Relative pay in the financial sector has experienced an extraordinary increase over the last few decades. A proposed explanation for this pattern has been that the demand for skilled workers in finance has risen more than in other sectors. We use Swedish administrative data, which include detailed cognitive and non-cognitive test scores as well as performance in high-school and university, to examine the implications of this hypothesis for talent allocation and relative wages in the financial sector. We find no evidence that the selection of talent into finance increased or improved, neither on average nor at the top of the talent distribution. A changing composition of talent or their returns cannot account for the surge in the finance wage premium. These findings alleviate concerns about a "brain drain" into finance at the expense of other sectors, but they also suggest that rents in finance are high, increasing, and largely unexplained.