The cyclical variation in office construction, vacancies, rents and values over the last decade has been enormous throughout the world. Special factors magnified the variation in numerous countries --the U.S. savings and loan debacle and the Finish reaction to the collapse of the USSR to name two, but the variation has been large in almost all countries (Goetzmann and Wachter, 1995, and Renaud, 1994) . Australian office markets are a good example. A major construction boom developed in the

more »

... e and late 1980s, causing large quantities of office space to come on line in the early 1990s. When this coincided with a severe recession, vacancy rates skyrocketed and real effective rents plummeted. The vacancy rates in Sydney, Melbourne, and Perth all rose by over 20 percentage points between 1989 and 1992. The likelihood of an extended period of excessive vacant space and diminished cash flows on occupied space collapsed values. The recovery since then has also been strong. By the end of 1995, over half of the increase in vacancy rates in these cities had been reversed. Some have attributed the severity of property cycles to developers building too late in the boom and well into the bust, attributing the error to either prolifigate lenders or egomaniacal developers. Recent work by Grenadier questions this error hypothesis. Given the value of the option to wait before developing and the high costs of adjusting to shifts in space demand, he argues that overbuilding in markets for assets with long gestation periods is consistent with optimal development strategies (Grenadier, 1995 and 1996) . While developers may well act optimally, we doubt that the observed severity of property cycles can be attributed solely to the option to wait and high adjustment costs. Our analysis of the Sydney office market suggests a major role for the failure of investors to understand the workings