Returns from the count showed 49.89 million people usually resident in England and Wales. The provisional mid-1991 population estimate published by the Office of Population Censuses and Surveys (OPCS) yesterday was 50.95 million.
Many of the missing million were traced by follow-up postal surveys and by interviewers after census night. But the latest figures show 'under-enumeration' of 965,000. In 1981 the census under- enumeration was 214,000.
The latest figures show 200,000 of the under-enumeration arose from missed or misclassified dwellings and 177,000 people missing from households which under-reported the number of occupants.
The 1991 report includes a new category of under-enumeration. For the first time the OPCS acknowledges that 572,000 people both failed to return the census form and also 'eluded' follow-up census interviewers.
The OPCS population report said there was 'an apparent shortfall of people aged 1 to 44, and 85- plus, particularly males aged 20 to 29, in the census who also eluded the census validation survey interviewers.' The number of males aged 20 to 29 who eluded interviewers was 220,000.
Politicians and academics have argued that many thousands of people evaded the census because they had failed to register for the poll tax and did not wish to be traced. They therefore lost the vote in the 1992 general election.
The OPCS said: 'It has been suggested people would avoid the census in the mistaken belief their names would be put on the poll tax register.
'We have been able to find no firm quantifiable evidence that this is the case. Some people have avoided or been missed by both the census and the census validation survey, in spite of the fact we have experienced interviewers. It is impossible to be be sure why it has occurred.'
However, Eileen Howes, census liaison officer for the London Research Centre, who advises London boroughs on public finance implications of the census, is convinced a significant number of the missing million is because of the poll tax.
'According to census enumerators we have spoken to, a significant amount of under-enumeration was because of the poll tax. The OPCS cannot prove it and they have no evidence because they didn't ask the question.'
She added: 'The most surprising figures are for those who the report says eluded the interviewers, especially men aged between 20 and 29. It is highly likely many of these and also more in other categories under-counted were due to the poll tax factor.'
The 1991 population figures have been recalculated to take account of the under-counting and show the population has increased by 1.32 million since the 1981 census. Two-thirds of the growth was due to the excess of births over deaths and the remainder to inward migration from the new Commonwealth and Europe.
One of the most significant changes was an increase of 507,000 or 6 per cent in the population of pensionable age - women of 60 and over and men 65 and over. The numbers of people aged 75 to 84, and 85 and over increased by 17 per cent and 50 per cent respectively.