How machine learning can help alleviate the labor shortage in the US

Experts have debated the causes of the US labor shortage. But one thing is painfully clear: there is a staggering disparity between the number of available jobs (over 10 million) and the number of job seekers (about 6 million).

In this short article, we take a step back and look at how we got here, the multiple factors that led to such disparity, and some of the solutions being implemented to combat this problem. In particular, we will look at machine learning (ML) and how it is being used to alleviate both the causes and consequences of the US labor shortage

The current labor shortage in the US

According to the American Chamber of Commerce, labor participation has fallen in recent years from 63.3% to 62.3%. While a 1% reduction in the number of skilled workers participating in the workforce might not otherwise pose a huge national problem, it comes after a pandemic that saw more than 30 million workers lose their jobs.

The sectors hardest hit include leisure and hospitality, food service, durable goods manufacturing, education and healthcare. But there is virtually no industry that has not been affected.

What are some causes of the labor shortage?

The COVID-19 pandemic has indeed shaken up the labor market. Studies show that about a quarter of a million people of working age have died from the disease, half a million people have left the workforce due to lingering health effects from the virus, and a similar number of workers have gone straight from illness to retirement.

This reduction in the workforce should have been compensated by job seekers wanting to enter the market, but that has not happened. Instead, the US saw an increase in the monthly quit rate across all sectors. In some industries, such as leisure and hospitality, the monthly quit rate exceeds 6%. Traditionally more stable sectors, such as business and professional services, are still recording an alarming churn rate of over 3%.

Many employees have expressed a desire to continue working from home. For some sectors, such as healthcare and manufacturing, this is a difficult expectation. But this shift in employee expectations is only superficial. Childcare at work, a shorter work week, better work-life balance and continuing education are at the top of the list of what employees demand from their employers, and companies are slow to catch up and adapt to the changing employee-employer dynamics . This partly explains why, even though the nationwide hiring rate is much higher than normal, companies across all industries are still left with millions of vacancies to fill.

What is Machine Learning?

Although often used interchangeably with AI (artificial intelligence), ML is more precisely a subset or application of AI. In simple terms, ML is the application of big data where machines (computers) use mathematical models to develop a new understanding without explicit instruction.

For example, image recognition is a commonly used application of ML. Image recognition allows computers to recognize and match faces (“tagging” posts on social media platforms) or identify cancerous growths on an X-ray.

ML is also widely used in the financial industry in what is known as statistical arbitrage: the use of algorithms to analyze securities in relation to established economic variables.

ML also allows computers to examine large data sets, identify and extrapolate causalities and correlations based on their predictions and probabilities. Predictive insights help to get the most out of data. Applications of this predictive ability can be found in real estate pricing, product development and other fields. Predictive analytics can also help job seekers and recruiters find better matches than they have before.

How does machine learning help with the US labor shortage?

The current labor shortage in the US coupled with the alarmingly high quit rate has shown us that there is a problem: workers are struggling to find a job that suits them.

Recruiters and job seekers are increasingly turning to advanced algorithms and big data statistical analysis to overcome this problem.

ML has the ability to analyze large amounts of data — in this case, employees who resign or are relieved of their duties versus those who persist or get promoted — and identify the common traits, characteristics, and skills. With this insight, recruiters can more quickly and accurately filter candidates who are unlikely to succeed in the position they are applying for. The result is a faster and smoother job search that is much more likely to lead to positive outcomes.

In addition to refining the matching process, ML has a positive influence on the speed and duration of the recruitment process. The inordinate amount of time a job seeker spends applying for a job and then applying for a job that he is unlikely to get or be happy with can only make the job seeker feel worse. When we face a crisis of unfilled vacancies and a high attrition rate, we need job seekers who are excited about the hiring process and not frustrated by it.

The evolution of the online job portal

Traditionally, an online job portal was where job seekers could view available jobs in their location or industry, read the various descriptions and requirements, and then take steps to apply. While that’s still a staple of today’s online job portals, the more successful ones go a few steps further.

By uploading a resume to an online job portal that uses ML, the job seeker can be directed and oriented to jobs that best suit their skills and experience.

However, ML can do even more than that. Having the required skills and experience is not enough to guarantee that the available position is a good fit. We must take into account the personality and priorities of the job seeker. ML can do that too. By having the job seeker fill out a questionnaire, personality test, or problem-solving tests incorporating gamification, the online job portal using ML gains valuable insight into how the job seeker thinks and what kind of company or position they are. more likely to be successful.

In a nutshell

In the US, there are millions more job openings than people looking for work. And the high hiring rate can barely keep up with the staggering number of workers leaving their jobs. Advances in ML allow computers to analyze large amounts of data to identify causation and correlation that can help recruiters and job seekers find matches that are more likely to succeed in both the short and long term.

Gergo Vari is founder and CEO of Lensa, Inc.