Here is a chart that has got me thinking. It shows the relationship between currency undervaluation and growth spurts in two sets of economies: those in Asia, and those in Africa. Basically, it depicts the "typical" (average) trend in under- or over-valuation in the ten years preceding and ten years following the onset of growth accelerations in the two regions. (See here for how growth accelerations are defined and their timing determined.)
Notice the difference?
In Asia, growth accelerations are preceded by a period of sustained increase in undervaluation, which is maintained during the period of high-growth. In Africa, by contrast, growth accelerations are preceded by growing overvaluation; the currency always remains in the overvalued zone (= negative undervaluation).
Here is the simplest explanation. In Asia, growth is typically engineered by increasing the profitability in manufacturing and other tradables. But in Africa the typical growth spurt is preceded by aid inflows and other transfers, which appreciate the exchange rate, and render future growth less sustainable. This is the so-called Dutch disease.
Africa has been having a good few years recently. To my mind, the key question with regard to sustainability of growth has to do precisely with the real exchange rate: is the real exchange under- or overvalued, and has it been appreciating or depreciating?
UPDATE: Luis Enrique is smart! (See comments below.)