Union Front Group Needs a Lesson in Basic Statistics

Anyone who’s taken an entry-level course in statistics has heard this phrase: “Correlation does not mean causation.” For instance, Americans might be drinking more diet soda and buying more cars, but that doesn’t mean that one trend has caused the other. It’s a lesson that the Restaurant Opportunities Center (ROC)–a union-founded activist group–should have reviewed before releasing its latest report on the tipped minimum wage.

ROC’s report claims that states with a higher tipped minimum wage have higher restaurant sales per person than the rest of the US–and that eliminating the tip credit is thus a “recipe for success.” But not only does ROC fail to prove that tip credit levels are a major driver of restaurant sales—its numbers actually make the opposite case.

ROC offers up this graph to support its argument:

 

For readers who aren’t steeped in statistical jargon, the small letter r at the bottom of the graph represents the strength of the correlation. For instance, an value greater than 0.7 would represent a strong positive correlation. In this case, ROC is reporting a negative r value, suggesting that sales decrease as the tipped minimum wage rises. This is obviously a typo on ROC’s part, reflecting an embarrassing lack of attention to detail.

Still, that leaves open the question of whether ROC’s big headline on the tipped wage is actually backed up by its data. All of the numbers ROC used are available from the Census Bureau and the Restaurant Association, so EPI pulled data from 2013 together to check ROC’s claim. No surprise: Comparing tipped minimum wages and restaurant sales yields an r value of only 0.2, which equates to a negligible or weak correlation.

In other words, ROC is wrong – it’s not true that sales will necessarily increase as the tipped wage increases.

This is consistent with both common sense and with the academic research on the tipped wage. To the extent that any relationship exists between a higher tipped wage and restaurant sales, it’s most likely due to employers in high-cost states passing along their increased costs through higher prices. But higher prices can also mean fewer sales, which is why employers will avoid price increases if possible — and why the relationship between a higher tipped wage and restaurant sales is weak. Instead, the research shows employers are forced to reduce employee hours and staffing levels to adapt to the increase in labor costs. That might keep prices steady, but it also means fewer highly-paid tipped jobs are available.

ROC might not have intended to, but if anything this new paper disproves its weak argument in favor of a higher tipped wage. This is the same group, of course, that tells others how to run their businesses but couldn’t make its own restaurant successful. Maybe ROC’s operations management is the same group that did the research for this report.