Matthew Dimick is Associate Professor of Law at the University at Buffalo School of Law.  He can be reached at [email protected].

Last week, the US Department of Labor released its latest union membership statistics.  In 2016, the rate of union membership among wage and salary earners—or union density—was 10.7 percent, down 0.4 percent from the previous year.  Unsurprisingly, union membership remains low, and far below its historical high point.  In 1964, for example, nearly a third of workers belonged to a labor union.

Union density is not the end-all and be-all of unionism, but it’s hard to overestimate its importance for succinctly capturing the strength of a labor movement.  Union members pay dues, are more likely to vote (and probably more likely to vote Democrat), and are more likely to participate in job actions and other tactics necessary to put economic pressure on an employer.  Even more broadly (and loftily), union membership initiates workers into an institution of community for both transcending differences (e.g., between races and genders) and developing broader, collective interests.

It’s not surprising therefore that goal number one of the labor movement has been to reverse the decades-long decline in union density.  This goal takes on existential proportions in light of the continued legal and political assaults on labor unions.  Last year, public sector unions—constituting no less than 40 percent of total union membership—dodged a bullet when the death of Justice Scalia led to a deadlocked Supreme Court in the Friedrichs case.  Had Justice Scalia lived, it is almost certain that the Court would have allowed public sector workers to enjoy the benefits of collective bargaining without having to join the union or otherwise contribute to supporting it financially.

Unfortunately, with a Trump presidency, labor supporters weren’t allowed much of a reprieve.  And given the larger political climate and vicissitudes of national labor law, this should impel the labor movement to seek out other ways of building union membership.  One of these ways, as I discuss in this post, is to adopt what is called the “Ghent system.”

On its face, the Ghent system—which exists in Belgium (where it originated in the eponymous town), Denmark, Finland, and Sweden—is simply a way for the government to turn over the administration of unemployment benefits to labor unions.  The system is voluntary, rather than compulsory.  To get unemployment insurance, you have to visit your occupation or industry’s union office and sign up.  But the fee is relatively nominal, since most of the financing is through government tax revenue.  Despite—or perhaps because of—its voluntary nature, coverage and take up is better in the Ghent system than in compulsory systems like the US or other European countries.  The funds are closely regulated and kept separate from other union finances.

Dig down deeper, however, and you find that the Ghent system has an enormous impact on union membership rates.  It makes about a twenty-to-thirty percentage point difference in union density between otherwise similar countries, such as Scandinavian Sweden and Norway and Benelux Belgium and the Netherlands.  These findings have been confirmed repeatedly in several social science studies (see, e.g., here and here).  No one seriously doubts the substantial positive impact of the Ghent system on union density. Confirmation this clear rarely comes in the social sciences.

But perhaps it shouldn’t be so surprising.  The Ghent system solves a free-rider problem habitually faced by labor unions.  For several reasons, unions cannot exclude individual workers from the collective benefits—improvements in wages, hours, and working conditions—they deliver.  If everyone gets the good as a matter of course, then why pay for it? And if nobody pays for it, the union’s ability to provide the benefits is threatened.

One way of solving the free-rider problem is by using “selective incentives”: a separate and “private” good that is made available only on the condition that people contribute to the collective good.  In US unions, the selective incentive is the job itself and is implemented with a union security agreement—the very device challenged in Friedrichs.

In the Ghent system, the selective incentive is unemployment insurance.  As already mentioned, to get unemployment insurance, one must sign up for it at the union office.  Interestingly, although historically union membership was required to get unemployment insurance this is no longer the case.  Nevertheless, the fact that one has to buy unemployment insurance from the union allows it to make its union-membership pitch.  The conferral of such a benefit also probably induces reciprocal incentives on part of the worker.  And one probably cannot exclude the role of some amount of social pressure to join the union while signing up for unemployment insurance.

Other big benefits come from the Ghent system.  First, it separates the decision to form a union from the decision to join the union. In the US, getting union members first requires getting the employer to accept a collective bargaining relationship with the union.  In a constantly shifting economy, this is bound to fail.  Separating these choices is probably the main reason why union density at the national level in Ghent-system countries is less different from union density within firms.  In the US, you have 100 percent density within unionized firms and zero everywhere else.  (Even in “right-to-work” states, union-density within bargaining units is typically quite high.)  Such an imbalance, I suspect, plays an enormous part in explaining particularly hostile employer resistance to unions in the United States.

In addition, with some influence over unemployment insurance, unions are able to bargain for security in the labor market.  Insurance probably works better, from an economics and labor-market perspective, than detailed restrictions on employers’ rights to terminate an employee.  Little surprise then that Denmark’s “flexicurity” system has been the envy of both the “social partners” and other countries in Europe.

What makes the Ghent system tantalizing from a progressive-federalist perspective is two things.  First, as readers of the blog will be aware, federal labor law preemption is famously (or infamously) vast, creating huge hurdles for local experimentation.  However, unemployment insurance appears to be an exception.  As the Supreme Court said in N.Y. Tel. Co. v. N.Y. Dep’t of Labor, the “omission of any direction concerning payment [of unemployment benefits] to [workers on strike] … implies that Congress intended that the States be free to authorize, or to prohibit, such payments.”  Thus, the deference given to states to administer their unemployment-insurance programs appears to give them more latitude under federal labor law preemption.

And, second, the deference states do get under the US’s unemployment insurance systems is broad.  As envisioned, that system is essentially a federal-state partnership that incentivizes states to adopt, finance, and administer their own unemployment-insurance systems under federal guidelines and oversight.  Federal certification is quite general, merely requiring a state to administer a program using “[s]uch methods of administration … as are found by the Secretary of Labor to be reasonably calculated to insure full payment of unemployment compensation when due” and “[p]ayment of unemployment compensation solely through public employment offices or such other agencies as the Secretary of Labor may approve …” (emphasis added).  The “such other agencies” language does not appear to prohibit nonpublic—labor union—forms of administration.

This deference makes a labor-law reform like the Ghent system much easier to enact.  Rather than requiring approval from both legislative and executive branches, a state only needs the endorsement of the executive branch.  Certainly, under a Trump Presidency, the Secretary-of-Labor requirement may make the Ghent system an impossibility in the short term.  But a Ghent-system reform is ideal under conditions like those of outgoing President Obama, who faced a Republican Congress during most of his administration. We typically think that those conditions will forestall any labor-law reform, but not so with a Ghent-type reform.  Furthermore, as I argue elsewhere, there may be other avenues for Ghent-type experiments.  The selective incentive need not be unemployment insurance.  What about job training?  The labor movement could also marshal its considerable resources to fund a fully private style Ghent system, not unlike some innovative worker organizations.  In any case, the Ghent system should be foremost in the minds of those contemplating a progressive-federalist strategy of labor union revival.