Today, the U.S. Census Bureau announced that a net of 2.8 million people migrated to the United States between 2023 and 2024. This is significantly higher than our previous estimates, in large part because we’ve improved our methodology to better capture the recent fluctuations in net international migration.

The Census Bureau continuously looks for ways to improve our methods to keep pace with changes in the U.S. population. Over the past several years, we have conducted research to make the methodology for estimating net international migration (NIM) more flexible and responsive to short-term fluctuations in migration.

For example, in recent years, there have been several notable fluctuations in international migration flows to and from the United States. At the height of the COVID-19 pandemic, there was a steep decline in immigration as countries restricted travel due to public health concerns. This was followed by a subsequent increase in immigration to the United States, especially among migrants seeking asylum. Accurately measuring these flows is important for developing the Census Bureau’s official population estimates.

To develop the Vintage 2024 population estimates released today and to better reflect current trends and potentially underrepresented populations, we used newly available administrative data to adjust the usually survey-based estimates of foreign-born immigration.

As usual, we also updated the estimates for prior years in the vintage – in this case back to 2020. Because of routine updates to the survey data and the improved methodology, the Vintage 2024 estimates for the 2022 and 2023 periods are also picking up on the recent, significant increase in NIM. 

In this blog, we describe the Vintage 2024 results and explain how we estimate NIM, including how we updated our methodology to improve the most recent estimates. 

First, let us provide some background on how we produce the annual population estimates. Each year we produce a new “vintage” of population estimates that consists of a time series from the current year back to the most recent census.

Typically, we develop the estimates beginning with the latest census as the base. For this decade, we’ve used a base that blends 2020 Census data with Vintage 2020 population estimates and the results of the 2020 Demographic Analysis. We then add or subtract current data on births, deaths and migration to the base population to reflect annual change.

Some of the data we use to create the estimates have a time lag. When we create a new vintage, we can incorporate newly available survey data and vital statistics into the estimates for prior years, in addition to data from new sources.

This process allows us to produce estimates in a timely fashion and to make them more accurate as more data become available. Because it’s more accurate, the new vintage supersedes prior vintages and can have different estimates for earlier years in the time series. 

The Census Bureau today released the first set of Vintage 2024 population estimates, which shows the national and state population totals as of July 1, 2024, and updated estimates for each year back to 2020.

The Vintage 2024 estimates of international migration are considerably higher than the Vintage 2023 population estimates (Table 1). We usually caution data users against combining data from different vintages within a single analysis because the most recent vintage supersedes any prior vintages. However, making comparisons across vintages – as we do below – is appropriate for assessing the impact of updates to the methodology or input data.

Table 1 reports the NIM estimates, which include the following components (not published separately):

• Foreign-born immigration (also referred to as “non-U.S.-born immigration” in our methodology statement) – Foreign-born people moving into the United States.
• Foreign-born emigration – Foreign-born people moving out of the United States.
• Net native migration – The net of U.S.-born people moving into and out of the United States.
• Net migration between the United States and Puerto Rico – The net of people moving between stateside and Puerto Rico.  
• The migration of the population born abroad of U.S. citizen parents.
• Movement of U.S. armed forces to and from other countries. 

The Vintage 2024 NIM estimate is the same as the Vintage 2023 estimate for the first two periods. But it is 69.5% higher in the third period (July 1, 2021, to June 30, 2022) and 101.7% higher in the fourth (July 1, 2022, to June 30, 2023). This increase was driven almost entirely by the estimate of foreign-born immigration, the largest component of the NIM estimates.

The NIM estimate for the July 1, 2023, to June 30, 2024, period is 2,785,517, which is the highest annual amount in the Vintage 2024 time series. 

The Census Bureau produces NIM estimates at the national, state and county geographic levels by age, sex, race and Hispanic origin. This is a significant challenge, and no other federal agency produces estimates of international migration with this much geographic and demographic detail. 

International migration is difficult to estimate because of its complexity and dynamic nature. The data available to measure international migration are often limited and the methods used to produce them require a number of assumptions (some of which are detailed in our recent working paper).

We use data from the American Community Survey (ACS) to produce estimates of foreign-born immigration. The ACS is an annual survey of approximately 3.5 million households, and it includes people living in housing units and group quarters. The ACS provides information about the demographic, social, economic and housing characteristics of the U.S. population and for small geographic levels.

We use ACS data on the foreign-born population whose residence one year ago (ROYA) was abroad to estimate immigration flows. This method has worked well during periods of relatively stable international migration (as we’ll demonstrate in a figure below). However, it has two limitations that could impact the accuracy of the estimates, particularly during periods of short-term fluctuations:

• The ACS data that are available when we produce the NIM estimates are at least 1 year old and do not reflect the most current trends. For example, the 2023 ACS 1-year file was the most recent file available when we produced the Vintage 2024 estimates.
• Recent immigrants may not be fully represented in the ACS data because of the ACS sampling frame and other coverage issues. Research indicates some foreign-born populations are underrepresented in the ACS. 

To overcome these limitations, we used newly available administrative data to adjust the ACS-based estimates of foreign-born immigration to better capture current trends and potentially underrepresented populations.  

Over the past several years, we have conducted research to make the NIM methodology more flexible and responsive to short-term fluctuations in migration. This process involves assessing various administrative data sources for validity, availability, completeness and the frequency at which the data are available. Subsequently, these data are incorporated into something we call the “Benchmark Database” that we use to assess changes in migration flows.

As part of this ongoing research, we recently refined our Benchmark Database. Currently, the Benchmark Database is compiled using U.S. federal administrative data on:

• Lawful permanent residents.
• Visa issuances.
• International student enrollees.
• Refugee admissions.
• Border releases.

Together, this database serves as an independent point of reference for the levels of immigration flows to the United States.

It is important to note that while administrative data may include inconsistencies, such as double counting, we partnered with leading migration experts to conduct a comprehensive evaluation of the strengths and weaknesses of the data sources. This led to major improvements in our Benchmark Database, giving us a better sense of current patterns of migration flows.

The Benchmark Database is a valuable tool for monitoring short-term changes in inflows and allows us to adjust our ACS-based estimates when appropriate. Additionally, our analysis sought to identify which cohorts were not being adequately captured. It revealed that recent humanitarian migrants, the group with the most significant growth in our benchmark data, were also the least likely to be included in the ACS.

To produce the Vintage 2024 estimates, we adjusted our ROYA-based, foreign-born immigration estimates upward to account for 75% of the humanitarian migrants in our Benchmark Database. We arrived at this adjustment factor after synthesizing different analyses of coverage error in the ACS and in consultation with internal and external migration experts. As noted in Table 1, this adjustment increased NIM estimates for the period of July 1, 2021, to June 30, 2023, in Vintage 2024 relative to the value from Vintage 2023.

This adjustment for humanitarian migrants was applied to the national total and then distributed down to states and counties using our usual method. As a result, the state and county estimates are also impacted by the adjustment. (Per our usual schedule, county level estimates will be released in March.)

Figure 1 displays the total number of foreign-born immigrants by year using the administrative data (bars), the unadjusted ACS-based, foreign-born immigration estimate (solid line) and the adjusted ACS-based, foreign-born immigration estimate (dotted line). 

As mentioned, administrative data were also used during the COVID-19 pandemic to make adjustments to account for fluctuations in immigration flows that weren’t yet picked up in the ACS.  

During the pandemic, immigration to the United States slowed considerably. But the ACS data we had available were from 2019 and, thus, did not show this trend. Therefore, we used administrative data from the Department of Homeland Security, Department of State, and other federal agencies to adjust the ACS-based estimates downward in Vintage 2021 and then upward in Vintage 2022.  

In Vintage 2023, the foreign-born immigration estimates in the ACS returned to prepandemic levels. However, there was a lot of migration activity at the United States-Mexico border. At that time, there was uncertainty about how many of these migrants were being allowed into the United States.

Additionally, a key immigration policy that limited migration because of public health concerns ended in May 2023. The Vintage 2023 population estimates extended to June 30, 2023, which was only one month after the policy ended. Consequently, we did not adjust the ACS-based estimates for Vintage 2023. However, subsequent administrative data showed that an adjustment was needed.

For Vintage 2024, we again adjusted the ACS-based, foreign-born immigration population using administrative data to reflect the continued increase in international migration, especially for humanitarian migrants. This resulted in a net international migration estimate of 2.8 million from July 1, 2023, to June 30, 2024.

International migration flows are a significant contributor to population growth and demographic change in the United States. However, data on recent migrants are often limited, making accurately estimating and projecting international migrant flows very challenging. 

The Census Bureau’s current method leverages survey and administrative data. We’ve developed this approach over decades of research and regular consultation with experts from academia and other federal agencies, and we continue to refine our methods. Currently, we’re researching the coverage of the foreign-born population in the ACS and developing adjustments by demographic characteristics and subnational geographic levels. Each vintage, we will do what we can to improve the international migration components of the population estimates to more accurately reflect current trends.