The Second Superpower: Referendums

In the 2012 David Butler Lecture, "The Second Superpower", YouGov President Peter Kellner advocates Representative Democracy, dismissing Direct Democracy and referendums as "flawed devices, used when politicians lose their nerve". Of course, Interactive Democracy supplants YouGov and all the other pollsters, to which millions of Pounds of tax payers' money flows, and this may account for his opinion, but it's worth listening to his arguments nevertheless.
He identifies several problems with referendums:-

1. If the choice is not a binary "for or against" and there is more than two choices, then measuring the result becomes difficult. Interactive Democracy can overcome this in two ways: Parliament should refine all the possible options as they do today; and an alternative vote type system can be used to measure the response to multiple options.
2. Once a referendum decision is made it is hard to change without another referendum, thus mistakes are difficult to reverse. It seems that Peter Kellner only considers infrequent referendums, but ID proposes very frequent referendums with devolved votes to your MP if you don't participate. This makes laws just as easy to change as today. But there's another facility, too: the government could refuse to implement the majority result at the risk of triggering a general election. They may take this option if the situation has changed or they have another good reason.
3. Referendums tend to maintain the status quo as there is a natural tendency among many to conserve the familiar rather than risk adopting something new that may turn out to be worse. That assertion may or may not be true but it hasn't stopped Switzerland from being a successful and progressive society. In my opinion it all comes down to the quality of the debate, the conservatives (small c) pointing out the risks, the progressives highlighting the benefits. The ID system (indeed any good government) should also facilitate using evidence from other places, local studies or setting up experiments to prove a policies worth.
4. Referendums stop politicians from being accountable. In ID there are special provisions to identify the arguments that politicians make when persuading us of their views prior to a referendum, by highlighting their contributions on the debating part of the site. This makes them more directly visible to their electorate who will judge them on their performance. It is their responsibility to research and present their views. The government is also accountable for many things, such as international relations or the implementation of policy, that we can judge them on in the usual manner.
5. How the referendum questions are phrased can vastly effect the result. This is undoubtedly true, but in ID the question is phrased by Parliament, government and opposition together, under the critical gaze of an involved electorate, savvy to their Machinations, who can mobilise real political power to press them for an honourable process. (I consider the pressure ID puts on politicans good for keeping them honourable.)
6. Peter Kellner also highlights the problem of the expectations gap: the difference between what the electorate expect and politicians have the power to achieve. By engaging the electorate in the debate, encouraged by easy access and real empowerment, the debate its self educates the population - Interactive Democracy IS education!

After all of his arguments, Peter does offer an alternative vision to cope with the general population's low opinion of politics (identified in his polls). He suggests the use of Citizens' Assemblies where people debate a topic, something he says is used successfully in Iceland, which also frequently uses referendums. And that's my point: let's have Citizens' Assemblies, too; they can run along side Interactive Democracy and reinforce it by offering a face-to-face debate that some may prefer. Perhaps the political parties or the media would fund them. Yet Citizens' Assemblies are much more difficult to access for most people than the slow written on-line debate, points collated for and against, that ID provides. This can be accessed from home, a library or even while on the move, at a time that suits you, and it facilitates careful deliberation.
Peter also advocates a People's Veto. This would stop any law if 50% of voters vote against it. That's not empowerment, it's a sop to public opinion and doesn't offer any real chance of implementation. It effectively bans my referendum vote unless a majority of others make an effort to get involved.
What Peter misses from his criticism of conventional referendums is that a low turn out enables a minority of the electorate to carry the day. ID avoids this problem by automatically devolving your vote to your MP if you don't happen to vote, making sure you are always represented and providing a balance of power between MPs and Citizens depending on the degree of public engagement.
Please click here to watch the lecture. I'd appreciate your views.