Swipe Up: Tech Workers and Home Values in Silicon Valley
- The typical worker at some of Silicon Valley’s leading tech companies lives in an area where the median home is worth well over $1 million.
- The typical individual who works in and around Apple’s Cupertino, California, campus now lives in a neighborhood where the typical home is worth five times more than the typical American home.
- Silicon Valley home values recovered much more strongly after the recession than home values elsewhere in the country; this is even more true in neighborhoods with high concentrations of workers from the area’s leading tech companies.
The tech industry is transforming the housing market in California’s booming Silicon Valley, as home values in the neighborhoods in which well-paid employees from the area’s technology giants tend to cluster rise more quickly than home values in surrounding neighborhoods.
The average private-sector employee working in the census tract containing Apple’s headquarters at 1 Infinite Loop in Cupertino lives in a census tract where the median home value is $1.14 million as of July 2015 (figure 1)[1], according to a Zillow analysis of home value and Census Bureau data. Similarly, the average private-sector employee working in the census tract containing the Googleplex in Mountain View lives in a census tract where the median home value is $1.28 million. Facebook workers live in areas where the median home value is about $1.25 million.
In the census tracts where Apple workers tend to live, home values are 34 percent above their pre-recession peak; in the census tracts where Googlers tend to live, home values are 29 percent above their pre-recession peak. The typical private-sector employee working in-and-around the Apple campus now lives in a neighborhood where the typical home is worth about five times more than the median home nationwide; as recently as 2010, homes in the neighborhoods where Apple workers lived were worth only three times the national median.
Interestingly, we also found that the gap between the homes where Apple workers[2] live and the homes where other San Jose-area workers live widened in summer 2007 – the same time as the release of the first-generation iPhone, a transformative product for Apple’s long-term corporate outlook. Comparing the average percent gap between the median home value of Apple-worker neighborhoods with the median home value of the entire San Jose metro area, the gap widened by 6.4 percentage points, to 20 percent, over the summer of 2007.[3] By contrast, the gap pre- and post-summer 2007 is much smaller (and not statistically significant) for Google and Facebook workers.
Methodology
By combining the U.S. Census Bureau’s Longitudinal Employer-Household Dynamics (LEHD) Origin-Destination Employment Statistics (LODES) data with Zillow’s data on home values across the country over time, we examined how home values have evolved in the areas where tech workers tend to live.
The LODES data allow us to identify – down to the census block – where workers both live and work. In places where a single, large employer dominates, we can very roughly identify the home values of workers working in and around particular corporate headquarters. Since the Census Bureau anonymizes the data to prevent the identification of specific individuals, we conduct our analysis at the census tract level, rather than the more precise census block level.[4]
For each of the tech-sector employers analyzed, we identified the census tract where the headquarters of the company they work for is located, then created weights based on the census tracts where individuals working in that tract live. We included only counts of private-sector workers when creating these weights. For example, Apple’s Cupertino campus is located in census tract 6085508101.
The total number of private-sector workers employed in the census tract containing Apple’s Cupertino campus in 2013 (as reported in the LODES data[5]) was 16,084. According to media reports, Apple’s total Cupertino headcount as of mid-2013 was about 16,000, not far off the total number of private-sector workers that the LODES data report for the census tract.
In the case of Facebook, which moved its headquarters in 2011, we used the census tract containing Facebook’s old downtown Palo Alto headquarters for 2011 and earlier, and the census tract containing its new headquarters for subsequent years. We start the Facebook series in mid-2004, roughly when the newly formed company moved into its first professional office space in downtown Palo Alto. As Facebook grew, its downtown Palo Alto footprint also grew, so its workers increasingly represented a larger share of total employment in that census tract.
Using Zillow data on home values over time, we also computed a median home value by census tract for each month. Next, we merged the Zillow data and LODES data with the Zillow monthly median home value in each census tract in April of each year, corresponding to the appropriate employer-census tract weight for that year, linearly interpolating the weights between each April. The most recent LODES data correspond to 2013, so we carried the 2013 weights forward through 2015, but will revise the analysis as new LODES data become available.
Finally, using the combined median home values and weights for each census tract over time, we calculated a weighted mean to compute a median home value time series for each employer. To smooth the series, we took a three-month, trailing, moving average.
[1] Data includes: both homeowners and renters; all home types, ranging from single-family homes to condos and coops; and workers of all levels, ranging from executives and engineers to food-service workers and janitors.
[2] We use the term “Apple workers” here as shorthand for employees working in the census tract containing Apple’s Cupertino campus.
[3] We tested this shift using a Chow Test, a statistical test that checks for structural shifts in a time series, and found that it is statistically significant at the 90 percent confidence level.
[4] Census Tracts are defined by the Census Bureau to be relatively stable geographic areas with a population between 2,500 and 8,000 people.
[5] In order to be counted as employed in the LODES data, an individual must have been employed at the reporting employer in the first and second quarters of the year.