Is the economy slowing down? Are we entering a recession? Are businesses and consumers becoming riskier? These are just a few of the glaring questions causing extensive debates! The question of whether the economy is slowing down is a multi-faceted one that requires a deep dive into factors such as GDP, consumer spending, interest rates, income and employment. In this post, we will examine the importance of unemployment in economic cycles and the necessity for fresh, granular and forward looking data.
Economic cycles start with an expansionary period that follows a contractionary correction, i.e., a recession. Looking back at recent historical recessionary events, a slowdown in cyclical economic trends does not happen abruptly but filters through the economy in different stages. This usually starts with inflationary pressure as the economic expansion matures, which triggers central banks to tighten their monetary policies to curb the supply of money in the economy in order to tame inflation.
High inflationary pressure in the United States started in the first quarter of 2021. This was largely caused by disruptions in the supply chain due to COVID-19 pandemic paired with substantial stimulus money injected into the economy through different means. The Federal Reserve responded by adopting quantitative tightening policies starting around November 2021, including several consecutive hikes in interest rates starting in March 2022.
Higher Fed fund rates have a complex and lasting effect on different aspects of the economy, including higher cost of borrowing that makes both credit and investment more expensive across the entire economy. The housing market is usually the first to respond to this higher cost as it is sensitive to interest rate changes. Reduced home prices and sales are the first-order effects of these policy changes that we have already observed through reduced mortgage applications data and a downturn trend in the number of issued housing permits as (Fig. 2).
As the demand for housing tapers off, it will next disrupt the construction sector as less housing means less need for hours worked in this sector. This trend started in September 2022 (Fig. 3).
A slowed-down housing market also indirectly affects the manufacturing sector, as less housing built means a decrease in the need for consumer durables such as appliances, electronics, furnishing materials, etc. This second-order effect is also manifesting itself through reduced outputs for this sector of the economy as it can be seen in the manufacturing production data (Fig. 4)
Fig. 4: US manufacturing production year over year growth
(Source: Federal Reserve)
In the last stage of an economic contraction, once businesses (starting with those in construction and manufacturing) experience a decline in activities and therefore revenue, they resort to job cuts, resulting in rise in unemployment.
A turning point experienced by unemployment has historically been a dependable indicator of an imminent recession. In addition, the Fed pays careful attention to unemployment signals as it can help them determine if an overheated economy has cooled down and thus to pivot their monetary policies. Therefore, it is critical to monitor unemployment indicators as it’s often the last leg of a slowing economy that can provide invaluable insights into the subsequent stages.
As argued above, monitoring unemployment indicators is crucial in understanding the state of the economy; however, accessible unemployment data is often stale. For example, the layoffs data provided by the U.S. Bureau of Labor Statistics (BLS) is lagged by 1-2 months. At the time of writing this blog (March 2023), the last data point provided by BLS is as of January 2023. This makes it difficult to act in a timely manner as the data reflect the state of the economy in January, whereas significant changes have taken place since then.
Fortunately, there is a way to access not only fresh but also forward-looking data around unemployment through the Worker Adjustment and Retraining Notification (WARN) act. The WARN act, which was enacted in August 1988, requires businesses to provide 60 days notification in advance of a plant closing or mass layoffs. Federal law requires a business to file a WARN notice if the business employs more than 100 employees and the plant closing or layoff will affect more than 50 employees. Additionally, some states have instituted more rigorous requirements
WARN data is notoriously difficult to collect as each state has varying formats and update frequency. At Livesight, we use machine learning with human-in-the-loop to collect, clean and process it, and using our proprietary and state of the art BusinessMatch product, enrich it with reliable industry code (NAICS or SIC) as well as refreshed firmographic data (e.g., headcount, revenue, and growth metrics). This data is an early indicator of upcoming layoffs and therefore a reliable predictor of unemployment.
There is no doubt that the tech sector has stolen the spotlight with a whopping 400% year-over-year increase in layoffs. Although pandemic over-hiring and squeezed profit margins due to higher interest rates and inflation seem to be the main drivers, there are a myriad of other reasons that can explain this trend through a secular lens and therefore might not be a robust signal in understanding the state of the economic cycle. As mentioned before, a more reliable sign of a slowing economy can be probed in the construction and manufacturing sector where the data over the past 12 months shows a significant spike in construction and manufacturing layoffs, which are troubling signs that could point to the onset of a steeper downturn (Fig. 5).
Fig. 5: Normalized number of early layoffs for manufacturing and construction sector
(Source: Livesight Early Layoff Indicator Data)
What makes seeing the signal through the noise even harder is that the impact of unemployment in any economic downturn is heterogeneous across different segments of the economy (e.g., industry and geography). For example, some industries historically have been robust to economic downturns such as education and healthcare, versus other industries such as construction and manufacturing that are more susceptible. The heterogeneity effect of unemployment can also be seen across different geographies. For example, COVID-19 layoffs did not impact many geographies in the United States, while others were severely impacted (Fig. 6).
Fig. 6: Early layoff 3 months trends2 for metro and micropolitan areas in U.S.A., as of March 2020
(Source: Livesight Early Layoff Indicator Data)
In sum, even though the national level unemployment rate is near all time lows (as in almost every recent recession), you have been WARNed (pun intended!). The data shows elevated risk of unemployment across critical segments of the economy (construction and manufacturing) and having access to highly granular and forward-looking data can be a powerful assistant in understanding the trajectory. With Livesight’s Early Layoff Indicator data, you can equip your risk mitigation arsenal with a powerful leading indicator of unemployment risk at very granular levels. This data is available at the business level (individual WARN notices) as well as aggregated data at the industry (NAICS, SIC), geography (MSA, ZIP), size (revenue and headcount bands), and any of these combinations.
You can access Livesight’s Early Layoff Indicator through two primary means: an API connection or an automated data batch delivery with options for daily, weekly, and monthly update frequency. You can also test up to 12 months of the most granular data for free. Contact us today to set up a demo call and get ahead of the risk curve.
