Switzerland offers a rare, high-resolution look at what continuous training, and more precisely work-related continuous training, actually does to people’s jobs and paychecks. Linking a national training census to register data on income and unemployment (2014–2018), researchers estimate that adults who took job-related, non-formal courses saw average annual earnings rise about 3.4% in the following years and faced a 2.1-percentage-point lower chance of becoming unemployed.
Those gains aren’t uniform. Lower-income workers captured the largest pay boost, while higher-income workers saw the sharpest drop in unemployment risk. The result: a double benefit: more earnings at the bottom, more security at the top.
In this article I will go deeper into the study (by Stefan Denzler, Jens Ruhose and Stefan C. Wolter) itself and also focus on the sustainable elements in it.
What the study examined – and why it matters
Switzerland is a useful bellwether: high training participation, a liberal labour market, and continuous training mostly organized and funded by firms and individuals rather than the state. Those conditions mirror many Anglo-Saxon settings more than continental or Nordic models.
But what counts as continuous training here? Non-formal, work-related learning – such as courses, seminars, conferences – taken for professional reasons within the last 12 months. In the 2016 micro-census, 67% of employed adults reported participating; they took a median of 2 courses, totaling a median 24 hours (mean 48), and 78% had employer funding.
Why we can trust the effects. The authors combine (1) Switzerland’s Micro-census on Education and Training (MET), (2) social-security register income data, and (3) unemployment-insurance records. They deploy a matched difference-in-differences estimator with entropy balancing, explicitly aligning pre-treatment income and unemployment trajectories (2014–2015) plus education, occupation, industry, demographics, and region. That design helps isolate changes attributable to continuous training rather than pre-existing momentum.
Main labour-market impacts (post-training average)
These are longitudinal estimates comparing the same individuals over time against a matched, reweighted comparison group, rather than simple cross-sectional differences.
| Outcome | Estimated effect | Interpretation |
|---|---|---|
| Annual earnings (log) | +0.034 | ≈ +3.4% average earnings increase after training. |
| Unemployment risk | –0.021 | –2.1 percentage points vs. a ~3.7% baseline (2015), roughly halving risk. |
| Unemployment duration | –0.136 months | About 4 days shorter time unemployed among those who do become unemployed. |
Who benefits, and how?
The average effect of continuous training hides different payoffs across the income ladder. The bullets I give below should be treated as a quick map of where training converts into higher pay versus lower risk.
- Bottom tercile (mean CHF ~28,649): earnings effect ≈ +0.080 (~+8%). Unemployment-risk change is near zero. The payoff here is income growth.
- Middle tercile (mean CHF ~68,345): earnings ≈ +0.023; unemployment-risk change is small. Mixed but positive.
- Top tercile (mean CHF ~137,492): earnings ≈ +0.003; unemployment risk ~–0.029. The payoff here is job security more than pay.
Education and occupation matter. Effects of continuous training are strongest among workers with vocational qualifications and among unskilled workers. People with general academic tracks show smaller earnings gains. That tallies with the idea that shorter, targeted courses may nudge vocationally anchored careers sooner than they move generalist trajectories.
Gains tend to be a bit higher for younger workers (20–39) on earnings; unemployment-duration cuts appear larger for men and for older cohorts in some splits, though many subgroup estimates come with wider uncertainty bands.
Little difference emerges between self-funded and employer-funded continuous training in terms of average earnings effects and unemployment risk. That suggests motivation and selection dynamics on both sides may net out in aggregate.
Splitting by total training hours (above/below median) yields similar average effects in this two-year follow-up window – evidence that relatively short, well-aimed courses can move the needle.
Switzerland’s continuous training ecosystem, distilled
In systems where employers share costs and workers co-invest time, continuous training becomes a practical labour-market instrument – responsive to technology shifts and firm-specific needs. The Swiss pattern shows how a privately organized CET (Continuous Education and Training) market can still deliver broad, measurable returns.
| Metric | Value |
|---|---|
| Participated in work-related non-formal training | 67% of employed respondents |
| Courses per participant | 2.6 average; 2 median |
| Total hours per participant | 48 average; 24 median |
| Employer financed | 78% of participants |
| Context | High participation (also 58% job-related non-formal training in AES 2016 across 25–64), liberal labour market, training mostly private. |
The design matters because trainees already differ from non-trainees. The study offers several solutions to neutralize those gaps.
The study uses a regression-adjusted difference-in-differences framework and entropy balancing to construct a comparison group that mirrors trainees on observed characteristics and on pre-treatment income/unemployment trends. That reduces bias from “go-getter” selection or prior training streaks.
The data spine:
- MET 2016 (training events, demographics, edu/occupation).
- CCO social-security register (annual earnings 2014–2018).
- SECO unemployment spells (monthly, aggregated yearly).
Note that the income data are yearly totals (no hourly wages), hours worked observed only in 2016 (full- vs part-time), and training before early 2015 isn’t directly observed – hence the emphasis on balancing 2014–2015 trajectories to absorb earlier differences. The authors also flag that some effects could be lower-bound if 2016 training partially “shows up” in conditioning variables.
Here’s how the results translate on the ground:
- For workers at the bottom: an ~8% lift in yearly earnings after taking work-related courses is material. It can offset stagnation, help weather inflation, and reduce the need for job-hopping.
- For high earners: wage jumps are small, but risk insurance is real. A roughly 0.029 drop in the probability of becoming unemployed reduces exposure during downturns or restructurings.
- For employers: continuous training investments correlate with lower turnover risk among higher-paid staff and faster productivity payoffs among lower-paid staff, even when course hours are modest.
Sustainability lens: where the green threads run
Continuous training effects connect directly to several SDG-aligned outcomes as I will show you now.
Economic sustainability (SDG 8: Decent Work & Growth). Cutting unemployment risk by ~2.1 points and trimming time spent unemployed reduces frictional costs, stabilizes household income, and eases pressure on social insurance. That’s macro-resilience, particularly relevant in transitions toward digital and clean-tech economies.
Social sustainability and equity. continuous training’s largest earnings impact lands with lower-income workers. That narrows gaps from the bottom up and supports mobility without waiting for lengthy degree programs.
Human-capital durability (SDG 4: Lifelong Learning). Short, non-formal courses that move real outcomes demonstrate a scalable format for keeping skills current while people stay employed. The Swiss model – heavy employer participation – shows a pathway to shared stewardship of skills without over-reliance on public budgets.
Environmental transition readiness. While the paper doesn’t estimate green-skill outcomes directly, the architecture – rapid, job-related upskilling with measurable pay and employment effects – matches what firms need to shift processes (energy efficiency, circularity) without displacing workers. In short: continuous training is a just-transition tool when it helps people adapt in place.
Courses taken purely for leisure show small, non-robust effects on pay and unemployment. The strong results concentrate among training pursued for work. That aligns with intuition: vocational purpose connects content to immediate productivity and labour-market signals.
What this study doesn’t do
Keep them in mind that one shouldn’t generalize or port the results abroad.
- No hourly wage decomposition. Annual income can rise through higher hourly pay or more hours worked. The dataset can’t split those channels beyond a full-time vs part-time flag in 2016.
- Short follow-up horizon. Effects are tracked up to two years after training. Long-run returns (or fade-outs) remain outside the window.
- Prior training not fully observed. Pre-2015 learning is not directly captured, so the design controls away differences via pre-trends. Residual non-linearities could remain, though the approach is conservative and may understate some effects.
Practical takeaways for policy and practice
As for policy-makers, the study also offers various concrete actions one can budget, procure, or negotiate:
- Back short, targeted, job-related courses. They deliver measurable outcomes fast, especially for vocational and lower-income workers.
- Encourage co-financing. Employer funding doesn’t dilute returns; effects look similar to self-funded cases, supporting shared cost models.
- Track pre-trends. Any CET evaluation should match on prior income and unemployment paths, not just demographics. The Swiss design shows how.
- Use CET in just-transition strategies. When industries need to decarbonize or digitize, tie support to job-related training that protects employment while upgrading skills.
Editorial verdict
A tight, register-linked design shows that work-related continuous education in Switzerland pays off, and rather fast too, in different ways along the income ladder. For low earners, it’s a pay lever. For high earners, it’s job insurance. For employers and policymakers steering through digital and green transitions, it’s a pragmatic instrument that raises productivity without leaving people behind.
