NCCER projects that around 41% of the current construction workforce will retire by 2031. Most firms are treating that as a recruitment problem, which is the smaller half of it. The harder loss is undocumented judgment: the estimator who knows which subs pad a number, the superintendent who can tell a real delay from a soft one. Headcount can be replaced through hiring. That knowledge cannot, and almost nobody is capturing it.
There is a specific kind of meeting that happens when a long-serving superintendent retires. Someone realizes that the sequencing logic for a particular type of pour existed only in that person’s head, and that the written procedure describes about a third of what they actually did. The company then spends two years and a few hundred thousand dollars relearning it through mistakes.
That meeting is about to happen a lot more often. The demographic numbers are well known and get quoted at every industry conference. What gets far less attention is that the loss is not evenly distributed, and the most valuable part of it is the part that was never written down.
This piece is about that second problem, because the first one already has a thousand articles and a recruitment campaign attached.
What the 41% figure actually measures
A projection of exits, not a prediction of chaos
The National Center for Construction Education and Research projects that approximately 41% of the current construction workforce will retire by 2031. That figure counts people leaving, not positions disappearing, so it sits on top of ordinary turnover and new demand rather than replacing either.
The supporting data lines up. Around 22.7% of construction workers are already over 55, and the median age reached 42.5 years in 2024 against 40.5 seven years earlier. ABC’s occupational breakdown notes that roughly one in five electricians is over 55, which is uncomfortable timing given how much precision electrical work the next five years require.
The visible consequence is already showing up in delivery. AGC and NCCER’s workforce survey found shortages are now the leading cause of project delays, with 45% of contractors reporting at least one delayed project in the past year, and 92% saying they struggle to fill open positions.
The part that cannot be hired back
Judgment built over decades does not live in your document management system
A firm can replace a set of hands in a few months if the market cooperates. What takes twenty years to rebuild is the pattern recognition underneath a senior person’s decisions. The estimator who prices a job low because they know how that particular client handles change orders. The project director who can look at a three-week-old programme and tell you which activity is actually going to slip.
None of that is in the QMS. The written procedure captures the steps, not the exceptions, and the exceptions are where the value sits. The gap is well recognized in workforce planning terms, where decades of institutional knowledge can leave a business in a single quarter when a cohort retires together, and without structured succession the loss is permanent.
The cohort effect deserves emphasis. These people did not join individually across thirty years. Many joined in waves, which means they leave in waves, and a firm can lose four senior operators inside eighteen months without any single departure looking alarming on its own.
- Client behavior. Which owners dispute, which pay late, which change scope after award.
- Subcontractor reliability. Who performs under pressure and who quietly falls behind.
- Estimating instinct. Where the historic numbers mislead and by roughly how much.
- Failure patterns. The sequences and details that have caused rework before on this type of build.
Why almost nobody documents it
Awareness is high, action is not
The Chartered Institute of Building surveyed its global membership and found 76% were aware of the workforce challenges facing the sector, while only 24% were aware of those issues being taken into account on projects. That gap between knowing and doing is the whole story in one statistic.
The reasons are mundane. Knowledge capture has no project code to bill against, which makes it the first thing cut when the schedule tightens. The people who hold the knowledge are also the busiest, and asking them to spend Friday afternoons recording what they know competes directly with delivering the work. And there is a quieter dynamic nobody says out loud, which is that undocumented expertise is job security for the person holding it.
None of those obstacles are insurmountable. They just need a named owner and a budget line, which is exactly what a problem with a 2031 horizon rarely gets when there is a pour on Thursday.
What capturing it actually looks like
Structure beats good intentions, and the cheapest methods work fine
Phased retirement is the highest-leverage move available. Letting experienced people reduce hours while staying attached to complex projects keeps the knowledge on site during the handover period rather than ending it on a single Friday. It also gives the successor someone to ask, which is worth more than any document.
Structured mentoring works when it is paired and scheduled rather than encouraged in principle. Naming a specific successor for a specific person, with time protected for it, converts a vague cultural aspiration into something a project manager can be measured on. Recording debriefs after difficult jobs is the low-cost version, and a transcribed conversation about what went wrong on a project is more useful than any procedure document.
Technology helps at the margins rather than solving it. Reality capture, model-linked as-builts, and searchable project records preserve the what, and they are worth having. The why still needs a person willing to explain it. Firms that expect a software rollout to handle succession will find they have digitized the easy half. Our broader look at where tools actually earn their place is in what construction technology really covers.
Workforce, technology, and the economics underneath both
The weekly briefing for people making these calls, not commenting on them.
Join 3000+ Readers
Four capture methods, compared
| Method | What it preserves | Cost and effort |
|---|---|---|
| Phased retirement | Judgment, in context, during live work | Moderate salary cost, very high return |
| Named-pair mentoring | Client and subcontractor knowledge, decision logic | Low cost, needs protected time to work |
| Recorded project debriefs | Failure patterns and exceptions | Very low cost, needs someone to index them |
| Reality capture and digital as-builts | The physical record, not the reasoning | Capital cost, only solves part of the problem |
It is an NCCER projection, widely cited by ABC, AGC, and industry associations, that approximately 41% of the current US construction workforce will retire by 2031. It measures people exiting rather than roles being eliminated, so it compounds with normal turnover and with demand growth from infrastructure and data center work. (Source)
ABC projects the US construction industry must attract around 349,000 net new workers in 2026 to stay in labor equilibrium, rising to 456,000 in 2027. Most of the 2026 requirement comes from retirements rather than expanding demand. (Source)
No. The European Union saw the share of construction workers over 50 rise from 24.7% in 2011 to 31.5% in 2018, and CIOB’s global membership survey found high awareness of the issue paired with very low incorporation of it into project planning. The demographics differ by market but the direction is consistent. (Source)
Record structured debriefs after difficult projects and index them so they are findable. A transcribed conversation about what went wrong, why it was not caught earlier, and what the person would do differently costs almost nothing and captures the exceptions that formal procedures leave out.
It addresses output, not expertise. Robotics and AI can help maintain productivity with fewer people on repetitive scopes, but they do not replace the judgment held by senior estimators, superintendents, and project directors. Treating automation as a succession plan leaves the harder half of the problem untouched. (Source)
Workforce shortages are now the leading cause of project delays according to the AGC and NCCER survey, with 45% of contractors reporting at least one delayed project in the past year and 92% reporting difficulty filling open positions. (Source)
Related reading
Will AI Replace Construction Workers?
What Is Construction Technology (ConTech)?
The Future of Construction Software: Are We Stuck with Legacy Systems?
NCCER, Apprenticeships and the Construction Workforce Gap
Westside Construction Group, ABC and AGC Workforce Data 2026
Fieldwire by Hilti, Tackling Labor Gaps in an Aging Workforce (BLS data)
HammerTech, Navigating Construction’s Age Gap (CIOB survey)
Acuity International, Workforce Planning and Knowledge Transfer
