promotion image of download ymail app
Promoted
紫龍 asked in 社會與文化語言 · 1 decade ago

英文翻譯(請勿用翻譯軟體)

This can be justified

theoretically by realizing that forecasts from a

structural model are just a function of current and

past data. If the function can be estimated consistently

then the resulting forecasts will have the same forecast

error variance, asymptotically, as if the function were

known (Stock, 2001).

In an attempt to illustrate the success of their

endeavor, econometric forecasters often perform static

simulations, in which the historical record is simulated

by the model, using actual values of exogenous and

lagged endogenous variables. Recognizing that the

actual values of at least some of these variables will be

unknown for the forecast horizon in question, it is

claimed that the exercise merely demonstrates the

performance of the model system, which is neither

necessary nor sufficient to make useful forecasts.

Arguably better is dynamic simulation, where actual

exogenous variables and model-generated lagged

endogenous variables are used as inputs, though this

too has been shown to be an invalid model evaluation

technique (Hendry & Richard, 1982). To properly

measure the forecast performance of an econometric

model, first, forecasts should be for outside the period

used for estimation and, second, all actual variables

that would be unknown at the time the forecast is

being made should be replaced by their forecasts.

In a review of forecasting, we have less interest in

the last two activities of econometricians, inference

and policy analysis. Structural models are often used

to make forecasts as well as structural inferences.

These are cause and effect relationships measured

over time. Historically, short-term and long-term

elasticities have been used. For demonstration they

rely on showing the effect of a change in a causal

variable or an impulse shock to the system that may or

may not happen. Therefore, there is no check on the

accuracy of the inference beyond tests of model

adequacy (specification and misspecification tests).

1 Answer

Rating
  • 1 decade ago
    Favorite Answer

    這可以合理

    理論上實現的預測,從1

    結構模型只是功能的當前和

    以往的數據。如果功能可以預計,始終

    然後由此產生的預測都會有相同的預測

    誤差方差,漸近,如果功能

    眾所周知, (股票, 2001年) 。

    在試圖說明他們成功

    奮鬥,計量經濟預測往往執行靜態

    模擬,在其中的歷史紀錄是模擬

    由模型,用實際值外源性和

    滯後內的變數。認識到

    實際的價值至少有一些這些變數會

    為未知的預測,在地平線上的問題,這是

    聲稱,這次演習只是表明了

    表現了該模型的制度,這是既不

    必要,也足以使有用的預測。

    可以說是更好的動態模擬,實際

    外生變量和模型產生的滯後

    內生變量是用來作為投入,雖然這

    太已被證明是一個無效的模型評價

    技術(亨德利&理查德, 1982年) 。妥善

    衡量預測的表現,計量經濟

    模型,首先,預測應為境外期間

    用於估計,第二,所有的實際變量

    這將是未知的在時間的預測是

    正在取得應改為他們的預測。

    在檢討預測,我們有少的興趣

    過去兩年的活動econometricians ,推理

    和政策分析。結構模型是常用的

    作出的預測,以及結構的推論。

    這些都是原因和後果的關係來衡量

    隨著時間的推移。從歷史上看,短期和長期

    彈性已被使用。示範他們

    依賴於顯示效果的改變,因果

    變量或衝動休克該系統可能或

    可能不會發生。因此,沒有檢查就

    精確的推理超出測試模型

    充足(規格和misspecification測試) 。

    • Commenter avatarLogin to reply the answers
Still have questions? Get your answers by asking now.