On one thing all the papers, whatever their political loyalties, seem agreed: the Kelly affair has damaged public trust in Tony Blair. A YouGov poll in the Daily Telegraph on 25 July found Blair to be less trusted than either Iain Duncan Smith or Charles Kennedy. An ICM poll in the Guardian on 19 August reported that only 6 per cent of voters trusted the government more than the BBC to tell the truth.
Conventional wisdom states that the voters' trust is essential in a democracy. Trust is the new economic competence. In the broadsheets, it is now mentioned twice as often - and, in some papers, three times as often - as it was four years ago. And it is better to be trusted than not trusted; we can agree on that. Nobody likes not to be believed.
But before he telephones for the removal men, Blair should ask himself two questions. Is he really losing voters' trust? And would it matter if he was?
On the first question, the polls are not quite as clear-cut as they might first seem. The two polls quoted in the first paragraph were taken a month apart. Each detected a fall in trust, yet the level of trust in the second poll was pretty much the same as in the first.
From what level of trust is Blair supposed to be falling? After all, he became Labour Party leader in 1994, and very quickly established a reputation as a ruthless and effective manager of the media. After the experiences of Neil Kinnock and John Major with the British press, many people sympathised with his approach.
Nevertheless, much of the argument of the 1997 election campaign focused on what Blair, once in office, would actually do. Would he be radical? Would he be socialist? Would he be a de facto Tory? Many inside the Labour Party distrusted him as much as those outside did.
Blair surfed to his first great landslide in 1997 on a wave of goodwill - but not trust. It was a mere two days into his premiership when the Economist accused him of "calculated dishonesty".
New Labour has always had the reputation of being more concerned with appearances than actuality. There is nothing new in what the Hutton inquiry has exposed about the Prime Minister's inner circle and its obsession with artful selection and presentation of information.
Yet Blair has always been written about as a man whose success is explained by the public's trust, and whose demise will be preceded by its loss. Such reports are no doubt influenced by his "hey, I'm a straight kind of guy" presentation of himself. But that does not make them accurate. Critics foresaw the toppling of Blair's government when a speech to the Women's Institute in June 2000 went badly wrong and he was heckled and given a slow handclap. This disaster generated column-inches galore - but is totally forgotten today.
In short, if Blair looks back through his scrapbook, he will find that he cannot be losing voters' trust, because he never had much of it to start with. That may not comfort him. But it brings us to the second question. Does the lack of trust matter?
A few weeks ago, Jeremy Vine's BBC Radio 2 programme ran a poll to find Britain's most honest politician. The winner was William Hague. Yet honest Hague was trounced by slippery Blair in 2001.
Hague's predecessor, John Major, who by many accounts was infuriated by Blair's false piety and naked opportunism, was generally agreed to be one of the nicest and most straightforward occupants of 10 Downing Street in living memory (at least before the revelations of his affair with Edwina Currie). Yet nice Major was walloped by slithery Blair in 1997.
Not that Major didn't have his own problems. In the New Statesman in 1991, the then political editor, Sarah Baxter, suggested that Major was "squandering his most precious asset - trust". Six months later, Major beat Kinnock.
In fact, if Blair looks back even further, he will find that trust seems to have very little relevance to winning elections. The table (above) shows election results since the Second World War, along with very rough summations of how the public perceived the two main party leaders. (I emphasise that these are presented as popular perceptions, not as my own assessments of the leaders concerned.) Elections are about much more than the party leaders, but a presidential focus is illuminating, given that the whole trust debate assumes that Blair's personal ratings are vital to the election result in 2005 or 2006.
Blair may be relieved to discover that on only four occasions out of 16 did the more obviously trusted candidate win the general election: Clement Attlee in 1945 and 1950, Edward Heath in 1970, and Major in 1992. In contrast, Harold Wilson, perhaps the shiftiest prime minister ever, won four elections against markedly more trusted candidates.
In other words, trust is not a big vote-winner. And that is the experience of many other mature democracies. In the US, the likes of Lyndon Johnson, Bill Clinton and Richard Nixon prospered, despite the lack of public trust. Nixon, known as "Tricky Dicky", managed to secure re-election after saddling himself with a vice-president, Spiro Agnew, who was even less trusted than he was.
In France in 2002, shady Jacques Chirac did not impress - but honest, plodding Lionel Jospin failed even to make it to the starting line. Thus has a French tradition, inaugurated by that connoisseur of African diamonds Valery Giscard d'Estaing and nurtured by Francois Mitterrand, been continued. If Blair wants further reassurance, all he need do is look at his new Italian friend Silvio Berlusconi, who is so distrusted that he is off the scale.
I am not suggesting that Blair can relax; he may still face trouble from rebellious backbenchers or from the voters, or even from his own conscience. But what is clear is that trust is much more complex than the current debate suggests, and that the links between trust and political success are intricate and hard to understand. The voters appear to accept that, to get things done or to keep coalitions together, it is necessary sometimes to dissemble and to conceal.
Trust is therefore not best represented by yes/no questions in opinion polls, as Onora O'Neill observed in her 2002 Reith lectures.
Trust has multiple aspects. Most obviously, you trust someone not in the abstract, but in a context. You trust them to do certain things. People trusted Michael Foot in 1983 to stick to his principles; it was just that they didn't want those principles applied in Britain.
Some people can be trusted on one level, but not on others. Margaret Thatcher was trusted to do big political things - smash the unions, retake the Falklands, impose the poll tax. But in her individual relations, she was not very trustworthy. Think of how she behaved over the Westland affair, or how she cheerfully broke "agreements" with Geoffrey Howe and Nigel Lawson at the first convenient moment. Major had the opposite problem: he was trusted as an individual, but was deemed ineffectual in the greater scheme of things.
Besides, the pollsters' binary notion of trust fluctuates according to quite bizarre stimuli. Why did Blair become more trusted when baby Leo was born? Why did George Bush become more trusted after 11 September 2001?
Because trust is such a complex thing, it cannot be generated artificially. Going around saying "I want to be trusted" would be about as effective as saying "I want to be loved". And perhaps that is just as well: if there was a particular behaviour pattern that would generate trust, then politicians (and others), Iago-like, would simply adopt it, severing the essential link between being trusted and being trustworthy. Sincerity is everything - if you can fake that, you've got it made.
The appeal of political leaders depends on a variety of qualities; apparent trustworthiness may be one, but it is not among Blair's. Neither does it seem to have been a quality of any conspicuously successful party leader since Attlee. Blair has benefited at various times from a ham-fisted opposition, from a natural aura of authority, from the appearance of managerial competence, and from a demonstrable commitment to a mild egalitarianism. These are the things he should be worried about losing.
Kieron O'Hara is a senior research fellow at the University of Southampton. His Trust: from Aristotle to Enron is published by Icon Books in February 2004