TSR Economics society

The ridiculous hours and a sense of morality?

Just to add to the debate but from the perspective of Nottingham...

Sorry - can't help you here, haven't read it.

Im thinking of learning an Economic model or two just out of interest. Could you suggest any, which aren't too difficult. For example, I've looked into the Solow Growth Model.

Solow and Solow-Swan type growth models are good. For something similar maybe look into the basic 3-equation New Keynesian macro models.

This sounds like my kind of thread - can't believe I have not seen it before!

Thanks very much, do you know any totally unrelated to the ones you just mentioned?

Have you ever looked at Bayesian Model Averaging? I've taken a look at it as it was mentioned in the article about China's rapid growth, but it didnt tell you how to actually use it. So I googled it, and it looked incredibly difficult. Having said that I am Lower Sixth.

Hi guys, I'm taking the IB next year.
I wanted to do SL Geography but due to timetable issues it is likely I will have to do SL Economics. I'd rate Geography 8/10 and Economics 6/10 in terms of personal interest at this moment in time. I don't know why I find Geography more interesting, I've never had an Economics lesson ever. Can anyone recommend Economics books for people with very little knowledge of Economics? I'd just like to see if I can become more interested, thanks.

I have a little question and if someone could help me out I'd be very grateful. I suck at econometrics so just need some clarification really on what I should do, if anyone is half decent at it they could probably help me.

I'm about to run a regression which would (in a simplified version) have "average changes in bond rates after a credit rating/outlook change" as the dependent variable and GDP as an independent variable. The problem I have come across is whether to have my averages yearly or just an average over the entire time period that I'm looking at (07-11). If I was to do it yearly my regression would be more accurate but due to the lack of change for the stronger countries with respect to credit ratings most of their average changes in bond rates would be 0. Because for example, between the period I'm looking at Germany only have their outlook changed once in 2011, so apart from that year the average for the other years would be 0. Would having quite a few zero's ruin my regression?

I just wanted your opinion on whether I should have my averages yearly or over the entire period? I don't want to run all my regressions and then find out it's inaccurate/wrong due to the sheer amount of zero's I'll end up with in my data!

It's been a long time since I did econometrics but I think I would start with the one where you're averaging over the whole time period. Then you would have a fairly straightforward regression with the number of observations equal to the number of countries. I'm not entirely sure what you mean by "doing it yearly" - would this be a panel regression with number of observations equal to number of countries times number of years? If so you would probably want to omit the years for which there were no credit outlook changes and run an unbalanced panel regression.

Cheers for the reply, I was planning on doing it over the entire time period anyway and just needed some clarification as my dissy tutor is **** at replying to e-mails. And I'd prefer to keep it simple.

By yearly I meant something like...

y
Avg. change in 07 (GR)
Avg. change in 08 (GR)
Avg. change in 07 (ENG)
Avg. change in 08 (ENG)
etc

x
GDP of Greece in 2007
GDP of Greece in 2008
GDP of England in 2007
GDP of England in 2008
etc

Some of the average changes for England would have been 0 and if I did it this way I was planning on just leaving the zero's in there so I'd have an equal number of dependent and independent variables. I just didn't know if the 0's would ruin my regression.

I would say that the zeroes would ruin your regression because it would drastically reduce the variation in your dependent variable. If I was doing this I probably wouldn't use years at all, but rather have each observation made up of variables like y = bond rate change after outlook change and x = GDP of country at that time, which would give number of observations equal to number of outlook changes. You could factor in time if you wanted and do a panel regression (like previously mentioned) but that would be a bit more complicated.

x

Sorry one last question. Youve been really helpful.

For GDP per capita I'm getting different values from these two sources (http://data.worldbank.org/indicator/NY.GDP.PCAP.CD AND http://www.tradingeconomics.com/greece/gdp-per-capita) I'm guessing it's because one's in current prices and one's at constant prices since 2000. Which one do you reckon I should use?