New projects are always a gamble.
You are going to put in effort in the hope of a more valuable return.
Private citizens have limited resources, so when a project doesn't workout in a reasonable timeframe they will cut their losses, abandon it and move on.
However, the public sector (making up about half the UK economy) have almost unlimited resources. If a project is not working out, they always have the option of doubling up - like a gambler betting on a coin toss, doubling his bet every time he loses so should he eventually win he will get back everything he has gambles so far, and the same again as profit.
The problem here is that projects are not coin tosses - there is no law of averages that says that a project will eventually come good.
So when a public sector project isn't working out, and it gets expanded (more money, more people, more scope) to double up the stake/(potential) return - it actually becomes less likely that any return will be had - but those behind the project will not want to lose face by admitting their initial mistake/failure so will keep doubling up, again and again until they find a plausible scape goat for a failure and then heap all the blame on them.
The risk of failure is what make businesses strive to succeed - in the public sector there is no such risk, there is always the taxpayers wallets to be raided.
This only goes completely off the rails when rank amateurs (the ilk of Gordon Brown) take it to a complete extreme and end up gambling the entire wealth of the country! At this point, only a war will divert attention from such in competence so we get dragged into foreign conflicts that are nothing to do with us - with our 'opponents' being the instantly created 'scapegoats' for the public sector failure.
'Good' is what is left when all 'bad' has been cut away - the public sector have no mechanism, reason or drive to cut away the bad. Quite the opposite, 'bad' has resources thrown at it in the futile hope that it can be made good - it can't it is precisely what stops 'good' emerging.