Arbitrage or Narrow Bracketing? On Using Money to Measure Intertemporal Preferences

Abstract

If experimental subjects arbitrage against market interest rates when making intertemporal allocations of cash, the data will reveal nothing about subjects' discount rates, only uncovering subjects' market interest rates. If they frame choices narrowly, market rates will not be salient and the experiment will uncover subjects' utility discount rates. We test arbitrage directly by forcing all transactions with subjects to be instant electronic bank transfers, thus making arbitrage easy and salient. We also employ four decision frames to test alternative hypotheses. Our evidence contradicts arbitrage, supports money as a valid reward, and suggests framing as a correlate with present bias.