Taiwanese residential construction experienced a massive rise and then fall in residential unit completions during the period 1992-1997. Completions rose nearly 200 per cent between 1992 and 1994 alone. A series of questions arise in relation to this phenomenon (referred to here as 'the Peak'): Why did it occur? How was this increase in construction achieved in such a short time? What were its impacts? Was it a supply-side or a demand-side phenomenon? And what are the significance and implications of the Peak? This thesis adopts a distinct methodological stance, being explanatory and interpretative rather than concerned with theory construction and testing. It is argued that this is appropriate as the Peak was a singular phenomenon, and not part of the regular fluctuations in a residential market cycle. Consequently, it is treated as an event and event-triggered, rather than as part of a continuing time-series. In addition, given that it was influenced by an enormous array of exogenous factors, including politics, the economy, and government regulations and policy, as well as by demographic, historical and geographical factors, a holistic approach is employed, as opposed to a reductionist one. The purpose is to build as complete an interpretation as possible. A major trigger for the Peak was the announcement of a residential density regulation (vohme control) in early 1992. This regulation, which threatened to dramatically reduce the profitability of development land, caused a massive acceleration in the rate of residential development approvals and ensuing completions. The regulation is viewed as interacting with a loosening of development and construction credit restrictions in late 1990 to precipitate a massive Peak in residential construction. This took glace over a very short period of time. This is analysed for each city and county in Taiwan as well as for Taiwan as a whole, the same factors applying throughout the country with the exception of Taipei City. In Taipei City, the peak occurred in a milder form in the 1980s rather than in the 1990s, due to an earlier implementation of volume control. This provides support for the overall interpretation of the thesis. After its announcement a grace period before the regulation was to come into effect allowed existing developers and landowners to propose and execute land development and construction under the old regulation. This period also induced new developers into the market, in part based on an expectation of higher unit prices resulting from volume control. This created a huge new stock of smaller-size apartments. Accompanied by falling prices, this outcome is consistent with the supply-side explanation of the Peak. Additional evidence of a supply-side phenomenon includes the high vacancies and unsold developer inventories that resulted. The rush to construction under the old regulation meant that the demand and supply market feedback loop could not effectively operate. The presence of simultaneous production peaks in all cities and counties is just one piece of evidence that this feedback loop was ineffective. A pooled cross-sectional time-series model, based on supply and demand factors of all cities and counties during 1982-1998, was used to further analyse this model of the Peak. Again the results are consistent with the supply- side interpretation. While credit loosening and volume control triggered the Peak, it is the existence of huge production flexibilities in the residential construction industry that enabled such a massive and rapid increase in construction. These flexibilities flowed, in part, from a dominant reinforced concrete construction method, as well as the Taiwanese subcontracting system, and residential construction industry networks. It is argued that these networks provided a form of quasi-public good that could be exploited. These same factors also facilitated the rapid entry of new developers into the industry. Conversely, this was associated with a decline in construction quality, as the limits of flexibility were reached. Some evidence for this was revealed by the 1999 earthquake. Imputed construction costs rose enormously during the Peak, with profits necessarily declining under the combined pressures of falling prices and rising costs. This resulted in bankruptcies and other exits from the industry. The impacts of oversupply -such as a high vacancy rate, growing developer inventories and falling prices -are expected to be long-term. The vacant stock is calculated as equivalent to ten to twelve years of pre- Peak construction. This thesis concludes that the Peak was a discontinuity phenomenon. It was triggered by a unique set of events, rather than being part of a continuous historical development. While developer behaviour was individually rational, collectively their decisions were disastrous for themselves and the industry