In a series of experiments we show that people learn to play the efficient
outcome in an open-ended rent-seeking game. This result persists despite qu
ite different experiment environments and designs, like different propensit
ies of competition, group sizes etc., and is interpretable as a resolution
of the so-called Tullock paradox which states that real-world rent-seeking
expenditures are much lower than what the standard rent-seeking model predi
cts.