Today we publish a guest article written by: Matthias Scaglione and Romina Soriaco-founder of data science and economics consulting firm Motio Research. This article briefly introduces the new monthly series of U.S. household income data officially launched by Motio Research in December 2023.
Motio Research, a data science and economic consulting company, officially launched a monthly series of U.S. household income data in December 2023. The new series is based on microdata from the Current Population Survey and follows the core methodology of the discontinued series: The 2010s attracted widespread national media attention.
As co-founders of Motio Research, we believe this monthly series provides a timely measure of the economic well-being of American families. It contrasts with published official estimates, which are reported by calendar year and where there are significant delays between data collection and release.
background
In October 2011, Sentier Research, a consulting firm led by two former U.S. Census Bureau officials, began publishing a monthly U.S. household income data series based on Current Population Survey (CPS) microdata. The CPS was the primary source of U.S. labor force statistics and had until then been used to estimate monthly household income.
This new series introduces for the first time monthly estimates of U.S. median household income, derived from households' reported pre-tax money income over the past twelve months. In comparison, official census estimates of median household income are reported by calendar year, and there is a significant lag between data collection and release, with data for a specific calendar year being released in September of the following year.
The new household income estimates are widely cited by national media and policymakers as an important monthly indicator of the economic well-being of U.S. households (see sample report) here, hereand here).Sheila Bair, who chaired the FDIC during the Great Recession, chose one chart in particular, showing the Sentier Median Household Income Index. Washington post As the most important chart of 2013 (Read the article). The Sentier series led Bell to conclude in December 2013 that “for a large number of American households, the economy has yet to recover.”
The Sentier Income Series continues to provide valuable information on monthly trends in U.S. household income following the Great Recession and subsequent economic recovery through December 2019. In early 2020, Sentier announced that it was ceasing operations due to financial difficulties, leading to the suspension of the series. Since then, monthly data on U.S. household income based exclusively on CPS microdata has not been released.
In December 2023, we launched a new series of U.S. household monthly income data that follows the core methodology employed by Sentier and makes improvements that enhance the stability and predictive performance of the series.
Data and methods
Our monthly household income series data are derived from monthly CPS microdata released by the U.S. Census Bureau. The CPS is a nationally representative survey with a sample size of approximately 60,000 eligible households covering the noninstitutionalized civilian population.
The official annual median household income series is calculated using a different data set that is a supplement to the CPS called the Annual Social and Economic Supplement (ASEC). The survey, which currently has a sample size of more than 75,000 households, aims to understand the poverty and income characteristics of the population during the previous calendar year.
The income question included in the monthly CPS requires households to report total income for the past 12 months, including “income from employment, net business income, farm or rent, pensions, dividends, interest, Social Security payments and any other funds.” “Income received from all household members aged 15 or over”. Respondents were asked to choose from sixteen income ranges.
Households included in the CPS were surveyed for a total of 8 months, participated in the survey for 4 consecutive months from the time they entered the survey, were then removed from the survey for 8 consecutive months, and were finally re-included in the survey for 4 consecutive months. Respondents can be grouped according to the month in which they participated in the survey, so there are eight groups with roughly similar numbers of respondents each month.
A key issue in calculating our monthly household income estimates is that “fresh” household income data are only available for the first and fifth months in which respondents participate in the survey, and for the remaining two consecutive months (2, 3, and 4; and 6, 7 and 8) carry forward the values reported in months 1 and 5 respectively. New household income data therefore account for approximately a quarter of the monthly sample, but given the high volatility associated with series relying only on the latest data, we decided to stick with Sentier's approach to exploit the entire sample.
The first step in calculating the household income series is to estimate actual income values based on the sixteen income brackets used by households to report income over the past twelve months. For the last 15 income ranges, we follow the method used by Sentier, where income values are derived from uniformly distributed income brackets. For the top bracket, we use Pareto interpolation, where the income values are derived from the standard Pareto distribution.
Because income data for four different months over the past 12 months are included in each monthly survey, inflation adjustments must be made at the micro-data level. We adjust household income values using the seasonally adjusted Consumer Price Index for All Cities (CPI-U). Household income values are adjusted for inflation using the 12-month average of the CPI-U, based on the month in which the household participates in the survey.
The resulting series are seasonally adjusted using X-13ARIMA-SEATS software, which is produced, distributed, and maintained by the Census Bureau.
Motio Median Household Income Index
The initial series available to the public is the Motio U.S. Median Household Income Index (MHII), an inflation- and seasonally adjusted index of median household income.An interactive chart of the index is available at motioresearch.com/household-venue-series/. The index is updated shortly after the reference month's CPS filing and the U.S. Consumer Price Index are released, usually in the second to third weeks of each month following the reference month.
Starting from a value of 100.0 in January 2010, MHII reached 112.8 in January 2024, equivalent to $77,397 (see chart below). The index initially fell, reflecting the effects of the Great Recession that continued into mid-2011. The index then experienced a long-term bottom until mid-2012 and then continued to rise until the COVID-19 pandemic struck.
The surge and sharp decline from March to October 2020 is largely attributable to the impact of non-answer bias in the CPS during the early months of the pandemic. Non-response bias in this case refers to the fact that low-income households are less likely to respond to the survey (see Ward and Anne Edwards, 2021). For practical purposes, we recommend taking the February 2020 value as the 2020 peak.
Since then, with inflation at historically high levels, real median household income has fallen and has remained below its pre-COVID peak since early 2021. The index reached its lowest value post-COVID-19 in April-May 2023 and has shown strong momentum again since June 2023. In January 2024, the index was 112.8, close to the pre-COVID peak of 112.9 observed four years ago. In February 2020.
Finally, and to reiterate, we believe that tracking monthly household income trends provides a unique vantage point for understanding the changing economic well-being of U.S. households. Our commitment is to provide accurate, timely, and comprehensive data and analysis to convey this perspective thoroughly.
refer to
Jason M. Ward and Katherine Anne Edwards. 2021. “CPS Unresponsiveness During the COVID-19 Pandemic: Explanations, Scope, and Impact.” Labor Economics 72:102060. doi:10.1016/j.labeco.2021.102060.URL https://www.sciencedirect.com/science/article/pii/S0927537121000956
The author of this article is Matthias Scaglione and Romina Soria.
