Where work and organizational psychology is actually bought

Three warnings that apply throughout

The eleven

1. Hiring demand for the function that employs most A&O people is below its 2020 level in both target markets. US HR postings index 94.7, German 82.1, against all-sector aggregates of 101.8 and 103.8. German HR postings are 65.8% below their May 2022 peak; the whole German market is only 3.5% above baseline. This is a contraction relative to 2020, not merely slower growth. A
2. The regulated occupation is numerically trivial and static. BLS projects US I-O psychology to add 300 jobs over ten years, with 400 total projected openings. BLS once projected this occupation as the fastest-growing in the entire US economy at +53.4% (2012–22 vintage); the current vintage says +6% — a 47-point downgrade. O*NET's “Bright Outlook” flag survives only on the percentage criterion. A C
3. The strongest paid demand is where a regulator, not a CHRO, is the buyer. De Nederlandsche Bank staffs organisational psychologists itself and has run 52–54 culture examinations since 2010; APRA imposed capital add-ons of AUD 500m each on NAB, Westpac and ANZ and USD 1bn on CBA; ECB governance-focused on-site inspections rose 22 → 31 between 2024 and 2025. These buyers are price-insensitive because the alternative is regulatory capital. A
4. The EU/US structural asymmetry is the single most important fact in the dossier, and it has not changed. Directive 89/391/EEC Art. 6(3)(a) obliges employers to assess all risks — read everywhere to include psychosocial risk, and made explicit in Germany by ArbSchG § 5 Abs. 3 Nr. 6. The US has no equivalent duty; OSHA's General Duty Clause has never been established for stress or burnout. EU psychosocial work is compliance spend with an inspectorate and a works council behind it; US psychosocial work is discretionary HR spend competing with every other line item. A
5. The 2025–26 deregulation wave removed reporting and safe harbours, not substantive prohibitions. CSRD scope cut from 250 employees/€50m to 1,000/€450m, ESRS datapoints cut over 70%; the AI Act's high-risk obligations deferred to 2 December 2027; its AI-literacy duty weakened from “ensure” to “support the development of”; Colorado's AI Act repealed; the EEOC's Uniform Guidelines in the process of rescission. But Art. 5(1)(f)'s emotion-recognition ban, Art. 4(4) of the pay directive, ArbSchG § 5(3) Nr. 6 and the BetrVG co-determination rights all stand untouched. Demand relocated more than it fell — into legal, works-council and risk budgets, which are smaller, more adversarial, and concentrated in fewer hands. A B
6. The strongest single EU driver is a job-evaluation mandate that nobody markets as psychology. Directive (EU) 2023/970 Art. 4(4) requires pay structures assessable on “objective, gender-neutral criteria” including skills, effort, responsibility and working conditions — the classic Genfer Schema factors. That is analytical job evaluation, i.e. work analysis: the oldest and least substitutable A&O competence. First reports are due 7 June 2027, and transposition failed almost everywhere (Germany has no published draft; Sweden withdrew its draft in March 2026), which compresses the build window rather than removing it. A B
7. Germany just named A&O psychology as a recognised route to a statutory safety role — the first crack in a structural weakness. Revised DGUV Vorschrift 2, § 4 Abs. 6: “Personen mit einem Studienabschluss in Physik, Chemie, Biologie, Humanmedizin, Ergonomie, Arbeits- und Organisationspsychologie, Arbeitshygiene oder Arbeitswissenschaft erfüllen als gleichwertig qualifizierte Personen…” — generic “Psychologie” does not appear. This converts psychological expertise into headcount an employer is legally required to buy under § 5 ASiG. Four cumulative hurdles apply, and diffusion depends on sixteen Länder authorities. A B
8. The cleanest funded mandate anywhere is a German works-council provision. BetrVG § 80 Abs. 3 Satz 2 (in force 18 June 2021): where a works council must assess the introduction or use of artificial intelligence, engaging an external expert is deemed necessary — only the person, scope and fee remain negotiable, and the employer pays under § 40 Abs. 1. It does not depend on the system being “high-risk,” so it is unaffected by the AI Act deferral. § 91 is separately notable as the only place in German law where “gesicherte arbeitswissenschaftliche Erkenntnisse” is the operative legal test — making the discipline itself the standard. A B
9. The best-evidenced growth in the problem is manager-specific, and it is recent. Gallup manager engagement: 30% (2023) → 27% (2024) → 22% (2025), while individual contributors were flat at 18 / 18 / 19%. Within 2024, managers under 35 fell 5 points and female managers 7. That divergence is the strongest differential in the whole dataset — though it rests on one vendor's instrument over three years. A
10. The best-evidenced decline is US diversity work, and it is largely already realised. Russell 3000 DEI headcount peaked near 12,300 in mid-2022 and stood near 10,800 in 2025; only 7.4% of departing DEI professionals moved to another DEI role. Executive Order 14173 revoked EO 11246, and the implementing regulations were rescinded at 91 FR 54444 (21 Aug 2026, effective 26 Oct 2026) — removing the compulsory floor under US diversity analytics. Board-committee DEI oversight nonetheless rose year over year, so the work is being renamed and relocated, not uniformly abolished. A B
11. The credential confers almost no market exclusion where the volume is. Coaching, BGM, Organisationsberatung, psychosocial risk assessment and personnel selection are unregulated in Germany; “Psychologe” is not a state-licensed title, and only psychotherapy is a reserved activity. Coach practitioners worldwide reached 122,974 in 2025 on $5.34bn revenue — but average revenue per coach fell from $52,800 (2022) to $49,283 (2025) in nominal terms. Regulation in this field reliably creates demand for a service and almost never for a credential. A B

1.1 Why demand for A&O psychology is close to unmeasurable

Neither the US nor the German occupational statistics can see this profession. Both failures are structural, not incidental.

United States. SOC 19-3032 “Industrial-Organizational Psychologists” has an OEWS-measured employment base that ranges from 610 to 1,280 across six consecutive years with relative standard errors of 6.4–14.2%. In May 2025 the estimate was 790 people — 0.00051% of US employment. Median annual wage moved −25.5% (2023→2024) then +76.6% (2024→2025). Any statement built on that series is inside the noise band. A B

Worse, BLS's own two systems disagree by a factor of five. The Employment Projections base for 19-3032 in 2024 is 5,600; the OEWS survey estimate for the same year is 1,050. The cause is documented — the National Employment Matrix adds self-employed, agricultural and private-household workers — but the practical consequence is that the two series must never be mixed, and the EP base itself rose from 1,600 (2012 vintage) to 5,600 (2024 vintage), which is a scope revision and not growth. A

Germany. The relevant code is KldB 2010 Berufsgruppe 816 “Psychologie, nichtärztliche Psychotherapie” — which sits under Berufshauptgruppe 81 “Medizinische Gesundheitsberufe”. An A&O psychologist is, administratively, a health professional. The one sub-code that would isolate the field, 8161 “Berufe in der nichtklinischen Psychologie,” is not published for any recent year. And most A&O psychologists are coded outside 816 altogether — as Personalreferent, HR Business Partner, Organisationsberater, Personalentwickler — landing in KldB 715 (Personalwesen) or 73 (Unternehmensorganisation). A B

The consequence for this dossier: the German market must be triangulated from obligation and cost of problem, not from occupational employment statistics. Where I do that, I say so.

1.2 What the statistics that do work show

The adjacent occupations — where A&O-trained people actually get coded — are large, precisely measured, and grew strongly from 2020, with a notable 2024→25 reversal in the largest.

Forward, the same occupations are projected at +5.0% (HR managers), +6.2% (HR specialists), +5.3% (comp/benefits) and +10.8% (T&D specialists) for 2024–34 — around and slightly above the +3.1% all-occupations rate, and nothing like the 2020–25 realised growth. C

Supply is flat, not surging. US research doctorates in industrial-organizational psychology: 183 (2019) → 200 (2024), inside a fifteen-year range of 169–222. There is no post-2020 inflection. Psychology bachelor's degrees rose +8.0% and master's +13.1% between 2019-20 and 2021-22. A

1.3 Job postings: the hardest available evidence, and it points down

Indeed Hiring Lab's open dataset is seasonally adjusted and indexed to 1 February 2020 = 100. It covers one platform, assigns sectors algorithmically from job text, and cannot tell you absolute job counts — but it is a genuine, daily, comparable 2020→2026 series, which is more than any official statistic offers for this field.

Three things follow, and none of them is comfortable.

• Within psychology, German demand is clinical. Therapy postings hit an all-time series high (191.6) on the last day of data; Community & Social Service sits at 202.8; nursing at 156.1. HR sits at 82.1. Nine of the top ten Psychologe vacancies in the German federal job database are clinical, therapeutic or rehabilitation posts. A
• The standard recommended pivot points into the weakest labour market in the dataset. Data & Analytics is the worst-performing of all 44 US sectors at 60.6 — down 39% on 2020 and 70% from peak — and sits at 75.5 in Germany. “Move into people analytics” is advice to enter a contracting market. A B
• Narrow, unambiguous A&O search terms return single- and double-digit counts in the entire German federal job database (22 Aug 2026): Arbeits- und Organisationspsychologie 14 — the top-ranked hit was a caretaker vacancy at an A&O institute — Eignungsdiagnostik 21, Arbeitspsychologe 32, Wirtschaftspsychologe 39. The broad terms inflate badly: Organisationsentwickler returns 499 results that are visibly almost all Business Analyst posts, and Betriebliches Gesundheitsmanagement returns 13,855 because the engine fuzzy-matches any posting containing Gesundheit or Management. Treat the big numbers as noise-inflated ceilings and the small ones as the real signal. A B

1.4 What genuinely emerged since 2020

Distinguishing new fields from new vocabulary matters here. The following have a legal instrument, a named buyer, or a measured problem behind them — not just a label.

1.5 What lost importance — and which claimed declines are unevidenced

1.6 New and intensifying problems, graded by evidence strength

1.7 What is not attributable to COVID

This was one of your explicit questions, and it produces a cleaner answer than the COVID story does. Ranked by evidence strength:

1. Demographic contraction of the working-age population. Destatis' 16th coordinated projection (December 2025): the population aged 20–66 falls from 51.2 million (2024) to 37.1–45.3 million by 2070 in every scenario. The 67-plus share rises from 20% to 25% by 2035, peaking at 25–27% around 2038. Old-age dependency rises from 33 per 100 working-age to 43–61. This is the only driver here that is official, monotone, entirely COVID-independent, and maps directly onto billable work — retention, age-appropriate work design, Workability, knowledge transfer. It is also slow: nothing creates a step change in 2026–28, so it is a structural argument, not an urgent one. A C
2. AI-driven intensification and monitoring in large firms. The OECD figures in §1.6, concentrated in exactly the segment that buys consulting: AI adoption among enterprises with 250+ employees is 55% (EU 2025), 57% (Germany), 37% (US). A
3. The German industrial recession. Price-adjusted GDP: −0.9% (2023), −0.5% (2024), +0.2% (2025). In 2025 manufacturing fell 1.3% with automotive and machinery named as loss-making, construction −3.6%, and — the most directly relevant line item in this entire dossier — business services −0.8%, which is where management and HR consulting sit. Corporate insolvencies reached 24,064 in 2025, up 10.3% and the highest since 2014, with Jan–May 2026 up a further 4.9%. The ifo automotive business-climate index stood at −21.4 in June 2026. Announced German industrial cuts: VW 35,000 by 2030, Bosch ~13,000 in Mobility by 2030, Continental 7,150 plus 3,000 in R&D, Audi 7,500, Siemens Digital Industries 5,600. A
4. Labour-market loosening. Covered in §1.5. Tightness peaked in 2022 and has declined for four consecutive years on both sides of the Atlantic. A
5. Regulatory reconfiguration. All of Part 3. Note that the largest single items — the UGESP rescission, the EO 11246 revocation, the CSRD cuts, the AI Act deferral — postdate 2024 and have nothing to do with the pandemic. A
6. Entry-level labour-market dislocation, cause contested. Brynjolfsson, Chandar & Chen (Stanford Digital Economy Lab, ADP payroll microdata through June 2026): employment for workers aged 22–25 in AI-exposed occupations is 19% below where it would be had it tracked less-exposed peers, driven by reduced hiring rather than increased separations, with divergence widening since August 2025 — and the authors explicitly call these “early, descriptive indicators … rather than causal estimates.” The NY Fed puts recent-graduate unemployment at 5.6% and underemployment at 42% (Q2 2026). SignalFire finds new-grad hiring down 25% versus 2023 and more than 50% versus 2019, and attributes it primarily to the end of cheap capital, not AI. Both explanations are quantified; neither is settled. A B
7. Skills-based hiring largely did not happen. Burning Glass Institute × Harvard Business School (Feb 2024): only 3.6% of roles removed degree requirements over 2014–2023, and where removed, non-degreed hiring rose 3.5 percentage points — a net effect across all hiring of 0.14 percentage points, roughly 97,000 incremental non-degreed hires out of 77 million, or “not even 1 in 700.” 45% of firms that announced skills-based hiring were “in name only” with no measurable behavioural change. A

2.1 Sectors with the strongest evidence of paid demand

2.2 Sectors where the demand claim is weak or unsupported

2.3 Adjacent market sizes, with reliability stated

3.1 Rollbacks and headwinds, with dates

This is the half of the regulatory picture most often omitted. Every item below is a confirmed reversal, deferral or narrowing, not a risk.

3.2 Is an EU psychosocial risks directive coming?

There is no EU directive specifically on psychosocial risks, and Directive 89/391/EEC covers them only indirectly — it does not mention psychosocial risks or work-related stress. A

The trajectory of the demand for one is instructive. The European Parliament asked for it in its resolution of 5 July 2022 on mental health in the digital world of work, § 30 of which “calls on the Commission … to propose a legislative initiative … on the management of psychosocial risks and well-being at work.” The Commission's June 2023 mental-health Communication — 20 flagship initiatives, €1.23 billion across instruments — referenced “a possible future EU initiative on psycho-social risks at work.” A

By late 2025 that demand had been converted into something smaller. The Quality Jobs Roadmap, COM(2025) 944 of 4 December 2025, does not propose a standalone directive. Instead the Commission commits to “review the Workplace and the Display Screen Equipment Directives … in particular, to better address psychosocial and ergonomic risks,” to report on OSH Framework Directive implementation by mid-2026, and to run “a Healthy Workplace campaign on mental health and psychosocial risks at work” in 2026–2028. The second-stage social-partner consultation on the Quality Jobs Act (20 July 2026) confirms the same approach, with trade unions still demanding binding psychosocial legislation. A

The soft-law channel is nonetheless real: EU-OSHA's Healthy Workplaces Campaign 2026–2028, “Together for mental health at work,” launches in October 2026 with the European Week for Safety and Health at Work on 19–24 October. Its stated objectives include promoting the assessment of psychosocial risks, with priority on health and social care. Campaigns of this kind reliably drive national labour-inspectorate programming and Berufsgenossenschaft campaign budgets for their duration — a three-year tailwind, budget-mediated rather than legally binding. A B

4.1 Inside Europe: the differences that matter

The ESENER 2024 country data is the single most commercially significant dataset in this dossier, because it measures the thing the market analysis actually turns on.

Right to disconnect is a genuinely divergent map. Pre-pandemic adopters: Belgium, France, Italy, Spain. Post-pandemic: Greece (Law 4808/2021), Luxembourg, Portugal (Law 83/2021), Slovakia (Act 76/2021). The number of national laws including a right to disconnect doubled during the pandemic. The Nordics and most of central and eastern Europe have no explicit statutory right. And effectiveness is poor where it exists: in 2021, 60% of French teleworking employees still had no formal right to disconnect, and only one in three Spanish workers said their company recognised it. A

Primary and official sources carry the argument throughout: BLS (OEWS and Employment Projections), NSF NCSES Survey of Earned Doctorates, NCES, Eurostat, EU-OSHA (ESENER, OSH Pulse), Eurofound (EWCS), Destatis, Bundesagentur für Arbeit, BAuA, GDA, GKV-Spitzenverband, DGUV, Deutsche Rentenversicherung, EUR-Lex, the Federal Register and eCFR, reginfo.gov, the European Parliament's OEIL procedure files, ECB and DNB supervisory publications, EASA, and company annual reports and 10-Ks. Peer-reviewed sources include Sackett et al. (2022), Bloom, Han & Liang (Nature 2024), Han/Shanafelt/Sinsky et al. (2019), Bisbee et al. and the Chisholm et al. (2016) paper behind the WHO/ILO figure.

Where evidence comes from an association, an insurer, a consultancy or an analyst house, that is stated inline and the reliability is assessed rather than assumed. No blogs, SEO articles or opinion pieces are used as central evidence. Law-firm alerts appear only to corroborate a legal fact already established from the instrument or the register.

Known limitations of this dossier

• Access constraints. Several primary hosts were unreachable during compilation — baua.de and gda-portal.de returned HTTP 403 on all content pages, gesetze-im-internet.de was blocked entirely, and Destatis GENESIS and LinkedIn are robots-disallowed. Where a German statutory text is quoted, it is reproduced from an official or semi-official secondary (GDA-Portal, VBG, BASI, bag-urteil.com, sozialgesetzbuch-sgb.de) and should be re-checked against the official portal before external publication.
• Two figures rest on stale reference years and are flagged as such: the German Produktionsausfallkosten set (€75bn production loss / €133bn gross-value-added loss, psychological disorders at 16.2%) is 2016 reference-year data and must never be presented as current. The current defensible German figures are 147.3 million AU-days from psychological disorders in 2024, 16.7% of all AU days, and 42% of Erwerbsminderungsrenten.
• One German market-sizing figure needs a caveat. The GDA Betriebs- und Beschäftigtenbefragung 2023/24 (n = 3,817 establishments plus 3,824 employees, fieldwork Sept 2023 – mid 2024, published 24 June 2025, reported here via BMAS and DGUV) gives 68% of establishments conducting any Gefährdungsbeurteilung, up from 52% in 2015, with psychische Belastung included in 65% of those cases — from which I derive roughly 44% of all German establishments, and a residual non-compliance of about 56% thirteen years after the obligation. That 44% is my derivation, not a published figure, and the base matters enormously: “any consideration of psychosocial factors” is a much weaker test than “complete and documented psychosocial risk assessment,” for which figures around 20–25% are commonly quoted. Establishment-weighted counts are also dominated by micro-firms, and employee-weighted shares run far higher. Any market sizing must state which of the three bases it uses. Also note there was no 2019/2020 GDA wave — the valid comparison is 2015 versus 2023/24.
• Series breaks to respect: the 2018 SOC revision means there is no May 2020 figure for SOC 19-3033; OEWS is a three-year pooled model-based estimate; the 2020–30 BLS projection vintage is explicitly COVID-distorted by its low base year; EWCS switched survey mode between 2021 and 2024; DAK switched to a Leitdiagnose method from 2024 and recalculated 2023; Eurostat's telework series breaks between 2020 and 2021; SED changed its field taxonomy between vintages; and Microsoft's Work Trend Index changed its sampled population between 2025 and 2026.
• The May 2025 OEWS release was delayed to 15 May 2026 by the federal government shutdown of 1 October – 12 November 2025. BLS states response rates stayed within the normal range and no methodology changes were needed.

A · Empirically supported developments

• HR job postings are below their February 2020 level in both the US (94.7) and Germany (82.1), against all-sector aggregates above baseline. German HR postings are 65.8% below their May 2022 peak.
• Registered unemployment in German Berufsgruppe 816 rose from 2,300 (2019) to 4,500 (2025), +96%, with a vacancy stock flat near 900 and a 4.1% occupation-specific unemployment rate. Psychology appears in neither the 2024 nor the 2025 Fachkräfteengpassanalyse.
• US I-O psychologist employment is between 790 (OEWS) and 5,600 (EP basis); projected to add 300 jobs by 2034; and BLS revised its own projection for this occupation from +53.4% (2012–22 vintage) to +6% (2024–34).
• Adjacent US occupations grew 22–44% from 2020 to 2025 and are projected at 5–11% for 2024–34. HR Specialists fell 0.5% between May 2024 and May 2025 — the only decline in the set.
• US I-O doctorate production is flat: 183 (2019) → 200 (2024), inside a fifteen-year range of 169–222.
• Only 29% of CFOs plan to increase HR budgets for 2026; 22% expect cuts; average functional budget growth falls from 2.4% to 0.7%.
• Gallup manager engagement fell 30% → 27% → 22% (2023–2025) while individual contributors stayed flat at 18/18/19.
• German retention intent (one-year horizon) fell from 73% (2019) to 50% (2024), recovering to 56% (2025).
• ESENER 2024: 21% of EU establishments use a psychologist; Finland 73%; 39% of establishments with 20+ employees have an action plan against work-related stress, up from 33% in 2019.
• Psychological disorders accounted for 147.3 million German AU-days in 2024 (16.7% of all) and 42% of Erwerbsminderungsrenten, while total absence days fell from the 2022 peak.
• The AI Act's employment high-risk obligations were deferred to 2 December 2027 and its AI-literacy duty weakened, by Regulation (EU) 2026/1744 in force 27 July 2026.
• CSRD scope rose to 1,000 employees and €450m; ESRS datapoints were cut over 70%; the CSDDD threshold rose to 5,000 and €1.5bn.
• 29 CFR Part 1607 is still in force, with rescission at Final Rule stage (RIN 3046-AB43), no NPRM stage, and final action projected November 2026.
• Sackett et al. (2022) lowered every major selection-method validity estimate, cognitive ability furthest (.51 → .31).
• Coach practitioners reached 122,974 worldwide in 2025 with average revenue per coach down from $52,800 (2022) to $49,283.
• Only 3.6% of US roles removed degree requirements over 2014–2023, for a net effect of 0.14 percentage points across all hiring.
• A randomised trial found two-day hybrid work cut attrition by roughly a third with null effects on performance grades and promotions over two years.

B · Plausible readings of the data

• There are two A&O markets, not one — discretionary work sold to HR is contracting; obligated work sold to regulators, works councils, safety functions and operational P&Ls is holding or growing. This is the organising claim of the dossier and it is an inference, though a well-supported one.
• The 2025–26 deregulation wave removed reporting and safe harbours rather than substantive prohibitions, so demand relocated to legal, works-council and risk budgets rather than disappearing.
• Roughly 56% of German establishments still do not carry out a psychosocial risk assessment — my derivation from GDA 2023/24, base-dependent as noted in Part 5.
• Regulation creates demand for a service and almost never for a credential, because every commercially valuable A&O activity in Germany sits in the unregulated zone. The DGUV Vorschrift 2 change is the first development that attacks this.
• ATD's numbers read as “less learning delivered at a higher unit price,” consistent with a shift toward expensive facilitated work — but not growth in volume.
• The defensible global core of the psychometric market is plausibly $3–6bn — smaller than Korn Ferry alone — with the ten-fold spread in published estimates being a definitional artefact.
• The Nordic paradox implies the market is created by regulation that makes employers able to see psychosocial problems, not merely obliged to address them.
• “45% in name only” on skills-based hiring is the base rate to apply to any client's announced people-strategy initiative.
• Synthetic LLM respondents are adequate for descriptive means and inadequate for conditional and moderator inference — a genuine technical moat for survey-based organizational research.

C · Forecasts and scenarios

• BLS projects US I-O psychology at +6% and 400 total openings for 2024–34; the 2025–35 vintage was not yet released as of this compilation.
• Destatis projects the German working-age population falling from 51.2 million (2024) to 37.1–45.3 million by 2070 in every scenario.
• OECD projects mental ill-health reducing EU GDP by an average 1.7% a year between 2025 and 2050, at roughly €76bn annually — a different construct from the older “up to 4% of GDP” cost-of-illness figure, and not a continuation of it.
• The AI Act's 2027 deadline is likely to compress procurement into the first half of 2027.
• A standalone EU psychosocial risks directive before 2028 looks unlikely on the current evidence; the 5 October 2026 plenary sitting on procedure 2026/2023(INL) is nonetheless a datable catalyst.
• DGUV Vorschrift 2 diffusion depends on sixteen Länder authorities' willingness to issue Einzelfallgenehmigungen — expect slow and uneven uptake.
• UGESP rescission is likely net strongly negative on validation volume and mildly positive on rate for a small litigation-expert elite, with professional-standards bodies becoming the de facto authority on validity norms.
• Sweden's withdrawal of its pay-transparency draft is the first sign of substantive retreat on Directive 2023/970 and is worth monitoring.

Official labour-market and employment statistics

• BLS Occupational Outlook Handbook — Psychologists
• BLS Occupational Employment and Wage Statistics (OEWS), May 2019–May 2025 vintages
• BLS Employment Projections 2024–2034 and the 2012–2022 vintage
• O*NET 19-3032 Industrial-Organizational Psychologists
• NSF NCSES Survey of Earned Doctorates, tables 13 and 3-1
• NCES Digest of Education Statistics 2023
• NY Fed — Labor Market for Recent College Graduates
• Bundesagentur für Arbeit — Blickpunkt Arbeitsmarkt: Psychologie
• BA Fachkräfteengpassanalyse 2024 and 2025
• Destatis — GDP 2025; 16th coordinated population projection; insolvency statistics
• Eurostat job vacancy statistics; employment by NACE (lfsa_egan2); telework (lfsa_ehomp)
• Indeed Hiring Lab job postings tracker (open data); FRED IHLIDXUSTPHUMARESO
• IAB-Stellenerhebung (methodology only; figures not obtained)

Occupational safety, health and working conditions

• EU-OSHA ESENER 2024 and ESENER-3 (2019)
• EU-OSHA OSH Pulse 2025 and 2022
• EU-OSHA Healthy Workplaces Campaign 2026–2028
• Eurofound EWCS 2024 overview report; The never-ending workday in flexible-working Europe
• BMAS / DGUV — GDA Betriebs- und Beschäftigtenbefragung 2023/24
• BAuA — Sicherheit und Gesundheit bei der Arbeit (SuGA) 2024
• GDA Arbeitsprogramm Psyche; GDA-Leitlinie psychische Belastung
• DGUV Vorschrift 2 and DGUV Regel 100-002
• National Safety Council — Work Injury Costs 2024
• NIOSH Science Bulletin 2024 on psychosocial hazards; AIHA on the reversal of NIOSH staffing cuts

Health, absence and prevention (Germany)

• AOK / WIdO Fehlzeiten-Bilanz 2025 — including the eAU statement
• DAK Psychreport 2025 and 2024
• TK Gesundheitsreport 2026
• GKV Präventionsbericht 2025; REHADAT-Statistik BGF series
• BPtK; DGPPN Zahlen und Fakten

Law, regulation and registers

• Regulation (EU) 2026/1744 (Digital Omnibus on AI); AI Act Art. 5 and Art. 27
• Directive (EU) 2023/970 (pay transparency)
• Directive (EU) 2024/2831 (platform work)
• Regulation (EU) 2022/2554 (DORA)
• Commission — revised ESRS, 3 July 2026; Omnibus I in the Official Journal
• COM(2025) 944 — Quality Jobs Roadmap
• EP procedure 2026/2023(INL) — psychosocial risks; EP resolution of 5 July 2022
• EU-OSHA OSHwiki — member-state psychosocial law tiers
• Arbetsmiljöverket AFS 2023:2; Arbejdstilsynet Executive Order 1406/2020; Arboportaal PSA; LOI n° 2021-1018; INAIL stress lavoro-correlato; PIP — Polish mobbing reform
• 29 CFR Part 1607 (UGESP), eCFR; RIN 3046-AB43; RIN 3046-AB45
• 91 FR 54444 — rescission of EO 11246 implementing regulations; EO 14281, 90 FR 17537
• NY State Comptroller — Local Law 144 enforcement audit
• California FEHA automated-decision-system regulations; Texas HB 149 (TRAIGA)
• MHPAEA 2024 final rules

Supervisory, sector and corporate

• DNB — supervision of governance, behaviour and culture; CEPR account of DNB's programme; FRC case study
• ECB Annual Report on supervisory activities 2025; SSM supervisory priorities 2026–2028
• EASA — rules on mental fitness of air crew
• NSI National Health Care Retention & RN Staffing Report; Han et al., Annals of Internal Medicine 2019
• Korn Ferry FY2026 10-K; WTW FY2025 results
• CNIL €32m fine, Amazon France Logistique
• Gartner — CFO budget plans for 2026; Deloitte 2026 Global Human Capital Trends
• Training Magazine 44th Industry Report; ATD State of the Industry 2025
• ICF Global Coaching Study 2025 and 2023

Research, surveys and academic

• Sackett, Zhang, Berry & Lievens — design implications of the revised validity estimates; the 2022 JAP paper; SIOP TIP summary
• Bloom, Han & Liang — hybrid working RCT, Nature 2024
• Bisbee et al. — synthetic replacements for human survey data
• Chisholm et al., Lancet Psychiatry 2016 — the source of the WHO/ILO figure; WHO/ILO policy brief 2022
• OECD Employment Outlook 2023 — AI, job quality and inclusiveness
• Brynjolfsson, Chandar & Chen — Canaries in the Coal Mine?; SignalFire State of Talent
• Burning Glass Institute × HBS — Skills-Based Hiring 2024
• Revelio Labs — DEI headcount; Harvard Law School Forum — DEI disclosure trends 2025
• Gallup State of the Global Workplace; Gallup Engagement Index Deutschland
• Pew Research Center — US workers on AI at work
• Eurostat — AI use in enterprises 2025; US Census BTOS; Federal Reserve FEDS Note on divergent AI adoption estimates
• CHE DatenCHECK — Psychologie study places; DGPs on the PsychThG reform; BDP on title protection