Well, really, theoretical ones are fine, but if you can use current statistics/factors that's even better. It's probably not worth worrying about for two mark questions though, it just gives whatever you're arguing some evidence which (to me, at least) makes answering the question a bit easier.
Often on exams they will ask for current knowledge, policy mix questions, for example, tend to do this. They also have questions like 'Over the past decade...' or 'In the past 3 years...' which would mean that in order to get full marks, you need to actually have some examples of what happened in that time. Usually those types of questions are asked about budgetary or microeconomic policy, but could be about anything on the course.
Remember though, the exam was written in Feb this year, so while you are expected to know some stuff that happened this year, it won't ask for things that happened in the last couple months specifically. It's more likely to ask about general trends or our current economic situation. For example, they asked about the drought in 2003, oil prices last year, and there were questions as well on the participation rate. While you could answer these questions purely on a theoretical basis, it would be really quite difficult to quantify your claims or demonstrate sound analytical skills if you haven't read a little bit about these topics.
You definitely need to do the last few exams, since 2004 at least, to get an idea of what's expected.