L2 earnings potential: German well above average, French above average, Spanish below

This article in the Economist on the economic benefits of knowing a second language cites a 2002 paper that agrees with a constant refrain of mine: that parents should not have their children learn a language simply because it's popular, widely studied or spoken. At the end of the day an employee looking for an economic benefit from knowing a language only needs to find one job in order to make this happen, so languages like Norwegian (pop. 5 million) should not be looked down upon if knowing it means a good job at a company like Statoil, for example.

The chart in the article shows German at the top of increased earnings, French is below yet still above average, and Spanish is below. Spanish is easy to understand. Why would a company pay a premium for a fluent Spanish speaker when they can just find someone that speaks it as a mother tongue? As the article puts it:

Non-Latino Americans might study Spanish because they hear and see so much of it spoken in their country. But that might be the best reason not to study the language, from a purely economic point of view. A non-native learner of Spanish will have a hard time competing with a fluent native bilingual for a job requiring both languages. Indeed, Mr Saiz found worse returns for Spanish study in states with a larger share of Hispanics. Better to learn a language in high demand, but short supply...

And yes, of course one should learn the language one likes the best. But simply adding the GDP of countries that use the language together and claiming that it is therefore the one that makes the most economic sense is incorrect.