The holding of vacant land has an option value. Existing theory reckons a linkage between the expected wait and the stochastic nature of development costs, but few empirical studies can address it due to data availability. This research merges developer characteristics with complete microdata on land acquisitions and developments of an emerging district in Taipei, and it adopts the Accelerated Failure Time model to examine the timing to build. The expected wait is longer for developers with more paid-in capital or with less development capacity. Most significantly, those involved in subsidiarization tend to shorten the wait. The results are robust under alternative specifications and consistent with the theoretical predictions of costs. Additionally, a policy incentive on the plot ratio substantially shortened the district's development process. The data have no censoring and truncation issues typically encountered in survival analysis. Further tests shed light on the severity of biases the estimation would suffer if the data had those issues.

持有闲置土地具有选择价值。现有理论认为预期的等待时间和开发成本的随机性之间存在联系，但是由于数据的可用性，很少有经验研究可以解决这一问题。这项研究将开发商的特征与有关台北新兴地区土地收购和开发的完整微数据相结合，并采用了加速失效时间模型来检验建造时间。对于具有更多实收资本或具有较小开发能力的开发商，预期等待时间更长。最重要的是，那些参与子公司化的人往往会缩短等待时间。在替代规范下，结果是可靠的，并且与成本的理论预测一致。此外，对地积比率的政策激励大大缩短了该地区的 的开发过程。数据没有生存分析中通常遇到的审查和截断问题。进一步的测试揭示了偏差的严重性，如果数据存在这些问题，估计将遭受损失。