When relative rates are misleading

Mark points me to this article.

To me these effects are hard to interpret in terms of relative growth rates given that rich areas of the country remained richer (on average) than the poor areas of the country. Let us consider a thought experiment to see what I mean by the problem of relative rates.

One region, call it sunbelt, has income of $1,000 per capita and this income increases by $20 (2% growth per year).

Another region, call it rustbelt, has an income of $500 per capita which increases by $15 (3% growth per year).

Clearly the rustbelt region has better growth (in % terms) but has a lower growth in absolute wealth. Now let us presume that the absolute growth rates remained constant over a decade.

At the end, despite always having lower growth, sunbelt has $1,200 in income while rustbelt has $650 per capita. The relative difference is smaller (as 650 > 50% of 1200) but, in absolute terms, the people in sunbelt now make $550 per year more than the people in the rustbelt.

Are we sure that % growth rates are always the best metric of economic progress?