Follow Insights Blog


CoreLogic Econ


Google Searches Shed Light on Rental Prices – Part I

National Average Rental Prices Unlikely to Fall This Summer

Bin He    |    Housing Trends

In a white paper published in 2009, Google claimed that Google Trends is able to help in predicting the present, in the sense that search volume data may be correlated with the current level of economic activity1. A growing number of academic studies have demonstrated that Google Trends can provide insights into home prices and Consumer Sentiment Index2, 3. Using CoreLogic rental price data4 and Google Trends for the category “Apartment & Residential Rentals,” a new analysis from CoreLogic shows a strong correlation between Google search interest and real rental prices (dollar per square foot)5. In Part I of this series we focus on the national-level analysis, and in Part II we will expand the analysis to several large metro areas.

Decomposition of Google Trends

Decomposition of Google Trends

Google Trends data does not report raw search volume; instead it provides shares of queries for a given term (or several terms) relative to all search terms in a geographic area. Google Trends for the category Apartment & Residential Rentals reflects search interest for terms including rent, apartment for rent, houses for rent and other similar terms. In the CoreLogic study, the Google Trends data is normalized so that it starts at 0 in January 2012, with the data following this date indicating the share of queries relative to the baseline. Figure 1 shows the Google Trend for Apartment & Residential Rentals and the real rental price per square foot at the national level. Even though the correlation between these two time series was somewhat masked by significant seasonal patterns in the search data, we can see that as the search interests for Apartment & Residential Rentals increases, so does the rental price per square foot. It is not surprising to see the close relationship between rental prices and Google Trends since the search interest for rentals will eventually translate into actual demand for rentals. As the demand goes higher, so do prices. However, Google search interest does not provide a measure of the supply of rental properties, and hence this data alone cannot possibly explain all of the dynamics in rental prices since this is a reflection of the intersection of tenant demand and landlord supply.

As mentioned, there is a clear seasonal pattern in the search interests for Apartment & Residential Rentals. As Figure 2 shows, by applying a structural time series model, the observed Google Trends can be decomposed into four components: a long-term trend, a repetitive seasonality, a cyclical pattern and outlying random errors. The seasonality pattern shows search interests for rentals ramps up during the spring and summer and peaks in July, in part due to student demand as they graduate and look for housing or prepare for housing at the start of fall semester.

Real rental price

Real rental price apprec

The ultimate goal is to compare the long-term trend in Google search interest, which is the red line in Figure 2, to the real rental price change over time. In order to do so, we normalized the real rental price per square foot so that it has the same definition as the normalized Google Trends. The normalized real rental price represents the real rental price appreciation since January 2012. In Figure 3 we plotted real rental price appreciation since January 2012 and the long-term trend in Google search interest in the same chart. From January 2012 to February 2016, the long-term Google search interest for rentals has increased by 15 percent while in the same timeframe the average real rental price has increased from $0.96 per square foot to $1.11 per square foot, which is a 15.6 percent appreciation. The correlation between the real rental price appreciation and long-term search interest registers at 0.89, signaling a very strong relationship (1 indicates a perfect correlation). From Figure 3 we can also observe that the rental price lagged the Google Trends by a few months. The Google search interest was flat until September 2012 and then started to rise. On the other hand, the rental price did not pick up until the beginning of 2013 – indicative of the time required to translate search interests into actual demand. Based on the Google Trends we see in the first four months of 2016 and the fact that supply is still not picking up, rental prices are unlikely to fall in the near future.

1 Hyunyoung Choi, Hal Varian, Predicting the Present with Google Trends, Google Inc., 2009.

2Lynn Wu, Erik Brynjolfsson, The Future of Prediction: How Google Searches Foreshadow Housing Prices and Sales, 2009.

3Marian Alexander Dietzel, Nicole Braun, Wolfgang Schaefers, Sentiment-Based Commercial Real Estate Forecasting with Google Search Volume Data, Journal of Property Investment & Finance, 2014.

4CoreLogic rental trends February 2016 data was used in this study. The rental trends represent the whole market.

5Nominal rental price is adjusted to real rental price using Consumer Price Index (CPI) less shelter to exclude the portion of rental growth simply due to general inflation. The CPI less shelter has not changed too much from 2012 to 2016, so results are robust for both real and nominal rental prices.

© 2016 CoreLogic, Inc. All rights reserved