Sunday, 20 June 2010

Corporate nannies

Nick Smith worries about the growing use of on the job drug testing by corporates.
Recreational cannabis use is rife in New Zealand.

There are, according to Auckland University's Alcohol and Public Health Research Unit, about 1.5 million pot smokers in this country, about 18 per cent of people aged 15-45.

More than half the adult population have inhaled at one time or another.

Illegality aside, it is uncontroversial to assert that what people do in their own home and in their own time is their business, not that of the Government or their employer.

So, corporate examination of people's personal lives by drug testing represents, at the very least, a gross invasion of privacy.

There is also an issue of fairness.

Cannabis stays in the system for up to 30 days and a positive urine test does not mean an employee is under the influence; it means that at some point in the last month they have consumed an illegal drug.


Nobody is arguing in favour of smoking on the job but workplace drug testing is being extended way beyond its initial, although dubious, safety justification.

If it were the Government rather than private companies implementing such policy there would be a public outcry at nanny state intrusion.

Perhaps this can be seen as another example of outsourcing government functions to the private sector.

Welcome to the new corporate nanny state.
When we chatted about it earlier, I asked Nick why an employer would ever test employees where it didn't make a difference for job performance and where it does annoy the employees. In short, Becker's discrimination model applies: companies compete for workers, and companies with an irrational fear of employees who are not less productive but who smoke marijuana on their own time will wind up having to pay more to attract staff. On the job drug testing is a disamenity in the overall pay packet requiring compensation.

I generally start from a model where I expect all employers to be completely selfish jerks and then work out why they’d give employees anything: from decent coffee machines to air conditioning in the office. They’re competing with other employers over a total compensation bundle, and if some amenities can be provided at cost lower than their monetary value to employees, then they’ll be provided with a lower monetary salary package. If some other disamenities, like mandatory overtime or working odd shifts or having to take drug tests, increase employer revenue by more than the monetized value of the disutility they impose on workers, then they’ll also provide those disamenities and higher wages. If we had monopsonistic employers with lots of power, then they’d have lots of scope for providing even inefficient disamenities, but that doesn’t characterize most of the workplaces out there. So I could buy inefficient employee drug testing in the “company town” in days of yore, but they don’t much exist anymore and mobility is cheap.

So why then would companies put in place drug tests where the workers' performance isn't likely to be an issue? Can we build a model generating inefficient workplace drug testing - testing where the value to the employer in improved productivity is less than the disutility to workers either of lifestyle changes or privacy violations?

There are a couple of ways. First, if customers put high value on that the product is delivered by a firm with drug testing, then it will be done regardless of whether drug use has any effect on workplace productivity. I don’t buy that story because I think it requires implausibly large customer willingness to pay for something that doesn't have any real effect on product quality. Second, you could maybe build a Glenn Loury statistical discrimination story leading to self-reinforcing equilibria where drug users underinvest in human capital because they expect workplace discrimination (and so drug users are in fact less productive, but only because of the inefficient self-reinforcing equilibrium), but I don’t particularly buy that one either.

Nick's finished piece has a better answer: union regs.
Madison Recruitment's Justin Pipe confirms the trend, saying increasing numbers of professionals are being asked to submit to drug and alcohol testing as part of companies' overall risk management policy.

Usually, Pipe says, it is firms that already have identified ''safety sensitive'' areas, such as operating heavy machinery or onerous driving requirements, that institute such intrusive policy.

On the basis that it is unfair to single out only one employee group, these businesses demand that all workers submit to regular drug and alcohol tests, he says.

Hays' Jason Walker disagrees, noting that such practices are still largely confined to the construction and roading sectors but, he says, testing in these areas are on the rise.

Greg Lloyd, Engineering, Printing and Manufacturing Union general counsel, sides with Pipe. His members, irrespective of the nature of their work, are being asked to provide urine samples for testing.

The EPMU took unsuccessful court action against Air New Zealand over testing of pilots and, Lloyd says, the airline now extends its policy to all members.
If pilot testing is efficient, and sufficiently so that the benefits of it outweigh the costs of imposing testing on all workers given a union constraint of an all-or-nothing regime, then I have a hard time blaming the company for pursuing testing. I can find fault with the union, but I have a hard time finding fault with the firm. If pilot testing is only efficient because of the tort regime, then blame should go to the tort regime.

I just have a hard time seeing why a firm in a competitive labour market would implement an inefficient employment practice. Some might do it by error of course.

And, I would be annoyed if the University here started imposing mandatory drug tests. I don't use anything that would upset them, but I'd still reckon it an intrusion eroding about 5% of the value of my overall pay bundle.

Any better models out there that would generate firms implementing mandatory employee drug tests when those tests are in fact inefficient?


  1. In my work I deal with a number of firms that perform drug testing due to the fact they operate large amounts of heavy machinery. A number of those firms perform random drug testing of all employees mainly for the reasons of perceived fairness that you outlined. Anecdotally it appears to be a case of "if we can't drink on the job, why should the execs be able to do so at their corporate lunches." And in some respects I can understand that position, but on the other hand I'd consider drug testing an invasion of my privacy.

    In other industries drug testing across all employees is possibly valid regardless of whether you actually operate the heavy machinery or not. Working in industries like construction there is probably a fair chance you'll end up visiting a construction site at some point, regardless of your role, so random drug testing of all employees kind of makes sense for those organisations.

  2. I think it's worthwhile here to separate out different sources of benefit to the employer: drug use causing employee inefficiency, and drug use being correlated with employee inefficiency.

    If employers believe (rightly or wrongly) that those people who consume illegal drugs in their spare time are worse workers (lazier, whatever), then it would make sense for the employer to screen and discriminate against those potential and existing employees.

    Whether that's efficient or not just comes down to whether the employer's assumption is right.

  3. @Duncan: The bigger point I'd think would be that marijuana persists, so tests can't distinguish between on the job use and something from three weekends ago. But again, nobody forces you to work for a company running those tests, and if firms have to compete for workers, I'm not sure why they'd do it if it were inefficient.

    @Matthew: It would also depend I suppose on whether there are better signals of employee quality. But the existence of a testing regime could induce a separating equilibrium at hiring time, which could work towards the kind of Loury equilibrium noted above.

  4. I would have marginally more time for testing of this nature if it could directly determine the level of impairment, but that simply isn't the case. I also wonder just how much intoxication in the workplace is a real vs perceived problem...?

  5. DOL did a study of industrial deaths from 2000-2005 and identified drug use as a contributing factor in around 4.5% of deaths, the largest factor was 'human error' at 43%. So perception probably outstrips reality.

    @Eric Possibly not everyone perceives drug tests as an invasion of privacy or do not consider it as a big enough concern in their job considerations or over estimate the risk of intoxification in the workplace; maybe the firms with drug tests can hire those people. I mean people put up with the security theatre at airports when it patently doesn't serve any real purpose or can be defeated with very little effort (ala Bruce Schneier).

  6. @Duncan: The testing is only inefficient to the extent that it's costly, and the biggest cost would be annoying employees; if the employees don't care, then testing might not be inefficient.

    I'd be reluctant to draw comparisons to airports - government regulatory requirements that are awfully hard to avoid compared to choosing an employer partially on testing policies.

  7. My friend made the good point that it's quite possible the cost of animosity employees feel towards the policy is outweighed by the benefit of increased compliance after enforcing the power imbalance inherent in the employer-employee relationship. A kind of "Don't forget, we own you," kind of thing.

    This is plausible, I think.

  8. @Matthew: I'd buy it if folks couldn't choose other employers.

  9. For some workers, with desirable skill sets, switching employers on a whim may be simple, but a lot of poorly skilled employees, or perhaps those with highly specialised skills, might not find it that simple to change jobs out of dissatisfaction with their employer's testing policy. In these cases the employers have their workers over a figurative barrel. I'd expect to see a higher rate of testing in environments with pretty much captive employees compared with workplaces that need to complete more strongly for warm bodies. In a highly competitive market my guess is that there is a stronger incentive to not piss off your employees, and in these cases, if testing is mandated in some way, I'd also expect less serious repercussions of a "failed" test...

  10. @Lats: We don't really need workers to be able to turn on a dime: we only need that they've an increased chance of leaving for other employment or that the firm would have a tougher time in a new recruitment round. If the labour market is really slack, then employers will have more room to chisel on parts of the agreement (like on testing); if it's really tight, then employees are the ones with more room to chisel.

  11. In that case I agree with you. I would find such a testing regime to be sufficiently invasive that I would choose to work elsewhere. Similarly, amd much to my Julie's disappointment, I am sufficiently put off by their draconian entry requirements (fingerprinting, etc.) that I refuse to travel to the US. I'd rather spend my vacation dollars somewhere where I feel welcomed.