The Labor Market is a (Very Flawed) Matching Market
While the labor market is constantly changing and adjusting according to the current economy and supply and demand details, it has always been classified as a type of matching market. It’s a bipartite graph where one set of nodes consists of potential employers and the other set consists of job seekers. The sources I’ve referenced discuss the idea of improving the matching market so job seekers are matched to jobs based on merit and interest, which would increase profits- which makes complete sense, but what doesn’t make sense immediately is why the job market is suffering, even when we have this piece of common sense.
The fact is that the internet is a double-edged sword in the job search: things like LinkedIn and Handshake have made it incredibly easy to mass apply to a bunch of jobs or internships, but these same platforms have caused problems that weren’t as prevalent a few decades ago, when job searching was mostly in person and through things like newspapers. One thing is that with how easy it is to mass apply online, too many people are applying to too many jobs. In our matching market analogy, the set of company nodes becomes a restricted set as the amount of nodes that are job seekers increases, and prefer only certain companies over others. Companies must now spend extra time and resources going through an enormous amount of applications, when in the past, before the internet, people mostly applied based on their in-person networks and proximities to jobs. Also, with how easy it is to mass apply to things, people are applying to jobs where their skillsets don’t match up with the company’s needs. Again, time is wasted sifting through these applications when previously in time, it was time consuming for both companies and job seekers to play their roles in the job market, so job seekers placed greater care in doing research on jobs they’re most compatible with.
Of course, there is no lie that the internet has been instrumental to many people in finding new jobs quickly. However, more strides need to be taken so a more ideal perfect matching can be obtained in the job market, where everyone gets something they’re happy with. One aspect that has been made worse by the current job market is that low-wage, low-skill workers are put at a disadvantage, since they likely don’t have the same access or ability to use the newest technologies that make applying to jobs easy. Many companies have been utilizing AI that scans through resumes and tosses ones that don’t have certain keywords, and while these technologies are getting better, there is no doubt that they are still flawed. AI is always biased based on the data it trains on, and this historically means that people of disenfranchised groups (women, POC, veterans, etc) suffer. AI also is not good at predicting unexpected changes to the job market, like sudden crashes, sudden events that open/close many job openings, and the like.
One last point that deserves discussion is that the current class at Cornell (and this will likely apply to many future classes) is graduating into a recession. The set of nodes representing companies will become an even tighter restricted set, and it will be even more important for matching technologies to improve so that the right people are selected for the right jobs and worker retention improves. We also need to improve technologies so less bias remains in our AI tools and people of all socioeconomic standings can get matched to a job appropriate for them and the companies.
Sources:
How job-matching technologies can build a fairer and more efficient U.S. labor market
The value of search-and-matching models for the labor market