There've been a gazillion U-turns by both the Conservatives and LibDems since they entered government, but none so massive than the LibDem pledge to abolish tuition fees, only to double them as soon as they got into power.
The response from the LibDems like Cable is to bleat that they "didn't expect to be in government" when they made the pledges - in other words, they felt they could safely trick voters with any number of unrealistic policies in order to get elected. Not only that, they seem to feel that the voters they tricked will nod sagely and decide that it was OK that the LibDems were pledging all sorts just to get elected, as long as they didn't expect to be in power! And that their pledges to their coalition partners are more important than pledges to the voters!
How much do broken pledges matter? Quite a bit, I think. One example is David Cameron's pledge to hold a referendum on the Lisbon treaty, and the subsequent U-turn in late 2009. The Conservative lead started to plummet as some of their voters simply abstained in disgust. They had been 20 points ahead in mid 2009, but when the general election results came in, they failed to win a majority.
Breaking pledges in government is even worse. Labour was very, very careful to stick to it's manifestos - the Labour government even delayed the passage of the tuition fees bill, so that fees were in the 2005 manifesto and voters could vote against if they didn't like it (and Labour had a mandate for them when they got elected). I recall only one pledge broken in 13 years of government and that was the pledge not to raise income tax - higher rate tax was raised to 50% with effect from April 2010, a month before the general election. And the only reason that took place before the election was because it was impractical to hold the general election on 2nd April and then the council elections in May. So one broken promise right at the end of 13 years in government.
What hurt John Major's government so badly was making a great fuss of how Labour would raise taxes (the so-called "tax bombshell") and accusing Labour of wanting to pull out of the ERM and then breaking both pledges just six months after the election, leaving voters feeling like they'd been tricked.
Major and the Tories also thought that as the broken promises happened right at the start of that parliament, the electorate would forgive them five years on when the came to vote in the next election. You hear a lot of LibDems and Tories saying something similar now. "The next election is not till 2015".
But not only did voters not forgive Major, the longer they were denied a say, the more furious they got. What had been shock in 1992 and 1993 had hardened into implacable fury by 1997, and when they finally got to vote the electorate blasted the Tories into their worst defeat since 1832. John Major would have been better off holding the election in late 1993/early 1994 before momentum against him had started to build.
LibDems and Tories dismissing voters annoyance by saying "people will have calmed down by 2015" are making the same mistake Major did.
General elections are great safety valves. Let voters have their say early and you take a minor hit. But if you blatantly fib to get elected, and then dismiss voters concerns with a patronising "they'll get over it", anger just builds and builds and builds. The longer people have to wait to have their say, the bigger the blow-out. Luckily for Labour the coalition are too stupid to grasp this. Incidentally, this is why fixed parliaments are a bad idea, and fixing the length of parliaments for five years is even worse. We should retain the right to have elections at any point (in practice this means four year terms and the ability to consult the voter in unusual circumstances).