英语论文网

ary variables. You told me my model shouldn’t contain both types of variable –what do I do?
Warwick Business School 8
‘Traditional’econometric methodology
The starting point here is based solelyon the underlying economic/finance THEORY.
From this theory a THEORETICAL MODEL is derived.
An ECONOMETRIC MODEL is constructed from this by
Proxyingtheoretical variables with actual/available data.
Attaching a random error term to the model to capture measurement error and the effects of excluded variables.
The econometric model is then ESTIMATED using an appropriate estimator (e.g., OLS if the assumed error terms are independent,homoscedasticand uncorrelated with the included variables).
The model is subjected to rigorous TESTING:
The error terms are tested to ensure they satisfy the assumptions (misspecification testing). If they don’t the model is re-specified until they do.
Once this is achieved the parameter restrictions implied by the theoretical model are tested.
When satisfied that the theory is correct on the basis of the tests the model is used for its intended purpose e.g., POLICY EVALUATION and/or FORECASTING.
Warwick Business School 9
Example: Purchasing Power Parity (PPP)
(THEORY) A fundamental relationship in international finance is PPP. This states that the prices of identical bundles of goods (converted to a common currency) is the same across countries (no arbitrage condition):
THEORETICAL MODEL:
ECONOMETRIC MODEL:
**lnlnlnPPSSPP−=⇒=()()() . 0, , 0,error term NOISE WHITEa is ~ln~ln~ln22*210stuuEuEuEuuPPSstttttttt≠===⇒+++=σβββ
Domestic priceForeign price
Proxy variables (time-series data)
Nominal exchange rate
Warwick Business School 10
Example: PPPESTIMATION: The statistical assumptions on the error term are sufficient for OLS to be a valid estimator (i.e., unbiased and consistent).TESTING: After several attempts (possibly adding in lagged variables and/or additional variables) a model which satisfies the assumptions is derived.An F-test could be used on the final model to test the PPP hypothesis:Assuming success at the testing stage we can 代写留学生论文/留学生论文代写proceed to policy analysis/forecasting with the estimated model.()()[]0210under 3,2~321:HTFTURSSURSSRRSSFH−−−==−=ββRestricted RSS obtained from regressingon a constant*~ln~ln~lntttPPS+−RSS from the unrestricted regression（详情见附件）

本论文由英语论文网提供整理，提供论文代写，英语论文代写，代写论文，代写英语论文，代写留学生论文，代写英文论文，留学生论文代写相关核心关键词搜索。