When you think of policies that are going to address inequality of wealth, you have to be very thoughtful about what economists call "incidence of taxes." If most of the savings is being done by capitalists, and you tax the return on capital, then they will have less to invest. That would mean, over the long run, that the rate of interest would go up. That would therefore undo some of the intent to lower the income of capitalists.

If you get a flow of credit increasing, as we've seen in the last few years - that flow of credit didn't go to more wealth accumulation as we normally use the term in economics, as capital goods. What you got is an increase in bubbles of one kind or another.

The issue is: $1 trillion or $2 trillion is a lot of money. If our objective is to have stability in the Middle East, secure oil, or extend democracy, you can do a lot of democracy buying for this sum. To put it in context: The whole world spends $50 billion a year on foreign aid.

They [free market policies] were never based on solid empirical and theoretical foundations, and even as many of these policies were being pushed, academic economists were explaining the limitations of markets for instance, whenever information is imperfect, which is to say always.

America has had to turn to foreigners to finance its debt - not surprising since household saving in the last years has plummeted to zero. China is one of the largest holders of American debt.

I've always been sceptical about the notion that the market is a person you can engage in an argument with, and that that person is an intelligent, rational, well-intentioned person: it is fantasy. We know that ... the market is subject to irrational optimism and pessimism, and is vindictive ... You're dealing with a crazy man ... Having got what he wants he will still kill you.

There was a hope then by some people that what we call trickle-down economics would work. That if you made the economy pie bigger, everybody would benefit. Twenty-five years after NAFTA, we know that that is not true. We should have known then that it was not true.