Did you read the full article? They have an paragraph on that under the section Methodology. IMHO this addresses your concern. What's your problem with that reasoning?
>The numbers for the map come from historian Angus Maddison who uses Sestertius records to find wheat equivalent figure for national disposable income and makes comparisons based off that.
But all that says is that some guy used historical records for Roman income figures. It doesn't tell you anything about how to relate income in Roman times to income today.