Nearly 5,000 registered voters in Rhode Island appear on a federal list of dead people, according to a Providence Journal analysis of the state’s central voter registration list.

The Journal compared the Social Security Administration’s Death Master File to the state’s voter list and found 4,991 matches. The presence of dead people on the voter list exposes the system to fraud. Someone can show up at a Rhode Island polling place, where identification is not checked, pose as a dead person, and vote.

It is too soon to tell whether any of the 4,991 dead people were recorded as voting in Tuesday’s election. That information will be available in several weeks when local boards of canvassers issue a report to the secretary of state about who voted.

Johnston, North Providence and Smithfield led the state in the proportion of dead people on the voting rolls. Johnston had 20.34 dead people per 1,000 registered voters; North Providence had 17.03, and Smithfield had 14.85 — all more than double the statewide average of 7.28.

In Johnston, the chairwoman of the Board of Canvassers said the town is powerless to remove dead voters it may find on the list.

“We have no legal recourse to remove them,” said Chairwoman Elizabeth D’Agostino. She said that the state Department of Health periodically advises the town of people who recently died and that the town removes those people. But, she said, no one else can be removed from the list.

D’Agostino said Johnston may be at the top of the list because it has a large number of nursing homes.

In North Providence, Lisabeth Marwell-Bussick, director of the Board of Canvassers, said the town follows a similar procedure with notifications from the Department of Health. Also, voters who miss two consecutive federal elections are classified as inactive. Inactive voters who miss two more consecutive federal elections are then removed.

Marwell-Bussick said she was at a loss to explain why North Providence had so many more dead people on its voting list than the state average. She said she would review death records the town has received over the years from the Board of Health to see whether any voters were overlooked.

The Journal found people on the voting list who died as recently as May and as long ago as 1988.

The party breakdown for dead people registered to vote was: Democrats, 2,394, Republicans, 391, and unaffiliated, 2,206.

In comparing the two lists, The Journal considered it a match if a voter’s first name, last name and date of birth were exactly the same as on the Social Security list and if the middle initial either matched or was missing from one or the other list.Deaths of people enrolled in the Social Security system are supposed to be reported to the federal government, which since the 1930s has been maintaining the Death Master File. Since the 1960s, newly reported deaths have been stored in a computer database. Because not all deaths are reported to the Social Security Administration, not all people who die are listed in the master file. The administration also cautions that, because of errors, some of the people listed in the master file may, in fact, still be alive. The only reliable way to confirm the information in the master file would be to examine death certificates, which are not public records in Rhode Island.

The Journal obtained its copy of the Death Master File from the National Institute for Computer-Assisted Reporting, which bought a copy from the federal government in July. The paper got the central voter registration list from the secretary of state’s office on Oct. 16.