Guest Contribution: “Gauging Recessions with the Jobs-Workers Gap”

Today, we are fortunate to present a guest contribution written by Paweł Skrzypczyński, economist at the National Bank of Poland. The views expressed herein are those of the author and should not be attributed to the National Bank of Poland.

In February Goldman Sachs introduced the jobs-workers gap (Briggs & Hatzius, 2022). This measure is defined as the difference between the total number of jobs (i.e. employment plus job openings) and the number of workers (i.e. the labor force). We present the most recent reading of this gap and show a simple rule of using it in order to track the business cycle.

Formally, we calculate the jobs-workers gap for month  as ,

where Et is household employment level, JOt is job opening level and LFt is labor force level. Notice that the gap calculated in this manner is scaled in number of workers. Alternatively one can express this gap relative to labor force level, i.e. JWGt/LFt as plotted in Figure 1. Notice that job openings data is available from Dec 2000, which means that some proxy of it is needed to calculate the jobs-workers gap for months prior to Dec 2000. To achieve that one uses the Help-Wanted Index developed by Barnichon, 2010 (data available on this website). Namely for months prior to Dec 2000 the Help-Wanted Index is used, while job openings time series for Dec 2000 and further months.

Figure 1. Jobs-Workers Gap (Percent)

Notes: Calculations based on BLS releases The Employment Situation – August 2022, September 2, 2022, Job Openings and Labor Turnover – July 2022, August 30, 2022 and historical data for the Help-Wanted Index by Barnichon, 2010.

As of July 2022 the jobs-workers gap was at 5.6mn workers, meaning that employment and job openings (total labor demand) exceeded labor force (total labor supply) by 3.4%. Historically this reading was near an all-time high level of 3.6% (5.9mn workers) reached in March 2022. While job openings data for August is not available at the time the August household employment and labor force data is released, we make a working assumption that job openings did not change in August compared to July. Alternatively one can use job postings weekly data from Indeed to make a bridge regression and nowcast August job openings. Assuming constant level of job openings in August and incorporating August employment situation release (household employment increased by 442k m/m and labor force increased by 786k m/m) we find that the jobs-workers gap was at 3.2% (Figure 1) down 0.2 percentage points from July level.

What magnitude of change from the current level of the jobs-workers gap would be sufficient to generate a recessionary signal? To answer this question we propose a following measure. Namely, we calculate the change of the three-month moving average of the jobs-workers gap, i.e. JWGt/LFt , relative to its maximum during previous twelve months. . Smoothing allows us to get a less volatile time series which is more useful for tracking the cycle. The resulting measure, called the jobs-workers gap business cycle indicator (JWGBCI), is scaled in percentage points and plotted in Figure 2. Notice that derivation of this indicator is similar to what is done with the three-month moving average of the unemployment rate when calculating the so called “Sahm Rule”. Along with the JWGBCI time series we plot a recession threshold resulting from Berge & Jordà, 2011, framework. Namely we calculate a threshold value of the JWGBCI that maximizes the utility of the classification into periods of recession and expansion, so that the benefits of hits equal the costs of misses in magnitude. In other words the recession threshold value found using this method, which is -0.93 percentage points, is the optimal level of the JWGBCI that separates recessions from expansions.

Figure 2. Jobs-Workers Gap Business Cycle Indicator (Percentage Points)

Notes: Calculations based on BLS releases The Employment Situation – August 2022, September 2, 2022, Job Openings and Labor Turnover – July 2022, August 30, 2022 and historical data for the Help-Wanted Index by Barnichon, 2010.

If we were to declare that the recession began, we would need to see the JWGBCI fall to -0.93 or below. As of July 2022 the JWGBCI was at -0.19 percentage points versus -0.15 percentage points in June. Our conditional “guesstimate” based on the unchanged level of job openings in August was at -0.22 percentage points. Thus we conclude that despite labor market conditions eased somewhat at the beginning of Q3, they remained tight and visibly distant from levels associated with recession. Notice that for example if we were to observe the JWGBCI to reach -0.93 in September, we would need to see the underlying jobs-workers gap fall 2.2 percentage points m/m to the level of 1% (i.e. equivalent of 1.6mn workers gap in terms of August labor force level). Such a big move in one month was not seen, apart from the initial period of the COVID-19 pandemic, and given current economic developments seems highly unlikely.

Overall, we show that current labor market conditions remain tight and that they would need to deteriorate strongly in order for the JWGBCI to generate a recession call. Softening of the JWGBCI readings do not mean that the recession is imminent, unless the recession threshold is reached. Therefore we would interpret future values of this indicator above the recession threshold as being consistent with the soft landing scenario.

References

Barnichon R., 2010, Building a composite Help-Wanted Index, Economics Letters, 109 (2010): 175-178.

Berge T. J., Jordà Ò, 2011, Evaluating the Classification of Economic Activity into Recessions and Expansions, American Economic Journal: Macroeconomics, 3 (2): 246-77.

Briggs J, Hatzius J., 2022, More Jobs than Workers: A New Measure of Labor Market Tightness, US Economics Analyst, Goldman Sachs Economics Research.

 

This post written by Paweł Skrzypczyński.

23 thoughts on “Guest Contribution: “Gauging Recessions with the Jobs-Workers Gap””

  2. Macroduck

    This looks like an alternative to the Sahm rule, a concurrent indicator of recession. The Sahm rule moves gradually enough that one can find hint of coming recession. Maybe here, too. It’s nice to know we’re not in recession. It’s useful to know the extent of the risk.

    By the way, I think Powell’s impersonation of Volcker was part showmanship. Fed policy doesn’t work if we don’t let it; expectations of easing were showing up, limiting the impact of rate hikes. Powell and company don’t want to use overnight rates to reinforce tightening if language will do – too damaging. So let’s hope Powell makes this paper required reading, and that Fed folk look for hints of recession risk.

  3. Macroduck

    Here’s a picture of the Skrzypczyński rule, the Sahm rule (inverted) and a bailing-wire Beveridge curve:

    You may find similarities. Play with the buttons. It’s fun.

  4. pgl

    https://jabberwocking.com/chart-of-the-day-net-new-jobs-in-august-2/

    Kevin Drum’s latest notes that the labor force participation rate rose by more than the unemployment rate rose which is what Macroduck and I noted means the employment to population ratio rose (household survey employment rose by 442 thousand). But his graph shows more. Yes the unemployment rate since Biden took office has declined a lot even as the participation rate has increased.

    So all you Biden bashers (Bruce Hall) and RECESSION cheerleaders (JohnH and his BFF Princeton Steve) take note – the employment to population has increased considerably since Biden took office.

  6. Moses Herzog

    Very interesting, some of it still above my head and needs more absorption and thought. I don’t know if it was pure coincidence or not, but I also think it’s very cool Professor Chinn and Pawel posted this near to Labor Day.

  7. ltr

    https://news.cgtn.com/news/2022-09-02/Chinese-mainland-records-373-new-confirmed-COVID-19-cases-1cZzIomIrxm/index.html

    September 2, 2022

    Chinese mainland records 373 new confirmed COVID-19 cases

    The Chinese mainland recorded 373 confirmed COVID-19 cases on Thursday, with 318 attributed to local transmissions and 55 from overseas, data from the National Health Commission showed on Friday.

    A total of 1,665 asymptomatic cases were also recorded on Thursday, and 24,226 asymptomatic patients remain under medical observation.

    The cumulative number of confirmed cases on the Chinese mainland is 243,822, with the death toll from COVID-19 standing at 5,226.

    Chinese mainland new locally transmitted cases

    https://news.cgtn.com/news/2022-09-02/Chinese-mainland-records-373-new-confirmed-COVID-19-cases-1cZzIomIrxm/img/1024fbe16ed14f7a9681e194ac0698c8/1024fbe16ed14f7a9681e194ac0698c8.jpeg

    Chinese mainland new imported cases

    https://news.cgtn.com/news/2022-09-02/Chinese-mainland-records-373-new-confirmed-COVID-19-cases-1cZzIomIrxm/img/9edf4791d0fa47c488426def8fdf9ab8/9edf4791d0fa47c488426def8fdf9ab8.jpeg

    Chinese mainland new asymptomatic cases

    https://news.cgtn.com/news/2022-09-02/Chinese-mainland-records-373-new-confirmed-COVID-19-cases-1cZzIomIrxm/img/3e3e701636174579a29181a71d3d8939/3e3e701636174579a29181a71d3d8939.jpeg

      1. Barkley Rosser

        There you go again, ltr. Incorrigible. But I guess with that UN report out, you need to just keep pounding away with the same old number.

      2. Ivan

        Since May 27’th China has reported 19612 diagnosed cases of Covid but not a single death. That is a miracle. UNBELIEVABLE. Yes literally, UNBELIEVABLE

        You must reveal the brilliant strategies in medicine and recordkeeping that allows for such unprecedented success in fighting what in the rest of the world is a deadly virus killing about 1 out of every 500 diagnosed.

        Also please explain the brilliance of dear leader’s policy to shut down the economy in an attempt to fight a pandemic that according to his own released numbers is much less deadly than the annual flue? If nobody dies from this why all those sacrifices?

  10. pgl

    Poor little Melania was violated:

    https://www.msn.com/en-us/news/politics/melania-trump-felt-violated-by-fbi-agents-contaminating-her-bedroom-during-mar-a-lago-raid-report-says/ar-AA11qpPn?ocid=msedgdhp&pc=U531&cvid=fe459795993a4c5d99e2344ac9253a6a

    FBI agents searched Melania Trump’s closet during the Mar-a-Lago raid last month.
    The former first lady felt violated by strangers rummaging through her belongings, friends told Radar Online.
    She bought new underwear because she felt as though her belongings had been “contaminated,” her friends claimed.
    Former first lady Melania Trump felt violated after FBI agents “contaminated” her home while executing a search warrant at Mar-a-Lago last month, her friends claimed, per The Times of London and Radar Online.

    What a effing princess. Hey bimbo – you were the moron who decided to marry a mob boss and traitor. There are very brave CIA agents overseas risking their lives who have been put in mortal jeopady by your meal ticker. No one gives a damn about your little panties.

    1. Moses Herzog

      The least they could have done was check Stormy Daniels’ and Karen McDougal’s closet first. What does a girl have to do to get respect these days?!?!?! Isn’t it bad enough Putin is killing off all of donald trump’s East European pool of future wives??

      1. pgl

        It seems the FBI searched the room of their son. I wonder if Barron had any naughty pictures of Stormy Daniels.

  11. pgl

    BLS has released the JOLTS report for July. The August report comes out on October 4.

    https://www.bls.gov/news.release/jolts.nr0.htm

    Note job openings are about twice the number of people unemployed. Of course that is a level concept. JOLTS also reports hires and separations. Hires exceeded separations by around 0.5 million in July.

    JOLTS indicates a very strong labor market that got stronger in July.

    Don’t let the cherry pick spin from Princeton Steve and his new BFF confuse you. They are not even trying to be honest.

  12. pgl

    Bain Capital dragged into a scandal involving sex with underage cheerleaders down in Lindsey Graham’s backyard:

    https://www.greenvilleonline.com/story/news/local/2022/09/02/rockstar-cheer-greenville-sc-lawsuit-additional-claims-sexual-misconduct/7958495001/

    Rockstar Cheer founder Scott Foster, who died by suicide last week, has been accused of a wide range of misconduct and illicit sexual behavior with minors in a federal lawsuit attacking the culture and organization of competitive cheerleading. Foster is accused of soliciting sex and explicit photographs from athletes under his care. However his behavior, the lawsuit alleges, is a symptom of a much wider problem. The suit, filed Thursday in Greenville, names several defendants, including Foster’s estate, his widow Kathy Foster, Rockstar Cheer and Dance Inc., national cheerleading organizations, the United States All Star Federation (USASF) and Varsity Spirit LLC, and corporate entities Charlesbank Capital Partners and Bain Capital. According to the lawsuit, Bain bought Varsity Spirit from Charlesbank in 2018. In over 44 pages, the complaint, filed on behalf of six anonymous plaintiffs by attorneys from the Columbia-based Strom Law Firm, details an exploitative system with little accountability that “created, organized and propagated a system of young-athlete abuse against innocent victims.” “This was a factory of abuse designed specifically to generate two things: a constant supply of underage victims for Scott Foster and his fellow predators and a billion dollar revenue stream to Varsity Spirit, USASF and Bain Capital,” said Strom attorney Bakari Sellers. “Instead of protecting these young men and women, they victimized them and cashed their checks.”

  13. pgl

    Even Bill Barr has turned on Trump:

    https://www.theguardian.com/us-news/2022/sep/03/william-barr-trump-mar-a-lago-fbi-justice-department

    Former attorney general William Barr came to the defense of the FBI and the justice department’s (DoJ) judicial request to search Donald Trump’s Florida home and country club compound last month, saying Friday that documents seized in the search appeared to support the department’s claims of a national security risk.“ As more information comes out, the actions of the department look more understandable,” Barr told the New York Times in an interview. “It seems to me they were driven by concern about highly sensitive information being strewn all over a country club, and it was taking them almost two years to get it back,” he said. The former attorney general, who resigned in December 2020 as Trump ramped up his claims that the election has been rigged, also dismissed Trump’s call for an independent review of materials confiscated from Mar-a-Lago by a “special master”. “I think it’s a crock of shit,” Barr said. “I don’t think a special master is called for.”

    1. pgl

      There are more than 480,000 job vacancies across the country, but with unemployment at an almost 50-year low, employers are struggling to fill the gaps.

      Relative to their population these are huge gaps. But no worries – Princeton Steve will find some Aussie statistic to manipulate so as to suggest an Aussie hard reset!

Comments are closed.