The rail wage

Sydney’s train drivers are asking for a 6% annual pay rise over the next few years.[1] Some of our Very Serious People have argued that, unless the drivers increase their productivity, they deserve nothing.[2] Bleeding-heart liberals point out that driving a train is a difficult and stressful job, and that living in Sydney is expensive, so productivity shouldn’t matter.

In this case, the bleeding hearts are correct. Any insistence on productivity gains is pure politics, based on economic ignorance.

Sydney has had pretty strong productivity growth over the last decade or two, according to ABS figures that were coincidentally released as the train drivers voted to strike. The improvements in productivity have been in tradable sectors, like manufacturing and IT. Nontradable sectors, like transport, have shown no improvement.

In a situation like this, you’d expect to see the Balassa-Samuelson Effect: for the economy to reach equilibrium, workers in the nontradable sectors need to get a significant pay rise. Otherwise, they will quit and join the tradable sector instead.[3]

Today’s risible decision by the Fair Work Commission to prevent the train drivers’ strike—their reasoning, as reported, was that the strike could only be allowed if it had no impact on anyone—will only postpone the effect. If the drivers are forced to return to work, economic theory would predict lower work effort, more absenteeism (practise that wet cough for your sickie call on Monday morning, one journalist advised), difficulties in recruitment, and higher rates of staff turnover. With overall unemployment low, now is a great time for rail staff to try their hands at a new career.

Eventually, all those factors will force the government to give drivers a fat pay rise. And when it comes, the workers should ignore the Very Serious People, enjoy every dollar, and raise a toast to Balassa and Samuelson.


  1. According to the union, their main beef is with compulsory overtime and other types of overwork, but they also want a wage rise. Image credit for this post: Wikimedia Commons ↩︎

  2. For example, a hack from the AFR put that case last week. The government apparently tried to make any wage rises conditional on productivity improvements. In turn, the union argues that its members have already made them. ↩︎

  3. At least, that’s what the back-of-the-envelope Balassa-Samuelson model predicts. In reality, retail workers and hairdressers won’t leave their jobs on Friday and become web designers on Monday, so you’d expect to see steady upwards pressure on wages for at least a few quarters. ↩︎

Jamie Hall

Jamie Hall

Data scientist and software engineer at Kaggle. Previously a China hack and quant modeller at the RBA. These are my personal opinions.

Read More