Melanie Phillips, among others, says:

The single most crucial factor behind all this mayhem'is the willed removal of the most important thing that socialises children and turns them from feral savages into civilised citizens: a fully committed, hands-on, there-every-day father.

Unlike many on the left, I don't regard this as self-evidently wrong - it's a hypothesis like any other. So, what is wrong with it?
Clearly, it's incomplete in itself, just as attempts to blame the riots on poverty are; fatherlessness, like poverty, has existed for years and yet the riots are only occurring now. This account must, then, be accompanied by a story of imitative behaviour or cascades.
Even with this accompaniment, though, I have my doubts.
The first issue it raises is: are fathers really a source of moral behaviour in children? Granted, a good dad can be a good role model. But equally, a bad one can be a bad role model and so it might be better if he leaves; if I'd regarded my dad as a role model, I'd have become a gangster. Net, it's not obvious why having a father around should be a good thing. And three things speak against them being so.
First, Daniel Coyle points out in The Talent Code that losing a parent can have the opposite effect from turning a child feral. It can instead act as a 'primal cue' that unleashes the energy that propels people to greatness. Michelangelo, Washington, Lincoln, Darwin and Newton were all products of lone parent families and they didn't turn out too bad.
Secondly, Judith Rich Harris has argued that, for most families, parenting matters less than a child's peer groups.
Thirdly, empirical research by Ian Walker has found that whilst having a lone parent is correlated with poor academic achievement (and by implication the correlates of that such as poor behaviour), this correlation vanishes when we control for the fact that lone parents are poorer. Income, then, matters more than fathers.
The second issue is: are fatherless youth really immoral, or is it instead the case that their moral code is limited? I suspect - I'd welcome evidence - that many of these rioters don't lack a moral code. It's just that their code extends only as far as their mates or 'gang' and not to neighbours. But as I've said, morality is endogenous; it arises from social processes. Why is it that social processes work to make people loyal to a few friends, but not to others? The Left, of course, can answer this; poor youth feel no obligations to society because society has done nothing to benefit them.
But there's a third issue. Let's assume that all the above is false, and that fatherlessness causes people to be amoral. Why does this amorality lead to rioting?
The answer is that, to paraphrase David Friedman, people riot for the same reason that I write about economics - because the benefits outweigh the costs. And one reason why the costs of rioting are low is that such people have little to lose.
Which brings me to a curiosity. In principle, Ms Phillips' theory is consistent with a Leftist answer to the riots - improve people's economic situation and they'd have more to lose by rioting and so they'd not do it. Strangely, though, no-one - neither Ms Phillips nor her critics - seems to see this link. I'm tempted to think that, maybe, talk about the riots is motivated by something other than a desire to understand or prevent them.