How to use time series analysis to analyse budget components such as National Debt, GDP and unemployment trends over years? Which parameters to be captured and which software can be used in this case?
In my experience, it is important to begin with a basic model that interconnects these various variables. For example, an open-economy loanable funds model is quite useful and versatile for this purpose.
The estimating equation that can be logically derived from the model can take a number of general forms, including: [with R being a nominal or perhaps ex post or ex ante real long term (10 years or more) interest rate yield]:
R = R(Deficit/GDP or Debt/GDP, Expected Inflation, the Unemployment rate OR change in per capita real GDP, money supply measure (such as M2)/GDP, the ex ante real short term interest rate [such as a Treasury bill rate, net capital inflows/GDP...)
The variables should be subjected to unit root tests and the empirical estimation can accordingly take the form of something a simple as two stage least squares or the co-integration tools as developed by Johansen--or both--perhaps as robustness tests. However, it is important to specify variables like the deficit or the debt and the money supply and net capital flows as a percentage of GDP before the unit roots tests. These variables should be scaled by GDP so they can be judged relative to the size of the economy. In any case, EVIEWS is very user-friendly for the estimations.
You can run Johansen Jesulius cointegration, Vecm, Granger causality by using any software namely eviews or stata, microfit. Firstly you must be clear with the theories for understanding their relationship and impacts on budget, then go for stationarity and the above mentioned test
- Badges
- Science topic
- Similar topics
- Correction