Hi Aidan,
I'm not an expert in this topic, but the lopsided distribution seems plausuble, assuming a lot of Grand Rapids households have modest incomes.
One thing you might consider: the ACS has a variable called GRPIP that already calculates rent as a share of income for you, and the CB publishes housing cost burden tables from it. Would those tables help answer your question? This alternative method only makes sense if you're estimating housing cost burden rather than constructing a synthetic affordability distribution.
If you stick with your custom calculation, make sure you've filtered out group quarters households, which can skew the results in some cities. Also, note that many households may have zero or negative income, which could be contributing to the distribution you're seeing.
------------------------------
Mark Mather
Associate VP
PRB
------------------------------
Original Message:
Sent: 05-20-2026 12:45 PM
From: Aidan Rozema
Subject: Am I messing up here?
Hello all! I am a student who is still learning the ropes of using federal data. I have been trying to construct a distribution of what people in Grand Rapids, Michigan could reasonably afford in rent using the 5-year American Community Survey (sourcing data from IPUMS) by calculating HHINCOME / 12 * 0.3. I know that many Americans have low-incomes, but I keep ending up with this really lopsided distribution. I was wondering if anyone has tried something similar, and if they have any tips. I'm new to this, so I have the sinking feeling that I messed up somewhere...
------------------------------
Aidan Rozema
Student Researcher
University of Michigan
------------------------------