Conceptualizing "recession"

Officially, recessions are measured by how the economy is changing. I think this is wrong. While it is useful to know if an economy is getting better or if it is getting worse, the primary measure of health ought to measure it against what it could be, not against what it was in the last quarter.

Example: Imagine that the economy goes into a tailspin, that unemployment shoots up from 5% to 10%, and that GDP plummets from $ 14 trillion to $13 trillion. Now, imagine that things start to improve after about 1 year. However, imagine two different scenarios...

Come-back scenario: A year after turning downward, the economy starts to pick up again. In another year, it is back where it began, with unemployment at 5% and GDP back to $14 trillion. The dip and return took two years; but, as currently conceptualized, the recession would be measured as being 1 year long, because that is when the economy started to rebound. (i.e. the recession is measured from "peak" to "trough").

"New normal" scenario: Under an alternative scenario, a year after turning downward, the economy flattens, and then climbs almost imperceptibly. Then, another year out, it starts to climb a little faster, but is still quite laggard. Finally, 5 years later unemployment is back down at 5%, and GDP is back to $14 trillion. Even though the dip and return took so many years, the recession would be measured as being 1 year long, because that is when the economy stopped going any lower.

This way of conceptualizing a recession is faulty. We need a measure of economic health that meaningfully describes how an economy is doing, compared to its potential, not a measure that accepts the previous quarter as a "new normal" and rejoices in small upturns.