1. Intro To Time Series Analysis Flashcards
What is time series data?
It consists of observations on a given economic unit
Signal
The systematically prefictable component of the time series
Noise
The component whcuh makes time series difficult to predict
What does the signal consust of?
Trend, seasonal and cycle
What is the trend?
Thr persistent tendency of the times series to increase or decrease over time
What is the seasonality?
A type of cyclicality where the times series has a tendency to increase or decrease in a predictable or regular way
What is a lagged variable
If Xt: t=1,2,…,T then Xt-1: t=2,3,…,T is the value of the variable in period t-1 or lagged one period
What is a lead or forward variable?
The opposite of a lag variable
Differenced variable
The difference between the current value of the variable and its value in period t-1
What is significant of a a log first difference variable?
It gives the proportional change % in X between period t-1 and t, it is also known as the rate of return
Stationary variable
A variable that isnt explosive, nor trending, nor wandering aimlessly wuthout returning to its mean
What is true of the stationarity of most macro variables?
They are non staionary but theut first differences or returns tend to be stationary
When can standard regression analysis be carried out on variables?
When variables are stationary or when certain conditions are given
What is the autocorrelation function?
Autocorrelation is the correlation involving a variable and its past (lag)
What is the autocorrelation of order 0
The correlation between a variable and itself so it is 1
Correlogram
A sample ACF
Partial Autocorrelation function (PACF)
The correlation between two specified points conditional on all other autocorrelations that come in between
Cross correlogram
The correlations between two time series variables Y and X calculated at various lags
When are X and Y not cross correlated
When corr(Yt,Xt+k)=0 for all values of k
Dynamic relationship
A change in a given variable today has an impact on the same variable, or ither variables in future time periods
What are dristributed lagged effects
Effects which have an impact over many future periods
What are the ways of capturing a dynamic relationship?
- distributed lag (DL) model
- autoregressive model (AR)
- autoregressive distributed lags model (ARDL)
What is the distributed lag (DL) model?
A model where the dependent variable y is a function of current and past values of an explanatory variable x
What is an autoregressive (AR) model?
A model where the dependent variable y is a function of current and past values of y its own lags
What is an autoregressive distributed lags (ARDL) model?
A model where the dependent variable y is a function of current and past values of an explanatory variable x and its own lags y
What is the long run multiplier of a DL model?
Bo+B1+B2+…
What is necessary for an AR model to be stable?
Alpha1 must be between -1 and 1
