Exorbitant Changes in Three Parts

We document that the positive differential on international portfolio returns, one aspect of the U.S. exorbitant privilege, has disappeared in three parts. Part One: U.S. international liabilities used to be mostly in low-return bonds while its international assets were largely in high-returning equities, thus naturally producing a positive return differential. More recently, however, U.S. equity liabilities have increased sharply, reducing this compositional tailwind. Part Two: While the exorbitant privilege literature has focused on expected returns as proxied by the sample arithmetic mean, the geometric mean is also required to produce an unbiased estimate of expected returns. Incorporating geometric means greatly reduces the returns differential. Part Three is a combined switch a) from aggregate to comprehensive security-level data to more accurately calculate returns and b) from expected to actual realized returns that take into account the timing and magnitude of portfolio flows. The combined effect of these changes is that over the past two decades the U.S. portfolio returns differential was not 228 bps but zero, and it is expected to be zero for the next decade.