On September 12, 2018, the International Trade Administration’s National Travel and Tourism Office (NTTO) resumed regular publication of data on the number of overseas visitors arriving into the United States for personal or business reasons and issued revised figures for January 2015 through December 2017. This followed a suspension of data releases that began on April 9, 2018, pending resolution of underlying technical issues with a significant number of records received from U.S. Customs and Border Protection (CBP).

In announcing its decision to suspend publication of overseas arrival data, NTTO said its review of 2017 arrivals records from CBP had uncovered a meaningful and increasing number of non-U.S. citizens traveling on visas to the United States being categorized as U.S. residents. These travelers were consequently removed from the visitor count of overseas travelers arriving into the United States, resulting in an undercount of overseas arrivals.

BEA uses the I-94 arrivals data to calculate foreign trade statistics. Expenditures by foreign visitors are recorded as U.S. exports of “travel (for all purposes including education).” The I-94 arrivals data are also used to measure exports of air passenger transport, the fares paid by foreign visitors to U.S. air carriers for transportation to and from the United States. These travel and transport exports are included in BEA’s quarterly U.S. International Transactions Accounts and its monthly U.S. International Trade in Goods and Services report. The trade statistics are subsequently used in the National Income and Product Accounts (NIPAs), where the expenditures of foreign residents are recorded in Gross Domestic Product (GDP) as exports of services. In addition, these expenditures are used to derive Personal Consumption Expenditures (PCE). Because there is a lag between the month of overseas arrivals and when BEA receives the I-94 source data for that month, BEA routinely estimates overseas visitors using historical trends in the I-94 data series.

BEA worked closely with NTTO to assess the extent of the undercount of foreign visitors and to determine a methodology for adjusting the data. BEA and NTTO determined that country of citizenship is a reliable proxy for country of residence. Therefore, BEA adjusted the foreign visitor counts beginning with July 2016 by substituting the country of citizenship for the country of residence in the affected records. For January 2015–June 2016, country of citizenship information was not available to BEA, so BEA adjusted foreign visitor counts by redistributing visitors erroneously categorized as U.S. residents to other countries based on each country’s share of total visitors. These adjustments resulted in upward revisions to “other” business and “other” personal travel exports for 2015–2017 released in the annual update of the U.S. International Transactions Accounts on June 20, 2018.

BEA will incorporate NTTO’s revised I-94 arrivals data into its estimates of travel exports for 2015-2017 with the annual update of the International Trade in Goods and Services and the International Transactions Accounts in June 2019, along with other planned improvements to methodology and source data. Incorporating the revised I-94 arrivals data will likely result in small revisions to travel exports of less than +/- 0.5 percent per quarter.