Generous tax-free salaries topped up with non-residence allowances; remuneration to cover school fees; regular, paid trips home: these are among the privileges enjoyed by scientists at Europe's leading multinational laboratories. The labs' member states would like to trim these perks, but they have so far found this almost impossible to do.

The custom of extending the privileges of international civil servants to scientists was established by CERN, the European laboratory for particle physics near Geneva. CERN was founded in 1954 as a beacon of international cooperation in a continent only recently ravaged by war. Making it an attractive place to work was a high priority.

Over the next two decades, other 'Eurolabs' were set up along similar lines. The European Southern Observatory (ESO) in Garching, near Munich, was founded in 1962; the European Molecular Biology Laboratory (EMBL) in Heidelberg in 1974. Working in a foreign European country was then still an unusual and often daunting experience, so these labs adopted similarly generous salaries and benefits packages.

But as Europe moves ever closer to political union, and as national laboratories become increasingly international in their recruitment strategies, the landscape has shifted. Is it still appropriate to compensate European scientists so generously for living in some of the continent's most beautiful cities? Researchers whose careers have similarly taken them to another European country — but who don't happen to work at one of the élite Eurolabs — would surely answer no.

Europerks

At CERN, a young, unmarried graduate can expect to earn 47,500 euros (US$43,000); a married physicist with two children, a PhD and six years' experience might earn 63,000 euros; senior scientists typically command more than 80,000 euros. These salaries are all tax-free. At DESY in Hamburg, Germany's main high-energy physics laboratory, the equivalent salaries are 18,500 euros, 52,000 euros and 62,500 euros, respectively — from which up to 40% is deducted in tax and social security payments. Yet DESY seems to have no problem attracting foreign nationals: only one-third of its scientists are German.

Foreigners who work for Germany's Max Planck Society — who make up to 10% of its scientific workforce — have similar salary and benefits packages to those at DESY. And remember, researchers in Germany are among Europe's better paid: senior scientists working for the CNRS, France's national research agency, earn around 51,000 euros per year, before deductions for tax and social security payments. Again, around 10% of CNRS scientists are foreigners.

The advantages of working in a Eurolab do not end with tax-free salaries. For example, CERN pays school fees of some 10,000 euros per year per child for foreign scientists; the chosen schools can be in any of the labs' member states. Every two years, scientists from abroad get a paid trip home. Bizarrely, more than 200 of CERN's 2,600 staff also have diplomatic status.

CERN isn't even the most generous of the Eurolabs. Indeed, when the lab reviewed its salary and benefits packages in 1995 and 1999, it ranked only fifth out of seven among comparable European-level facilities. After its reviews, CERN's governing council decided not to tinker with the system, arguing that generous salaries are necessary to continue attracting the best physicists against strong competition from industry. Over the past decade, however, the member states that bankroll other Eurolabs have begun to question the favourable terms and conditions that their scientists enjoy — but they have struggled to shift the status quo.

Freezes

ESO's council, for instance, has tried hard to freeze salaries and squeeze privileges. From the early 1990s it departed from its policy of awarding annual pay increases based on recommendations of the Coordinated Organizations, a committee of the Organisation for Economic Co-operation and Development that helps set pay for international bodies including the North Atlantic Treaty Organization.

Instead, salaries were kept down to, or below, inflation. But ESO's staff association complained to the International Labour Organization (ILO), and won its case. ESO's council agreed to compensate, but then changed its rules to give the council more leeway in deviating from the recommendations of the Coordinated Organizations. The ILO now argues that ESO has not clearly justified its salary decisions, and more cases are pending. ESO has, however, managed to check some privileges. For example, school-fee allowance has been frozen at some 7,200 euros per child per year for the past decade.

The EMBL has similarly fallen foul of the ILO in its attempts to control salary costs — and earlier this month was forced to agree to pay in full a claim for backdated implementation of higher salary scales that will cost the lab some 1.25 million euros.

Neither officials nor scientists at the Eurolabs like to talk about the issue of salaries and benefits. But they should, because an important principle needs to be addressed. Although it would be ludicrous to suggest that employment packages should be reduced to the lowest common denominator, or that they can be standardized across different types of laboratories, it must be right to question the silent continuation of perks that no longer have a clear practical or political justification.

This is particularly relevant in the current harsh economic climate, in which the Eurolabs may be forced to consider eating into their research budgets. The EMBL's governing council, for instance, will decide next March whether member states will stump up extra funds to pay for the backdated salary payments. If not, the money will have to be found from within the lab's existing budget.

CERN, meanwhile, is currently in considerable financial difficulties, facing a major overrun in the cost of building the Large Hadron Collider (LHC), its next big accelerator (see page 841). The lab's council is now reviewing the costs of all experiments that are peripheral to the LHC's main aim, the search for the Higgs boson, with a view to making savings.

Among the experiments under review are those that do not use the main accelerator's beamline, in which particle beams are fired at fixed targets. Their annual running cost amounts to some 10 million euros — about the same sum that CERN spends each year on non-salary benefits.