Testing for Price Anomalies in Sequential Sales

We thank an anonymous referee for this point.

Developers may also systematically change asking prices over time in order to meet sales goals at different stages of project completion. As in many real estate markets, sales prices in our sample are not the same as asking prices. Nonetheless, we expect developers’ pricing strategies to reflect current market conditions.

In Singapore developers can begin preselling as soon as the construction permit is obtained; in contrast, developers in other presale markets such as Hong Kong can only start selling when construction is close to completion (Fu and Qian 2014).

The anti-speculation measures announced on May 15, 1996 include a tax of 100%, 66%, and 33% on the gains from disposal if the property is sold within the first, second, and third year of purchase, respectively. In addition, buyers of residential properties are required to pay 20% of the purchase price in cash. They are also not allowed to use their pension fund to cover stamp duties related to property transactions.

Our measurement of the sequential sale effect can also account for possible missing observations due to incomplete transaction data. The transaction database is based on caveats lodged by the purchasers to protect their interest soon after an option to purchase a property is exercised. Essentially, caveats are legal documents lodged by home purchasers through their lawyers with the Singapore Land Authority to register their legal interest in the property. Typically, caveats are lodged two to three weeks after a purchaser signs an option to purchase at the show flat. Since it is not mandatory to lodge a caveat, it is technically possible that the transaction database does not include all the units sold directly by the developers. However, any omission is likely to be small in practice since most home purchases involved mortgage loans, in which case the solicitors acting on behalf of the banks would insist on lodging a caveat to protect their clients’ interest in the property. Finally, while private owners can re-sell their units before the development is completed, these transactions are not included in our data. Our measurement of the sequential sale effect allows for missing transaction data from this source as well.

The number eight is considered a lucky number because it sounds like “prosperity” in Chinese. Conversely, the number four is considered an unlucky number because it sounds like the word “death”.

The mainland of Singapore measures 50 km from east to west and 26 km from north to south.

We thank an anonymous referee for suggesting this possible explanation.

The partitions of the sample into these sequence subsamples are the partitions that maximize second stage model fit. This requires re-estimating the MNL model and second stage price equations for all feasible partitions and then optimizing over goodness-of-fit in the selection-corrected price function. For the three development phase models, the breakpoint was varied using 1 percentage point increment from first sale to the completion of the development.

The probability of a sale in the second phase based on the second phase sample is 34.02%.