The "balance between worker protection and labor-market flexibility" is the policy debate's most over-rehearsed false binary. The Danish flexicurity model — easy hiring and firing, generous unemployment benefits, mandatory active labor-market policies — was designed in the 1990s precisely to demonstrate that the trade-off does not have to be steep. The empirical record of three decades has largely vindicated the design. Denmark's unemployment rate has tracked the OECD median through the 2008 crisis, the pandemic, and the post-pandemic period; worker turnover is high, but earnings scarring from job loss is meaningfully lower than in countries with stricter employment protection but weaker active policies. The OECD's Employment Outlook has used Denmark, the Netherlands, and to a lesser extent Sweden as the running benchmark cases for two decades.
The argument here is that the U.S. has the worst of both worlds: weak employment protection (the at-will doctrine still applies in 49 states) paired with weak active labor-market policies (federal ALMP spending at roughly 0.1% of GDP, against an OECD median above 0.5%). That is not a balance. That is a thin layer of statutory minimum-wage and anti-discrimination law over a system that offloads displacement risk almost entirely onto individual workers. The reform agenda is not "more protection" or "more flexibility." It is more of both, structured around portability.
The flexicurity evidence base
The original Danish flexicurity model has three pillars: flexible hiring and dismissal rules, generous unemployment benefits (typically replacing about 80–90% of prior wages for low earners, with a benefit ceiling), and "activation" — mandatory participation in job-search assistance and training as a condition of receiving benefits beyond an initial period.
The empirical evaluation literature, including the OECD's Back to Work series, Bredgaard and Larsen's research at Aalborg University, and Andersen, Holm-Hadulla, and Schimmelpfennig's IMF analyses, finds three things. First, the model substantially reduces long-term unemployment by churning workers through short spells rather than leaving them stuck. Second, it shifts employer behavior toward more hiring at the margin, because dismissal costs are predictable rather than litigated. Third, it depends critically on the activation pillar working — when activation is underfunded or poorly implemented, the system degrades toward expensive passive transfers.
The Dutch model is structurally similar but with stronger sectoral collective bargaining as the wage-setting mechanism. The German model is closer to "rigidity plus retraining" — stronger employment protection but with the Kurzarbeit short-time-work scheme that prevented the 2008 and 2020 unemployment spikes that hit other economies harder. Each is a viable variant. None of them looks like the U.S.
The U.S. position
The U.S. labor market is famously flexible in the OECD comparative ratings. The OECD's Employment Protection Legislation index, with all the caveats about how it is constructed, consistently places the U.S. at the low end among rich economies. At-will employment is the legal default in every state except Montana. The percentage of workers covered by collective-bargaining agreements has fallen from roughly 35% in the 1950s to about 10% in 2024, per BLS data. Unionization in the private sector is below 7%.
The flexibility side is, on its own terms, working: U.S. hiring rates and job-creation rates are high, unemployment spell durations are shorter than in many European economies, and labor markets clear quickly. The cost is borne by displaced workers, who experience large and persistent earnings losses. Steven Davis and Till von Wachter's 2011 paper "Recessions and the Costs of Job Loss" (Brookings Papers on Economic Activity) found that displaced workers in the U.S. experience earnings losses of roughly 1.4 to 2.8 years of pre-displacement earnings, depending on the local unemployment rate at the time of displacement. The losses persist for two decades.
What the U.S. system lacks, against the OECD benchmarks, is the active and passive labor-market policy infrastructure that converts displacement into transition rather than into prolonged earnings depression.
The four reforms that would actually rebalance the system
Modernized unemployment insurance with national minimum standards
Average U.S. UI replacement rates are about 38% of prior wages — well below the OECD median. State-administered UI systems produce huge variation: Massachusetts, Washington, and Hawaii replace meaningfully more; Mississippi, Tennessee, and Florida replace substantially less. The 2020 COVID Pandemic Unemployment Compensation experiment, which temporarily added $600 per week, demonstrated that higher replacement rates did not significantly reduce job-search activity (research by Marinescu, Ganong, and others is the most-cited evidence on this). A national replacement-rate floor of 60%, with duration tied to local labor-market conditions, is the most direct UI reform.
Tripled federal ALMP spending
The U.S. spends one-tenth what peer economies spend on active labor-market policies. The benefit-cost evidence on sectoral training programs, registered apprenticeship, and short-duration job-search assistance is strong enough that scaling is a higher-confidence intervention than most alternative uses of federal funds. See our flagship treatment of this in Reskilling for Real →.
Portable benefits for non-traditional workers
Roughly 36% of the U.S. workforce, by Upwork's recurring Freelance Forward survey and adjacent studies, performs some independent contract work in a given year. The traditional benefits structure does not serve these workers. Washington State's 2023 portable-benefits law for gig workers, paired with the existing ACA framework for health insurance, is the closest U.S. template for a benefits structure that travels with the worker.
An employer-side reform: stronger advance-notice and severance defaults
The WARN Act, the U.S. statutory advance-notice law, applies only to mass layoffs at large employers and has notable enforcement weaknesses. Strengthened advance-notice requirements (60 to 90 days for most non-cause terminations) and statutory severance defaults (one to two weeks per year of service) would import elements of European employment protection without producing the dismissal-rigidity that the strongest European protections create.
What flexicurity does not mean
The most common misreading of flexicurity is that it is about laxer employment protection. It is not. Danish hiring and firing rules are not dramatically lighter than U.S. at-will employment in practice. The difference is that the social-insurance backstop, the active-policy infrastructure, and the sectoral collective-bargaining wage floor make the system land softly on workers who are displaced. The U.S. has the lax employment protection without the soft landing.
The other common misreading is that flexicurity requires Nordic-scale taxation. Denmark and the Netherlands have higher overall taxation than the U.S., but the marginal cost of the ALMP and UI components is small relative to the entire fiscal envelope. The U.S. could fund a credible flexicurity-light system at a fiscal cost of well under 1% of GDP, given the low starting base.
The U.S. has the worst of both worlds: at-will firing without the active labor-market policy that catches displaced workers. That is not balance. It is offloading.
The labor-market-flexibility debate in the U.S. has been stuck in a tendentious back-and-forth between "more protection" and "more flexibility" arguments for thirty years. The empirical answer from comparable OECD economies is that both can be expanded simultaneously, if the policy mix is right. Stronger UI, tripled ALMP spending, portable benefits, and modest employer-side notice and severance defaults would move the U.S. system toward the OECD median without imposing the dismissal-rigidity that the harshest critics of European protection legitimately worry about. The political path to this set of reforms is hard. The policy design is not.
Updated May 21, 2026. This piece was substantively rewritten as part of NWLB's 2026 editorial refresh.
