Time Series Bullshit

Question 1

What does the 4QCMA represent

Answer 1

It contains only trend and cyclical components. The totaling process remove the seasonal effect and the averaging process removes the random component

Question 2

What does ratio to MA represent

Answer 2

Seasonal and random components

These are not pure seasonal indices because they are affected by the random component. So they are called specific seasonal relatives

Question 3

How to isolate the seasonal component from the specific seasonal relatives

Answer 3

1-arrange the specific seasonal relative according to their specific seasons

2- calculate the median for each season and then sum them together

3- determine the adjustment factor = (number of seasons *100)/sum of medians

4- obtain the seasonal indexes by multiplying each medians by the adjustment factor

Question 4

Deseasonalized

Answer 4

The seasonally adjusted data contains the time series components TxCxI

Question 5

How to isolate the cyclical and random component

Answer 5

1- we divide Y/TxS

2- we take a 3-période moving average of the CxI component which helps to eliminate the randomness so that the 3ma series is in effect the C component

3- we divide CxI by the 3ma to isolate the random component I. It should be white noise (truly random)

Question 6

What is the percent of trend

Answer 6

Y/T= CxSxI