It's not compatible to expect multilateralism to work and, at the same time, to expect to walk out with everything you wanted. This is a recipe for failure. If we prize the system, we have to come knowing that we will need to make compromises. Sometimes painful compromises.

Making strong infrastructural reforms, particularly in the area of social security, that's not an appealing prospect for any country or for any political structure. But that's a reality.

I believe that trade is essential for economic growth and development around the world, but I also believe that trade is imperfect.

Just having the economy picking up doesn't mean that we will get through or that we will turn the chapter on many of the problems that we see ahead of us.

Trade is not the cause for unemployment. In fact, the biggest drivers for unemployment are innovation and increased productivity. It has nothing to do with trade.

In most negotiations, you can't attribute success or failure in negotiations to one side.

The best thing we can do to secure the future of the global system, trading system, is to redouble the efforts to improve the system, to reform the system.

One thinks that one is winning when we slap tariffs or introduce barriers to imports from another country, and we think we win. But you lose when you export because the other countries are going to raise tariffs as well. They're going to introduce barriers as well. So you win with one hand and you lose with the other.

New technologies, innovative management, higher productivity, displacements in the labour market, increased migration - these are all provoking major economic, social, and political shifts. These shifts need to be better understood if we are to address them in a positive and effective manner.

I think educational systems have to be more nimble, have to be more adapted to today's realities where students can go in different directions and professionalize even faster. Constant retraining and reskilling and upskilling, whatever you want to call it, of the workforce.