It seems to me that the prime determinant of pay is how close the worker's "category" is to the money. So in a start-up intensive environment like the Bay Area, the devs are also sometimes the executives, and executives are often former devs, so devs get paid correspondingly.
Norway interestingly has almost the same pay as USA - $68,737 [2], probably because of high share of oil extraction in GDP (22% [3]) which skews the result. If you subtract oil share from the GDP, you get $56,195.88 nominal GDP per capita, which is almost the same as the American one ($57,220)!
Now Luxembourg is a tax haven/financial center with a population of half a million so I don't think it's a relevant comparison. Just due to population to get something close to true pay average you would have to ask a relatively (to other countries) very big percentage of their developers.
Same site for comparable data
[0] http://www.payscale.com/research/CH/Job=Software_Engineer/Sa...
[1] http://www.payscale.com/research/US/Job=Software_Developer/S...
[2] http://www.payscale.com/research/NO/Job=Software_Developer/S...
[3] https://www.ssb.no/en/befolkning/artikler-og-publikasjoner/_... page 40, graph
> GDP per capita seems to be a curious way to estimate if a local dev is likely to be productive.
Curious is putting it mildly. :) What is being objected to is your _measure_. Stop using the GDP per capita of an entire country as a proxy for software developer productivity. It's frankly stupid. Anyway GDP per capita using PPP (purchasing power parity) is seen as a fairer comparison. But even GDP per capita by PPP is a stupid metric to compare software developer productivity. I'm sorry for using the word stupid. But it's stupid.
Outside of entertainment where software is a direct consumer product, the developer's productivity comes from increased efficiency of use of other productive resources. You can't eat code, but you can eat food that comes from higher production due to better software. So software has a multiplicative effect on existing production. That is, GDP.
Now even added value of entertainment software (games etc) depends on total GDP, because people have to pay with something for that entertainment.
So average developer's productivity IS a function of GDP, with different coefficient depending on the structure of a economy.
A primitive non-mechanized agricultural economy would have a coefficient of near zero because there's almost nothing to automate.
>Anyway GDP per capita using PPP (purchasing power parity) is seen as a fairer comparison.
A fundamentally wrong metric because pay is nominal.
So why would the GDP per capita of their country be a useful tool in deciding who to hire?
A Swiss farmer is much more productive when measured in currency units, but probably not that much in milk volume.
However as we are talking about salary differences it's money that matters.
>So why would the GDP per capita of their country be a useful tool in deciding who to hire?
Why would it be? In this conceptual model the ability of a developer is how much he multiplies the output of whatever he's working on, but his productivity is the absolute value of added output. How could it be counted otherwise, in what? Lines of code?
If productivity didn't depend on location immigration wouldn't exist.
A remote dev working for a Bay Area company from Spain can be just as productive for his employer as one located in Los Gatos, CA. However the above method would categorize this dev as "objectively less productive", which seems counter-intuitive...
It wouldn't, it purports to explain the differences in local salaries, or more precisely salaries paid by local entities to on-site developers.
It's true I didn't specify that explicitly in the first comment, along with definition of productivity, so your reading of it was a reasonable understanding. It's a good thing you helped me clarify the intended meaning.
One assumption is that foreign demand (for non-local use) for local on-site developers is small enough to not change the workforce demand significantly. So it won't work for India or other common offshore destination, but it seems to explain pay differences between USA and Spain, Norway and Switzerland reasonably.