But what if the historical baseline has always been misleading, and undercounting a problem?
> Simply redefine "murder" to mean "persons killed by a woman wearing a blue hat on the third Tuesday of February".
This isn't a useful example, because the point of redefining unemployment is to get a more accurate view of it, whereas this redefinition of murder gives a less accurate and completely pointless view of the problem.
Then you create a new metric, but keep the old one around for historical baselines.
By the way, this is what the Bureau of Labor Statistics (BLS) is doing with the various unemployment metrics.
> The point of redefining unemployment is to get a more accurate view of it, whereas this redefinition of murder gives a less accurate and completely pointless view of the problem.
I see you are new to politics.
You need that historical baseline to assess the current situation.
To provide an extreme example, as extremes are often useful in illustrating failure cases, it would be trivial to say that the murder rate in the US has dropped by nigh-on 100%. Simply redefine "murder" to mean "persons killed by a woman wearing a blue hat on the third Tuesday of February".