More than three in five decisions to refuse family visas to those wanting to come to the UK from New York were overturned by managers before going to appeal, inspectors have said.
John Vine, the chief inspector of the UK Border Agency (UKBA), said he was concerned about inconsistencies, the lack of transparency, and the number of decisions overturned before an appeal in the office which deals with applications from the Americas.
Decisions over bids to settle in the UK also raised concerns, with more than a third of refusals being overturned by managers without going to an appeal.
In one case, the decision to refuse a woman permission to join a partner in the UK, who she met on the internet but with whom she could show no contact or communication, was overturned. She was allowed to settle in the UK, despite inspectors saying similar applicants had been refused and describing the case's credibility as "very weak".
Inspectors said the UKBA "chose not to address the chief inspector's specific concerns", responding simply that the entry clearance manager "was satisfied that the customer had addressed the reasons for refusal with her appeal". The UKBA also chose not to take any further action.
Mr Vine said he had "concerns about the lack of transparency in the UK Border Agency's inconsistent approach to evidential requirements".
"It is imperative that the agency ensures that its staff adopts a 'right first time' approach and makes correct and robust decisions," he said. "I found a high percentage of cases where the original decision was overturned before cases were allowed to proceed to appeal. This concerned me, as the agency had not conducted any analysis to establish whether or not the original decisions had been correctly made."
The problems were found in a review of the New York visa section's work for the 2010/11 financial year up to the end of February 2011. Refusals in 63% of family visit visa cases and 35% of settlement visa cases were overturned by a manager, inspectors found.
In their report, they said: "This prevented a significant number of cases from going to appeal as the original decisions were overturned.
"We were somewhat concerned that the percentage of cases where the original decision was overturned was so high and we found that no analysis had been carried out to determine any reasons for this, or if the original decisions made were of sufficiently high quality."