Question for streeteasy

According to white paper, new development are included by SE. Case-Shiller excludes them. Stock indices exclude IPOs. I imagine this is done because new developments represent a different transaction that typically demands a discount. But SE does not discern.

Also counter to Case-Shiller, flips are included by SE (resales within 6 months). I imagine Case-Shiller excludes because of the biases they introduce. E.g., each transaction is equally-weighted, but only the profitable flips show up. The unprofitable ones are held for 15 years.

No info as to the definition of outliers. Presumably they incorrectly miss some true pops and incorrectly include some renovations.

Calling streeteasy moderator.

300_mercer, thank you for your question and inonada, thank you for your reply.

Here are some answers to your questions:
1. What is considered an outlier? We did not use a particular benchmark value or a period of time. Anything that fell outside 3 standard deviations was considered an outlier.

2. How are new developments resales treated? Yes, we included new development sales but ONLY when there was a repeat sale. We did not include the first sale purchased from the sponsor. We only included the price change between the first and second sale, and subsequent repeat sales. Case Shiller includes them after the second repeat sale.

3. Flips. Case Shiller does not include flips. Since 1995, the average holding period in Manhattan is around 3.5 years. The Manhattan market is quite different from the single-family markets covered by Case Shiller where short-term investments are more likely to happen and properties are more likely to have high turnover rates.

Feel free to reach out to Sofia who heads up research at ss@streeteasy.com with any additional questions.

Does it mean that streeteasy index is understating the recent increases as large increases are being thrown out? There are no large decreases (at least I could not find them) which are outliers, to offset them. Naturally, the decrease during 2009 would have been understated as well which means that increase from the low is probably bigger than what streeteasy condo index shows. This seems to match the anecdotal evidence.

>> Anything that fell outside 3 standard deviations was considered an outlier.

I take this to mean that out of ~5000 condo resales in a year, about 15 outliers (high or low) are thrown out. It also means that most renovations are included without taking out renovation costs.

I don't think this would explain what you want it to explain, 300_mercer. Even if all the 15 outliers are on the high side, and they are 2x the price, it would only shift the index by 0.3%.