Time Series Analysis 30E00800

GD Star Rating

Was the course worth attending? Were the lectures useful? What questions did the exam have? Give us your honest opinion on the course!

One Reply to “Time Series Analysis 30E00800”

  1. I don’t understand how we can allow people to graduate from this school without doing this or an equivalent course (like econometrics for finance, not that basic econometrics course) . It goes to the very core of how to analyze financial and economic data. This is the bare minimum you should be able to expect anyone with a degree in economics to know.

    It’s not exactly the easiest course in this school, half of the students drop out. The course book is essential. Get it and scan it for yourself if you must (FYI it is legal to do this for your own purposes; it is illegal to share the scanned file), but I can’t imagine how you can survive this course without it. The lecture slides are quite useless as they mainly contain formulas and equations without very many explanations.

    The management of the course was a bit lacking. E.g. sometimes there were mistakes in the exercises and the corrected versions were posted quite late without the DL being moved (and sometimes you wouldn’t even notice they have changed). The last exercise round example answers were published a few days before the exam, and yes, the exam had hard questions from that set (see the last question in the exam). Again, not properly even announced, I was lucky enough to notice from Noppa a few days beforehand that they were there.

    There is a list of equations that you get in the exam. Don’t think that these are the important things that you should study hard for the exam, quite the opposite. The formulas you are given are the irrelevant ones you don’t need to memorize for the exam (you actually only need one of those formulas). YOU NEED TO MEMORIZE ALL THE FORMULAS THAT ARE NOT IN THE LIST OF FORMULAS, THOSE ARE THE IMPORTANT ONES THAT YOU NEED IN THE EXAM. The list of formulas you are given is nothing but a cruel joke.

    The exam has 5 questions, something like this:

    Which of the following are weakly stationary, weakly non-stationary or possibly both. Wrong answers will be scored -1. Then you get 20 equations or descriptions of equations for which you select stationary, non-stationary, both, no answer:
    – The sum of I(0)+I(1)
    – Random walk
    – Random walk with drift
    – (1-L)^2*Y = 1
    – ARCH (p), p>1
    – Trend stationary process
    – GARCH(1,1)
    – Y(t) = Y(t-1) + 2 + u(t)
    – (1-L)^2*Y = 1 + white noise
    – etc.

    Explain briefly (~15 concepts):
    – Wald decomposition
    – Strong stationarity
    – Cointegration (1,1)
    – Integration of order d
    – etc.

    Show that the following AR(2) model is stationary. Then you are given some AR(2) model and you start doing the math, but the trick is that if you blindly follow the math you are taught, you end up with an imaginary number. So after doing a lot of math you notice you need a completely different approach. Figure that one out hahaha. I don’t know if I answered this one correctly, but I decided to just investigate the decaying acf and concluded from that that the model is stationary since the analytic answer gave a complex (imaginary) root.

    4: [This is the only question where you need one of the equations in the given list of equations]

    A researcher has the following AR(1) model [some AR(1) model given]. Test for unit root.

    Another researcher has the following AR(2) model [some AR(2) model given]. Explain how unit root would be tested. You do not actually need to do the calculations but what would you conclude? Which model would you choose and why (4a or 4b)? The t-statistics are given below the coefficients.

    The model for the volatility of a stock is the following:
    σ(t)^2 = α(0) + α(1)*u(t-1)^2 + βσ(t-1)^2

    α(0) = 0.00005 α(1) = 0.05 β = 0.9

    What is the model?

    The following values were estimated for today (time t):
    Return on portfolio: -2%
    Volatility for today: 5%

    What is the estimated volatility (standard deviation) for tomorrow?
    What is the long-term average volatility?
    What is the average daily volatility for the next 10 days (t+10)?


Sähköpostiosoitettasi ei julkaista. Pakolliset kentät on merkitty *

This site uses Akismet to reduce spam. Learn how your comment data is processed.