I had a discussion with colleague whether would it be sensible to average quarterly unemployment benefit rate to an annual rate? Is it better to refrain from averaging the quarterly unemployment-related data as the quarterly data allows for accounting for the seasonal variation in the unemployment benefit claimants rates.
Assuming that it is permissible to average the data quarterly, which option would be advisable for producing small area estimates:
- Obtaining totals for each quarter and then using appropriate reference population to obtains the rate: $$ rate = \frac{total~claimants}{reference~population} $$ where $total~claimants = \sum\limits_{i=1}^4 q_i=q_i+q_{i+1}+q_{...}+q_4$ and then calculating average using reference population.
- Simply averaging quarterly rates.
