A new paper from George Mason University professors Noel Johnson and Mark Koyama makes a pretty serious point: A government’s ability to raise taxes is directly related to its ability to enforce law. But the researchers explain that conclusion in a rather unconventional way, by charting the decline of witch hunts in 17th-century France. Here’s the key paragraph:

This is a unique data set as other countries do not have information on both witch-trials and tax collection at the regional level during the sixteenth and seventeenth centuries. We find that regions with higher taxes were less likely to try witches and that the rise of the fiscal state across much of France during the mid-seventeenth century can account for much of the subsequent decline in witch-trials. These results are robust across a range of different econometric specifications and our findings are supported by additional historical and qualitative evidence.

Here’s what this all looks like in map form, comparing tax collection levels to the frequency of witch hunts. Central France had low levels of witch hunts (the lighter countries in the map on the left) and higher taxes (the dark blue countries on the right):

“Centralized legal institutions,” they wrote, “led to less judicial discretion, improved standards of evidence, and increasingly standardized law enforcement...had the inadvertent effect of reducing the number of witches tried.”

The authors’ research didn’t cover the question of what triggered the witch hunt phenomenon. On that, Johnson and Koyama said: “The underlying causes of the `witch craze' of the early modern period remain a puzzle which continues to defy explanation.”