Statistical Analysis for rents

[Randy Janoe]

Hopefully some on here are interested in quant or statistics and how it can be applied to rents. If it intimidates you, don't worry, I will attempt to be concise and provide a laymens terms summary.

Basic statistics crash course:

Obtain as many sample rent data as possible (Zillow works)

Define alpha for null hypothesis testing ( 0.05)

Run multiple regression analysis (#bed, #bath, sq ft) against active rental list prices.

Observe P-values for each metric

Result:

#Bed - 0.02

#Bath - 0.04

Sq ft - 0.02

All are below or equal to 0.05, reject null hypothesis (statistically significant).

Create multiple regression line using coefficients for each metric.

Input variables from project property and solve for rent!

Summary:

It makes sense directionally that a 3/2 1400 sq ft rents higher than a 2/1 800 sq feet, right? 

Have you ever looked at a rental comp and said well this one is 100 more sq feet or mine is a 3 bedroom and not a 2? Those comps may feel like poor properties to compare to but they are valuable data points.

The regression line will answer exactly what the extra 100 sq ft or bedroom is worth!

This can prevent two extremes: under pricing below market and over pricing and having an idle asset.

I believe this method will provide a proxy rent strong enough to do deal analysis with, particularly in a well diversified portfolio.

Let math do the heavy work and massage as necessary for finishes and how fast (or slow) the unit needs to be rented.