The U.S. census's 72-year confidentiality rule has a strange history

It's a rule that many genealogists plan their lives around.

Once a decade, the U.S. Census Bureau tries to gather the names, home addresses and other details of every person living in the country for a head count.

And 72 years after a national tally's Census Day, records with all of that information are shared with the public, including family historians eager to flesh out their genealogy charts.

This policy — called the "72-Year Rule" — was enshrined into law in 1978 and has become part of the current promise of confidentiality the bureau relies on to persuade households to get counted.

This year, on April 1, files from the landmark 1950 census — the first U.S. count to include baby boomers — will be released for the first time.

But exactly why seven and a fifth decades became the specific length of time to keep census records confidential has stumped researchers — including Jessie Kratz, the historian of the National Archives, which is in charge of making census records public.

"Everyone just said, 'You know, 72 years is a lifespan,' " Kratz recalls of the commonly cited explanation she first heard after starting to work at the federal government's archives.

But the average life expectancy in the U.S. was closer to 73 years around the time the 72-Year Rule became federal law, the National Center for Health Statistics reported.

72 years may be the result of bureaucratic happenstance

When asked for an interview about the 72 years' origins, the Census Bureau directed NPR to a webpage with no clear answers. The website does, however, highlight a 1952 exchange of letters upon which the 1978 law is based.

Sent between Wayne Grover, the archivist of the United States at the time, and former Census Bureau Director Roy Peel, the letters specify an agreement that following "seventy-two years from the enumeration date of a decennial census, the National Archives and Records Service may disclose information contained in these records for use in legitimate historical, genealogical or other worth-while research."

Kratz and some longtime census watchers say the letters are part of a spotty trail of clues for their theory about what's behind 72 years — bureaucratic happenstance.

The theory goes back to the early days of the National Archives, when the fledgling federal agency, Kratz says, "had to go around convincing all these agencies that, in fact, we were the correct repository for these federal records."

In 1934, the year the National Archives was established, the Census Bureau was not convinced. It had already set up its own system for accessing the information stored in Washington, D.C.

"These records, which cannot be replaced if lost or destroyed, would be of little avail to the public if transferred to the Bureau of Archives without keeping them in that building under the direct supervision of the Director of the Census and provision being made for a number of searchers and correspondence clerks," Census Bureau Director William Lane Austin wrote in a 1934 memo to the commerce secretary, who oversaw the bureau.

But less than a decade later — during World War II and in the same year of the bureau's move from downtown Washington to suburban Suitland, Md. — officials at the bureau had apparently changed their minds.

"Census records through the year 1870 are not considered confidential and may be used for legitimate research purposes by private individuals," read a National Archives press release announcing the transfer of the earliest batch of census records.

"The National Archives, like the Census Bureau, however, is forced by the pressure of war work to confine its research on these records to war-related requests such as those involving birth data needed for enlistments or by workers in war industries."

The year was 1942 — 72 years after the 1870 census.

Kratz, the National Archives historian, suspects officials at the time were using that 72-year timespan as a guide for when information could be disclosed as part of the 1952 agreement the Archives and the bureau struck over the regular transfer of census records.

A disagreement stalled the release of the 1900 records

"That would be my biggest guess, although I have not seen a piece of paper that says that," she says about the theory of why the bureau's director proposed "the lapse of seventy-two years" in 1952 (when, the U.S. surgeon general at the time told newspaper readers, average life expectancy was 68 years).

Kratz recently wrote a blog post that details a disagreement that emerged in the 1970s over how long census records should be kept confidential and who in the public could get access. It stalled the release of the 1900 census records until 1973, and in the end, the dispute was settled by Congress with the 1978 law.

Still, during a 1973 congressional hearing, Archivist James Rhoads told lawmakers that the National Archives did not "find any evidence in the files specifically as to why 72 years was picked" for the 1952 agreement.

More than a couple of years later at a 1975 hearing, the acting archivist at the time, James O'Neill, appeared to back up Kratz's theory.

"The 1870 census records were made available when they were transferred to the Archives in 1942, 72 years after the census was taken. This established the 72-year precedent for restrictions on population census records," O'Neill said.

There's a case to be made for a 77-year rule

For David McMillen, a former Census Bureau employee who later worked for the National Archives and has tried tracking down this history, the question of "why 72 years?" remains open.

"It's a funny one, isn't it?" McMillen says. "There is not a definitive answer to that question."

Margo Anderson, the leading historian of the U.S. census and author of The American Census: A Social History, notes that this is a common conundrum when digging into the decisions surrounding records for the once-a-decade counts.

"The people who did it the last time may not be around anymore," adds Anderson, who has also researched the complicated history of the lifting of privacy protections during World War II that allowed the use of census records to identify people of Japanese descent living in the Washington, D.C., area.

What is clear today is that the 1978 law does allow the 72-Year Rule to be renegotiated between the Census Bureau director and the archivist of the United States.

There are no signs of any active talks.

But if a person's lifespan is the rule of thumb the government wants to use, the latest data on average life expectancy in the U.S. could help make a case for a new rule for how long to shield census records: 77 years.

Copyright 2022 NPR. To see more, visit https://www.npr.org.