Arbitrage in the Philadelphia Housing Market

After nearly six months of searching, my wife Sophie and I recently bought a home in Philadelphia. Over this period, we learned about each other’s preferences and how to communicate those preferences peacefully. To prioritize our time, we decided to develop a checklist to evaluate each house. We developed a more nuanced understanding of what we were looking for through many conversations and property tours. In response, our checklist grew from eight criteria and ended with 14. For simplicity, we equal-weighted our checklist criteria and focused our list on immutable aspects, things that cannot be changed with money. Our final checklist is described below:

1. Location

2. Neighborhood

3. Immediate surroundings

4. School quality

5. Commuting to Radnor and Wharton

6. Southern exposure

7. Layout of home amenable to kids, remote work areas, guest room

8. Outdoor space

9. Parking

10. Distance to a Park

11. No obvious water damage

12. Owner-occupied

13. Proximity to grocery store

14. Noise level

Tactically our approach was to use Zillow at 6:00 am every morning and monitor for homes that met at least 80% of our checklist criteria. If we found a home, we’d try and see it that day. Apparently, other buyers had a similar approach. For example, we identified a home in Fitler Square early one Saturday morning in June that met about 90% of our criteria. We toured the property later that morning and made an offer by late afternoon. Two other buyers made offers, which triggered escalation clauses, and another buyer was willing to pay more than we were.

Somewhat discouraged, we continued looking. About two months later, Sophie happened to stumble across an intriguing opportunity. While reading an email from Redfin suggesting a few homes that met our browsing pattern, she found a home that looked interesting. She cross-referenced it with Zillow and to her shock, realized that the home was incorrectly listed almost a mile away from the actual location, in a different (and incorrect) school zone. After finding this data inconsistency we immediately went to tour the property, found it met more than 90% of our criteria and made an offer for the asking price later that day. The next morning, the seller accepted our offer. We had no competition, there were no other offers.

After finding this inconsistency, we further cross-checked the USPS Address data and Philadelphia Property Tax data. Again, both of these data sets had substantial inconsistencies. Even Google Maps did not have a record of the property. After speaking with our agent, we learned that many properties in historic neighborhoods like Logan Square, Old City, Rittenhouse Square, and Fitler Square exhibit this pattern. Because Philadelphia is hundreds of years old, and many plots have been subdivided many times. There is no complete unified database of all subdivisions. When subdivided plots go on sale, Zillow pulls data from MLS and occasionally parses the locations incorrectly. When this happens, you should move fast!

Red circles indicate potential arbitrage opportunities.

Now that we purchased the property, we are going to do a bit of shareholder activism to effectuate a liquidity arbitrage. We need to update Google Maps and fix the inconsistencies that lead to Zillow misidentifying the home. Assuming we fix the data errors, we can greatly improve the odds that more potential buyers will see the home relative to when we purchased it. More buyers means more liquidity and as Sam Zell always says, “liquidity equals value.”