Thursday, October 3, 2019

NYC Market Analysis -- Housing vs. Condo

(Click on the image to enlarge)

Michael is interviewing for a Housing Analyst position with a major economic research firm.

Question # 1
Interviewer: Though we are using the two Case-Shiller data series, let's assume for the sake of this interview -- they represent only the City of New York (NYC). Given that assumption, are these two market segments comparable?

Michael: No. The NYC housing market comprises the outer boroughs, meaning Brooklyn, Bronx, Queens and Staten Island, whereas the condo market represents primarily the Manhattan market, along with some newer condos in Brooklyn. Nonetheless, due to the shared geography, the two data series would be highly correlated. 

Question # 2
Interviewer: Use your best quantitative judgment and guess the level of correlation between these two data series.

Michael: (After a short pause) Going by the r-squared, I would say it would be at least 0.82.

Interviewer: Please give a minute so I can find out how close you are to the actual coefficient. Yup, it's 0.83976. I will say you are very close. Very impressive! 

Question # 3
Interviewer: The condo data shows a sharp fall in June-July, 2018. What would you attribute that fall to?

Michael: It's just an aberration in the data and has nothing to do with any drastic change in economic fundamentals. That's reason the data comes in and gets cleansed. That is why you see a 3-point jump in August, 2018 which is unrealistic too. In any case, any trendline would iron out that aberration in the data.  

Question # 4
Interviewer: Are the demand drivers very similar in those two market segments?

Michael: No. The NYC housing demand is driven by the local economic fundamentals, so it's a natural market. The Manhattan condo market, on the other hand, is heavily dependent on global demand, especially Asia and Europe. Therefore, any slowdown there heavily impacts the Manhattan condo market.

Question # 5
Interviewer: What would be the statistical impact if we remove those two months from the data series?

Michael: Once you remove outliers from the data, stats get better. In this case, the correlation coefficient would jump. 

Question # 6
Interviewer: In that case, can you guess the new coefficient without those outlier months?

Michael: In the vicinity of 0.87. 

Interviewer: Give me 15 seconds and let me find out. Actually, it jumps to 0.8780. Very close! Congrats!

Question # 7
Interviewer: Can you explain why the housing market has been holding steady while the condo market has been trending down?

Michael: The local market fundamentals are good so the housing market has been holding steady. There are two basic reasons why the Manhattan condo market has been trending down: (a) the cap on SALT, and (b) the tumbling of foreign buying.

Question # 8
Interviewer: Take a look at the above SD/COV table and tell me which segment is less volatile and why?

Michael: The condo market segment is less volatile as shown by both volatility measures meaning Standard Deviation and Coefficient of Variation (COV). The reason is simple: The condo market has been range bound since Q4-2017.

Question # 9
Interviewer: Would you recommend one or both of these segments to our clients? And, what is your forecast for these two market segments?

Michael: Given the low interest rate environment and good local fundamentals, I would recommend the housing market. Considering the reasons I previously talked about, I will not recommend the Manhattan condo market.

The housing market will remain sideways in Q3-Q4 of 2019, declining in Q1, 2020+, until the presidential polls start to project some meaningful directions. Of course, a correction is long overdue and is always healthy. The Manhattan condo market will continue to trend downward and I won't be surprised if the January, 2017 level is retested by the end of this year.

P.S. These are Case-Shiller’s seasonally-adjusted indices so the month-over-month comparison is fine. While using Case-Shiller’s seasonally unadjusted indices, one should compare July-2019 with July-2018 and July-2017, etc. 

-Sid Som
President, Homequant, Inc.

No comments:

Post a Comment