A Job Is Not Enough

Everybody who works in reentry says the same sentence. Get them a job.

I said it too. For a long time I believed it was the whole answer. Then I read the research, and the research says something more specific and more uncomfortable.

A job is not enough. The wage is the thing.

What the Data Actually Found

An economist named Kevin Schnepel tracked 1.7 million people released from California prisons between 1993 and 2008. He did not ask whether they got jobs. He asked what kind of jobs were available in their county on the day they walked out.

When construction and manufacturing openings went up at the moment of release, recidivism went down. Significantly. Measurably.

When food service and retail openings went up, nothing happened. No effect at all.

Sit with that. Those are the jobs we push people toward. Those are the jobs employers are most willing to give someone with a record. And in a study of 1.7 million people, they did not move the number.

The difference was money. In California, a person without any college who got hired into construction could expect more than $2,000 a month. In food service, about $1,000. That gap is the whole story. One wage lets you make rent, keep a car running, and stay. The other does not, and when it does not, people go back to what they know.

A separate study of more than four million people released across 43 states between 2000 and 2013 found the same pattern from the other direction. When wages for low-skilled men rose, reincarceration fell. The biggest effects, again, were in construction and manufacturing.

Two enormous datasets. Two different methods. Same answer.

Job quality is not a nice-to-have. It is the mechanism.

What the Trades Actually Pay

Here is the May 2024 federal wage data, straight from the Bureau of Labor Statistics.

Median for all American workers: $49,500.

Electricians: $62,350. Entry-level education, high school diploma. Training path, apprenticeship. Employment projected to grow 9 percent through 2034, with about 81,000 openings a year.

Plumbers, pipefitters, and steamfitters: $62,970. High school diploma. About 44,000 openings a year.

HVAC mechanics and installers: $59,810.

Welders, cutters, solderers, and brazers: $51,000. High school diploma. About 45,600 openings a year.

Electrical power-line installers: $92,560. High school diploma.

Elevator and escalator installers: $106,580. High school diploma.

Not one of those requires a four-year degree. Not one requires anything a person cannot start from a cell or a halfway house.

And the apprenticeship path is not a promise of future money. It is money now. You get paid while you learn. The Government Accountability Office reported that people who completed a registered apprenticeship earned average annual wages of about $80,000 in their first year after finishing, and roughly 90 percent stayed employed.

That is not a job. That is a career with a floor under it.

The Industry Needs Bodies

Associated Builders and Contractors ran their model in January. Construction needs to attract 349,000 net new workers in 2026. Then 456,000 more in 2027.

Their chief economist was blunt about why. It is not a boom. It is retirement. The people who built this country's infrastructure are aging out faster than anyone is replacing them.

So we have an industry with a structural shortage of workers, paying well above the national median, requiring no degree.

And we have millions of people who want to work and cannot get in the door.

Somebody explain to me why those two facts have not found each other.

It Works Inside, Too

RAND ran the largest meta-analysis ever done on correctional education. People who participated had 43 percent lower odds of returning to prison. For vocational training specifically, the odds of finding work after release were 28 percent higher than for people who got no training at all. Every dollar spent returned four to five dollars in avoided incarceration costs in the first three years.

Michigan built something called Vocational Village. Skilled trades training inside the walls, in a dedicated housing unit, with real equipment and industry certifications. The statewide three-year return rate for people released in 2020 was 22.7 percent. For Vocational Village graduates released the same year, it was 12 percent.

Alabama has its own version. J.F. Ingram State Technical College is the only technical college in the country whose entire student body is incarcerated or recently released. Welding. HVAC. Diesel mechanics. Nineteen certificate programs. Roughly 3,000 students enrolled in 2023.

The training exists. The pipeline exists. The demand exists.

So why is anybody still going back?

The Part Nobody Wants to Say

I am going to be honest about the weak spot in my own argument, because the numbers deserve better than a sales pitch.

When RAND revisited their analysis in 2018 using only the highest-quality study designs, the recidivism benefit held. The employment benefit did not. And across multiple randomized trials, employment-focused reentry programs reliably increase how many people are working and reliably fail to reduce how many people go back.

That looks like a contradiction until you line it up with Schnepel. The programs placed people. They placed them in transitional minimum-wage work. Minimum-wage work does not reduce recidivism. It never did.

The programs were not wrong to try. They were aiming at the wrong target.

There is a second problem. Even after someone earns the credential, the American Bar Association's inventory counts more than 27,000 state occupational licensing restrictions tied to criminal records. Some are blanket bans. Many are "good moral character" clauses, which is a phrase written to mean whatever the board wants it to mean on a given Tuesday. A man can weld for four years in prison and be told he is not of good character to weld for pay.

And there is a third problem, which is the one I actually built a company around.

The Last Hundred Feet

Here is the thing that gets skipped in every conversation about workforce development.

Between the certification and the paycheck, there is a room. There is a table. There is a man across it who has your file, and one of the lines in it is the worst thing you ever did.

Two federally funded audit studies found that a criminal record cuts the odds of a callback or a job offer by roughly half. A more recent field experiment put the felony penalty around 37 percent. Either way, half your shot is gone before you speak.

Which means the half you have left is the conversation. That is the entire margin. That is where it is won or lost.

And that is the exact moment nobody trains for.

We train the hands. We certify the skill. We build the resume. Then we send a person into the highest-stakes twenty minutes of their life with no preparation for the only question that actually matters, which is: what happened, and why should I believe it will not happen again?

Ask a general-purpose AI to help with that and it will tell them to write "personal sabbatical" on the resume. It teaches people to sound like a liar. A hiring manager can smell that across a desk.

We do the opposite. We help a person tell the truth in a way that competes. We let them practice against an interviewer that pushes back the way a real one will, out loud, in their own voice, until the room stops feeling like an ambush.

That is not a nice add-on to trade training. On the evidence, it is the difference between a certificate and a career.

One More Number

The Bureau of Justice Statistics and the Census Bureau tracked 51,500 people released from federal prison in 2010 for four full years.

A third of them never worked a single quarter.

But watch what happened to the ones who did. In the first quarter out, the biggest employers were temp agencies and food service. Four years later, temp work and food service had shrunk, and construction and manufacturing had grown.

People find their way to the trades. They do it slowly, on their own, with no map and no help, and it takes them four years.

Four years is too long. Four years is where the relapse happens. Four years is where the violation happens. Four years is where a person gives up.

We can shorten that. Not with a poster about opportunity. With training that is honest about what the person is walking into and rigorous about preparing them for it.

What I Am Asking

I am not asking anyone to lower a standard. The trades cannot lower standards and should not. Bad work on a panel or a pipe kills people.

I am asking for the standard to be applied to the person in front of you instead of the file behind them.

If you run a training program, stop measuring completions and start measuring placements and wages. A certificate that ends in a $12 an hour job is a certificate that did not work.

If you are an employer in the trades and you are short, the Federal Bonding Program has been running since 1966. It gives you a $5,000 fidelity bond, no cost, no deductible, covering the first six months. Ninety-nine percent of those bonds have never had a claim filed against them. The risk you are imagining has been insured against for sixty years.

If you run a reentry program, the pathway is not a job. It is a trade.

Seventy million Americans carry a record. The industry that pays best without a degree cannot find 349,000 people this year.

The math is not hard. The will is the hard part.

Sources

Schnepel, Kevin T. "Good Jobs and Recidivism." The Economic Journal, vol. 128, no. 608, 2018, pp. 447–469.

Schnepel, Kevin T. "Do post-prison job opportunities reduce recidivism?" IZA World of Labor, 2017. (Source of the California construction vs. food service monthly wage comparison.)

Yang, Crystal S. "Local Labor Markets and Criminal Recidivism." Journal of Public Economics, 2017.

U.S. Bureau of Labor Statistics, Occupational Outlook Handbook, May 2024 wage data: Electricians; Plumbers, Pipefitters, and Steamfitters; Welders, Cutters, Solderers, and Brazers; Heating, Air Conditioning, and Refrigeration Mechanics and Installers; Electrical Power-Line Installers and Repairers; Elevator and Escalator Installers and Repairers.

U.S. Government Accountability Office, GAO-25-107040, Apprenticeship: Earn-And-Learn Opportunities Can Benefit Workers and Employers, 2025.

Associated Builders and Contractors, 2026 Construction Workforce Shortage Analysis, January 2026.

Davis, Lois M., et al. Evaluating the Effectiveness of Correctional Education: A Meta-Analysis. RAND Corporation, 2013.

Bozick, Robert, et al. "Does Providing Inmates with Education Improve Postrelease Outcomes? A Meta-Analysis." RAND, 2018.

Michigan Department of Corrections, May 2025 population and recidivism report; Brennan Center for Justice, The Data Behind Prison Reform.

Bureau of Justice Statistics and U.S. Census Bureau, Employment of Persons Released from Federal Prison in 2010, NCJ 303147, December 2021.

Bureau of Justice Statistics, Recidivism of Prisoners Released in 34 States in 2012: A 5-Year Follow-Up Period, NCJ 255947, July 2021.

National Institute of Justice, "In Search of a Job: Criminal Records as Barriers to Employment" (Pager audit studies); Agan and Starr, Journal of Law and Economics, 2018.

National Employment Law Project, Unlicensed & Untapped, citing the American Bar Association National Inventory of Collateral Consequences of Conviction.

U.S. Department of Labor, Federal Bonding Program.

J.F. Ingram State Technical College; Alabama Reflector, July 2024.