Turkish monetary reform, which took effect in January 2005, introduced the New Turkish Lira (NTL) by deleting six zeros from the former currency, the Turkish Lira (TL). Two experiments investigated how the introduction of the NTL might affect price estimation. In the first, conducted in December 2004, 202 students were first presented with high or low anchor values and then estimated the average price of a “new Turkish mid-sized car” in different currencies (TL, NTL and Euro). Although anchoring bias was not significantly different across familiar (TL) and unfamiliar currencies (NTL and Euro), price estimates in Euro and NTL were significantly higher than those in TL. In the second experiment, carried out 6 months later, 212 adult consumers estimated the prices of 13 items in one of three currencies. For five items prices estimated in Euros were significantly higher than those expressed in either TL or NTL. However, there were no significant differences between TL and NTL, suggesting that Turkish consumers had quickly adapted. Such ease of adaptation is consistent with a rescaling hypothesis: when one or more zeros are dropped from a currency, consumers rescale all prices relatively quickly rather than relearn them selectively through gradual exposure.
It is much harder, I think, when the nominal change is either uncertain or not common knowledge or not distributed evenly across economic sectors. In those alternative cases the signal extraction problem is multi-dimensional and not necessarily solved by a quick social conversation. In other words, I would not expect all nominal adjustments to run so smoothly. Still, this is a nice test of an old proposition from David Hume (and others).