kiwiPhrases/rentDistributionChanges

I compare the distribution of rents between 2009-2013 and 2013-2017 ACS surveys at the block group level in LA County. I believe this is a better metric of change than comparing changes in median rents because my metric explicitly accounts for the disappearance of low cost units and their replacement with high cost units. I want an index that shows whether the rent distribution has shifted upward or downward across the years but I do not have micro level data. Moreover, using this index, I wish to compare the size and direction of the shift between geographies so that geographies with a high index experienced a rightward shift in the rent distribution (thus became more expensive) while geographies with a low index experienced a leftward shift in the rent distribution (thus became less expensive). I am not aware of any such metrics so I propose one here and map it. In order to construct a metric for distributional change without actually having micro-level data, I 1) normalize bins for each block group, 2) take the difference between the 2017 and 2013 bins, and 3) take the difference between changes in the top two bins against the changes in the bottom two bins. Let b index the 8 bins (ie rent300 to 599 is bin 2, rent 2000 or more is bin 8), i index block groups, and t index survey years . In order to avoid overlap, I exclude the ACS 2011-2015 from the index so in effect, I only use the 2009-2013 and 2013-2017 ACS surveys for the index.

Jupyter Notebook