Replicating Income and Poverty Census Report with ASEC data



I am trying to replicate the data shown in Table 1 in the Census Bureau’s “Income and Poverty in the United States: 2016” report using ASEC data from IPUMS. I am using the weights in the asecwth variable and cannot seem to get the same values for number of housholds and median income as the census bureau reports.

Is it possible to replicate these figures using the public use data? If so, how?


Without being able to look at your code, it is difficult to tell exactly what is happening. That said, do make note of the following detail included in the POVERTY variable description: “Users looking to replicate official poverty rates should restrict the poverty universe to all persons except for secondary individuals under 14 for 1969-1979 and all persons except for secondary individuals under 15 for 1980 and on. To facilitate this, IPUMS-CPS has made available OFFPOVUNIV.” Additional detail and documentation are listed on the Replicating Official Poverty page.

Another detail to consider is IPUMS CPS datasets are labeled by year of collection, while the ASECs represent income from the previous year. So the poverty rate for 2016 comes from the 2017 ASEC sample.

Of course, if you continue to observe a discrepancy, feel free to email with more detail about your calculations.