The human loss from the coronavirus will not be reflected in the 2020 census because of a matter of timing, which could save a congressional seat for New York but cost Alabama one.

Because the start of the pandemic in the U.S. and the April 1 reference date used for the census fell so close to each other last year, the deaths that began in mid-March will not show up in the state population figures that determine political representation in Congress.

The timing will paper over the losses from the virus, which has killed around 44,000 people in New York state, including concentrations in some New York City neighborhoods in Brooklyn, Queens and the Bronx. Alabama has reported around 8,000 virus-related deaths.

New York still is projected to lose at least one seat, but the quirk in the calendar should ensure that the state gets the last of the 435 congressional seats by a margin of more than 20,000 people, and that would save it from losing a second congressional seat, said Kimball Brace, a redistricting expert at Election Data Services.

“When you get to that last seat or two, any little change could have an impact,” Brace said.

The once-a-decade headcount of every U.S. resident determines the number of congressional seats, and Electoral College votes, each state gets. Redistricting experts estimate that 10 congressional seats will shift among 17 states when the Census Bureau releases apportionment numbers by April 30.

The division of congressional seats is sometimes decided by relatively small numbers — just thousands or even hundreds of people.

Brace drew the conclusions using population estimates released in December. Getting that last available spot during the apportionment process would keep New York from losing its 26th congressional seat, but it would cost Alabama its seventh seat in the House.

The April 1 reference means people were required to use that date in answering census questions. If the reference date had been July 1, Alabama would keep its seventh seat by a more than 6,200-person margin. But that would cost New York a second seat, reducing its number of representatives in the House from 27 to 25, according to Brace’s analysis. By the start of July last year, New York had experienced more than 32,000 virus-related deaths.

Brace cautions that those scenarios are tempered by the accuracy of the December estimates and the 2020 census, which took place amid the pandemic and natural disasters. The count was also dogged by concerns that the participation of immigrants and Latinos could be suppressed by the Trump administration’s failed efforts to exclude people who are in the country illegally from the process of allocating congressional seats.

Both Brace and William Frey of the Brookings Institution predict that Texas will gain three seats, Florida two seats, and an extra seat each will go to Arizona, Colorado, Montana, North Carolina and Oregon.

On the flip side, Alabama, California, Illinois, Michigan, Minnesota, New York, Ohio, Pennsylvania, Rhode Island and West Virginia each stand to lose a congressional seat, according to Frey and Brace, though Brace estimates that New York could lose up to two seats.

The loss would be a first for California, the nation’s most populous state.

Other states have close margins for either gaining or losing a seat. Montana’s seat gain would be by fewer than 5,000 people, and Rhode Island is expected to lose a congressional seat by fewer than 17,000 people, according to Brace.

Even though the numbers will not be released for months, Alabama has been fighting to save its seventh congressional seat through a lawsuit filed more than two years ago. The case seeks to exclude people who are in the country illegally from the apportionment process, even though the Constitution spells out that every person in each state should be counted. President Donald Trump issued a similar directive last year, but President Joe Biden rescinded it upon taking office.

Including people in the U.S. illegally “will place Alabama at substantial risk of losing political representation,” the state of Alabama said Thursday in a court filing.

Alabama officials are optimistic their push to get residents to participate in the census made a difference, said Mike Presley, a spokesman for the Alabama Department of Economic and Community Affairs.

“When the Census Bureau releases the data from the 2020 count, we are hopeful that the state’s strong effort, which resulted in a higher self-response rate than in 2010, will lead to positive news for Alabama,” Presley said.

In New York, where more than 40% of residents live in New York City, the virus made it more difficult to reach hard-to-count minority communities. Advocacy groups were prevented from going door to door to convince people to answer questionnaires, and many residents of wealthy neighborhoods fled the city during the census, said John Mollenkopf, director of the Center for Urban Research at the City University of New York.

Plus, New York City has been undercounted in past censuses, countering any advantage New York may have had because of the April 1 reference date, said Melva Miller, CEO of the Association for a Better New York.

“We know New York City is a terribly hard place to count, and numerous people are historically undercounted,” Miller said. “In terms of whether we did better because we might have had some people counted, I can’t see that is the case.”