID-19 (which we assume sometime in the first half of 2021), services activities will likely grow only modestly and chances of another restrictions-led negative handle on activity growth data in the next six months remains elevated. We peg our probability of a recession within the next 12 months at 25%-30% (previously 30%-35%). The risk is closer to the top of the range given the elevated uncertainty over the path of the virus and an agreement on further stimulus.

Chart 2

We have already observed the recovery path fade in recent weeks as the virus cases have resurged widely across the U.S. (and Europe) and additional fiscal support was punted until at least post-election. Monthly data suggest that growth in the economy has already slowed from the outsized gains between May and July, so we expect real GDP growth to moderate to 0.86% quarter over quarter in the fourth quarter (+3.5% quarter over quarter annualized).

Leading Indicators

Most leading indicator activities rebounded strongly as the economy opened by varying capacity and localities in May following strict lockdowns and social-distancing measures. Out of the 10 indicators of near-term U.S. economic growth, which we closely look at, four indicators are in positive territory (was one in May publication), four indicators are in neutral (was one), and two are in negative (was eight) (see table).

Stable financial markets positive for growth

The Fed's highly accommodative monetary stance, including a lower-for-longer zero interest rate outlook (we have penciled in 2024 for first rate hike in the current cycle), continuation of large scale asset purchases (currently at $120 billion per month), and standing emergency credit facilities on various asset classes swung the yield curve (10 year minus three-month Treasury) pendulum to positive spread (positive slope is considered normal and expansionary, while negative usually precedes contraction) since March (currently at 50 basis points).

Chart 3

Chart 4

At the same time, credit spreads on both investment-grade corporate bonds and speculative-grade bonds have narrowed steadily since April when they climbed to near early-2016 stress levels. The New York Fed estimates that, within the first three months following the Corporate Credit Facility announcement, credit spreads retraced about 90% toward pre-pandemic levels, with two-thirds of the improvement occurring on facility announcement dates (i.e., before purchases even began) and that half of the improvement in bid-ask spreads since the peak in March occurred on the initial announcement date itself. Still, high-yielding corporate bond spreads remain slightly elevated compared with pre-pandemic level.

The national financial condition has become accommodative with the Fed's actions and provided additional space for increased leverage in the system. The capital markets opened up, taking pressure off the banking sector that has tightened credit standards for both corporate and household loans. The Fed's Senior Loan Officer Survey also pointed that demand for loans fell significantly.

Chart 5

Chart 6

The equity market continues to outperform since its March nose-dive. The S&P 500 index increased by nearly 14.8% from February, reflecting investors' confidence on the Fed's monetary policy and the government's stimulus measures to support economic recovery. The wedge between S&P 500 growth and value index is at an all-time high, which signals that investors are favoring high-growth companies over attractively valued stocks.

Chart 7

Chart 8

It is worth mentioning one of the explanations from our previous publication why the equity market has recovered as well as it has. The underlying backstop of monetary and fiscal authorities doing "whatever it takes" has been a key stabilization force for the financial markets. Corporate profits are a function of investment, dividend and buybacks, savings of households, savings of government, and the rest of the world (also known as the Kalecki Profits equation). The unprecedented government spending and transfers is making up for a substantial portion of investment collapse, a spike in household saving, and a decline in dividend and buybacks. If the government hadn't stepped up as it did, the story would have been different.

It then follows that the current positive stock market, as well as the emerging recovery, could easily fall flat if (i) there is a resurgence in pandemic spread with another round of stay-at-home directives, or (ii) there is policy fatigue before an effective treatment is available. A space to watch.

A string of positives from the production sector

That said, the business sentiment remained upbeat. The Institute for Supply Management (ISM) manufacturing index came in at 59.3 in October (the highest level since September 2018) following 55.4 in September, remaining well above the pre-COVID-19 level. ISM manufacturing new orders for October printed at 67.9, which was the highest since January 2004 and about 11 points higher than February's level. Further, core new orders (nondefense excluding aircraft) continued to display strength in domestic demand for manufactured goods, which increased for the fifth straight month in September, albeit the pace of monthly growth slowed to its lowest level since the recovery began in May.

Chart 9

Chart 10

Encouragingly, the inventory-to-sales ratio continued to fall in August after surging to a record high in April and was back to pre-crisis levels. As the economy recovered and sales picked up, inventory level decreased. This is indeed a positive sign before the upcoming holiday season for the firms to produce more goods, with the inventory-to-sales ratio staying low.

The freight transportation index has regained substantially since the pandemic drop but remains 3.7% below February levels. The improvement in freight transportation services indicate gradual pickup in the transportation sector, broadly supported by rail intermodal, rail carload, and truck tonnage. The August freight transportation index level was 7.4% below its all-time high in August 2019.

Chart 11

Chart 12

The housing sector was one of the bright spots in the U.S. economic recovery following a sharp contraction In April and May. The record low mortgage rate and change in household preference (locality), combined with already building pent-up demographic demand, bolstered housing demand. Building permits (which is considered a forwarding-looking housing starts indicator) staged a V-shaped recovery--in September, single-family building permits increased by 7.8% month over month to 1.12 million units at seasonally adjusted annualized basis. With demand outstripping the supply of homes, supply of existing home has fallen to near historic low of 2.5 months (in September), while the supply of new home was at 3.6 months (the second-lowest since August 2003). This will further encourage homebuilders to increase construction.

Chart 13

Elevated jobless claims and subpar consumer sentiment headwinds to holiday spending

The University of Michigan consumer sentiment index has started to recover from the trough since the economy reopened. Yet the October consumer sentiment index was still well below February levels with the resurgence of COVID-19 cases and waning of strong recovery in the third quarter. Consumers are still uncertain about their economic outlook, and the expiration of the unemployment benefits booster will deter consumer spending for those who are unemployed during this holiday season. Initial jobless claims fell substantially from April's high of near 7 million the last week of March to an 800,000 per week average in October but remained about four times higher than February pre-pandemic levels and 200,000 per week above June 2009 when the Great Recession ended.

Chart 14

Chart 15

Where Do The Coincident Indicators And Capacity Utilization Stand?

Monthly data on the key coincident indicators suggests that after an initial burst of recovery, the economic activity decelerated heading into the fourth quarter, confirming our expectations in the May report that the recovery will slow down past summertime. The level of utilization of productive capacity remains well below normal.

The economy added 11.4 million new jobs to the nonfarm payrolls from May to September (latest data available), after a dramatic loss of 22.16 million jobs in March and April. Although the COVID-19 recession destroyed nearly three times the job loss in one-third of the time compared with the last crisis, the initial recovery has also created new jobs at a much faster pace. Having said that, the pace of job creation has slowed down, with only 661,000 net new jobs added to the nonfarm payrolls in September, less than one-fourth the pace of the job gains in June.

Chart 16

Chart 17

By September, little more than half the number of jobs lost has been recouped but the next half is going to take longer time to recover as we now go through the "sticky" part of the unemployment pool. The standard measure of unemployment rate (U-3) topped out in April at 14.7% (versus 3.5% in February, 10% peak 2008-2009 recession) and now stands at 7.9%--at or higher than previous unemployment peaks in eight of the last 11 recessions going back to 1948. With an employment-to-population ratio of prime-aged (25-54) workers at 75%, labor capacity utilization stands at just about the lows of the Great Recession of 2008-2009. That's about 8.8 million missing prime workers compared to a full employment economy of 1999-2000 and 6.9 million missing prime workers compared with pre-pandemic high.

The industrial production index tells a similar story. The consecutive recovery from May to August has stalled since July, leaving the activity in the industrial sector still 45% below its peak in November 2019. Capacity utilization in the industrial sector (mining, utilities, and manufacturing combined) was 71.5% in September, materially below 2019 average of 78% (and historical average of 80%).

Chart 18

Chart 19

The real manufacturing and trade industries sales, however, have recovered to more than the pre-pandemic level. Spending recovery in the goods sector was quick out of the gates--by July, sales had already reached 100.6% of its peak in February, after a nose-dive in April. With any other recession in the past, this would be enough to pull the overall economy back to its pre-recession level, but not so during the current recovery. Spending in goods take up a smaller share of the overall spending (31% in 2019) but is more cyclical compared with spending in services. Had the larger service sector remained intact--like it generally does in normal business cycle recessions--the recovery in real manufacturing and trade industries sales would be nearly equal to a full recovery of the economy. However, given the touch-sensitive nature of the 2020 recession, the COVID-19-challenged service sector continues to weigh down the overall economy even as spending in the cyclical goods sector has completely recovered.

Chart 20

The spending power of households has been held together by unprecedented transfers from the government. The real personal disposable income per capita spiked to 113% of its pre-crisis level in April, reflecting countercyclical income support from the government, which was timelier and more generous compared with fiscal response in previous recessions. Having said that, through August, real personal income has been declining since April from a fading fiscal boost as the unemployment benefits booster has expired and temporary moratoriums on payments are set to expire by the end of 2020, if not already.

If we take income transfers out of the equation, the improvement in wages and salaries still puts overall income excluding transfers 4% below pre-pandemic level in August, following an approximately 8% peak-to-trough decline. In comparison, in the Great Recession, there was an approximately 8% decline (as well) in a much longer 21 months, and the recovery back to pre-recession peak took 26 months--a long round trip partly due to premature fiscal austerity.

Chart 21

Chart 22

Endnotes

(1) The last expansion lasted 128 months, the longest in the history of U.S. business cycles dating back to 1854. The previous record was held by the business expansion that lasted for 120 months from March 1991 to March 2001.

(2) NBER notes that the monthly peak (February 2020) occurred in a different quarter (2020Q1) than the quarterly peak. The committee determined these peak dates in accord with its long-standing policy of identifying the months and quarters of peak activity separately, without requiring that the monthly peak lie in the same quarter as the quarterly peak.

(3) An observation that is more than 3 standard deviations from the historical average, or a 3-sigma event, is sometimes called a "black swan"--something that is not at all expected, until you see one. Our calculation is based on the standard deviation of quarterly changes in the log of real GDP over 1947:Q2 to 2019:Q4.x

(4) Nonetheless, this may be less of an issue for our preferred model which uses monthly (not quarterly) categorial identifier based on NBER's announcement (which is not based on only GDP but several coincident indicators) and our belief that the recovery here onwards will start to show similar business cycle recovery properties as in the past now that the initial swing from the exogenous shock is behind us.