Perhaps he tugs at his hair, perhaps he kicks out at a soap box. Either way, as one of those who hoped that a blooming economic spring might bounce the Conservatives back up the opinion polls, he will be feeling glum and confused this weekend.
Conventional wisdom and historical evidence agree: the more positive voters feel about the state of the economy and their own personal finances, the more amenable they are to the government of the day.
The fact that for the past five years the economy has been growing without any noticeable effect on Conservative popularity didn't stop Tory politicians hoping. Time and again we were told that the problem was simply that voters hadn't realised things were looking up. Once the housing market shifts, once wages start rising, once job growth picks up, once income tax cuts feed through: then, they said, voters will return to the Conservative fold once more.
If you believe the economic commentators, all those elements are now in place - superficially at least. The nation found out this week that dole queues shortened dramatically in February. More important, we discovered that average earnings were picking up fast. House prices are bubbling upwards too. Yet there is little sign of any significant movement in the polls. It seems that in this election the economy is not driving public opinion after all.
Big deal, some will shrug. For many, the idea that economies won and lost elections was always too simplistic.
Maybe so. But there is something curious happening here. Cast a glance back through history and there is an identifiable parallel between economic confidence and support for the government of the day. Splash across the Atlantic to the US and the link is even clearer. So when the odd unusual election violates the trend, it is worth looking further for explanation.
In Britain and the US, only two elections seem to violate the thesis that economic confidence and political support go hand-in-hand: the 1997 election in Britain, and the 1992 election in the US.
In Britain consumer confidence and government popularity moved hand-in-hand, up and down, throughout the Eighties. The graph opposite shows how the Conservatives' lead over Labour fits neatly with Peter Kellner's feel-good index from 1987 to 1993.
Across the Atlantic, the correlation between elections and economies is even clearer. An academic, Ray Fair, tried - slightly tongue in cheek - to trace a relationship between the share of the vote won in presidential elections by the candidate for the incumbent party, and the state of the economy at the time. Examining presidential elections running back to the beginning of the century, he found that the level of growth and the level of inflation had a measurable impact on the percentage of the vote won by the party of the governing candidate.
His model proved a pretty accurate predictor of the 1980, 1984 and 1988 US presidential elections. Even in 1996, Bill Clinton won - exactly as Professor Fair predicted last summer - with 49 per cent of the vote. But in 1992, it fell apart. Just as in Britain today, the popularity of the government and the state of the economy, after a long (if turbulent) marriage, seemed finally to have divorced. What went wrong for George Bush in 1992 might provide a clue to the Government's persistent unpopularity here in the UK in 1997.
Prof Fair re-examined his model after 1992 to see if weariness explained the discrepancy. After all, since the Republicans had been running the Oval Office for nigh on 12 years, the desire for change could be stronger than the good feely-factor.
Bingo. The new results proved that yes, the longer the party was in power, the harder it became to win again. It doesn't take a leap of imagination to reach the same conclusion here: 18 years of the same faces is simply too long for many voters to stomach.
But there is another important factor that pops through the gaps in Prof Fair's research. Even taking into account the period in office, the 1992 US election doesn't quite fit the historical trend. So what else was different about 1992? One thing emerges above all others: trust. Read my lips, said Mr Bush, no new taxes. But that isn't what he delivered after getting elected in 1988. By telling porkies, or at least by breaking promises, George Bush violated the trust of the electorate. No wonder they wouldn't reward him for the state of the economy if they didn't feel they could trust the economic statements he made.
The same factor helps explain the uncoupling of the economy and support for the Government in Britain. The facts of the economy matter less when you don't trust the statements about them made by the Government. The strong statements made in 1992 first about not raising tax, and then about supporting the pound in the ERM, all crumbled spectacularly. Likewise the persistent weaknesses in the economy during the recession made a mockery of all those wild claims about economic miracles during the Eighties. Hardly surprising then that the link between government support and economic feel-good has not been patched together.
If the polls are proved right in six weeks' time, and if the explanation suggested by the US evidence is accurate, the lesson for politicians is straightforward. Voters are not mugs. They may not like it when the economy doesn't match up to their hopes. But they hate being lied to even more. Governments who want to be re-elected can't just hope for economic good news, they have to earn voters' trust on economic policy as well.