通过日收益测算波动性 不知道除了rolling window, midas,garch-family,还有啥？希望大家给点儿提示

Mixed Data Sampling (MIDAS) regression is an estimation technique which allows for data sampled at different frequencies to be used in the same regression.

More specifically, the MIDAS methodology (Ghysels, Santa-Clara, and Valkanov, (2002) and Gyhsels, Santa-Clara, and Valkanaov (2006), and Andreou, Ghysels, and Kourtellos (2010)) addresses the situation where the dependent variable in the regression is sampled at a lower frequency than one or more of the regressors. The goal of the MIDAS approach is to incorporate the information in the higher frequency data into the lower frequency regression in a parsimonious, yet flexible fashion.

The following discussion describes EViews’ easy-to-use tools for single equation MIDAS regression estimation. We begin by offering background on the approach. Next, we describe how to estimate a MIDAS regression in EViews. We conclude with examples.

MIDAS Estimation in EViews

With built-in tools for working with multi-frequency data and an intrinsic understanding of the relationship between various time series frequencies, EViews offers an ideal platform for MIDAS estimation.

To perform MIDAS estimation in EViews, open the equation dialog by selecting Quick/Estimate Equation…, or by selecting Object/New Object…/Equation and then selecting MIDAS from the Method dropdown menu to bring up the MIDAS estimation dialog:

http://www.eviews.com/help/helpintro.html#page/content%2Fmidas-MIDAS_Estimation_in_EViews.html%23

SpecificationThe Specification tab is used to specify the variables of and form of the MIDAS equation and to set the estimation sample.The Specification edit field is used to specify the low frequency dependent variable followed by a list of low frequency regressors from the same page as the dependent variable. The low frequency regressors should include any desired lags of the dependent variable. Note that explicit ARMA terms are not permitted in this estimation method.

The Higher frequency regressors edit field is used to specify the higher-frequency regressors. The syntax for these variable is pagename\seriesname where pagename is the name of the page containing the series, and seriesname is the name of the series. Note also that series expressions are allowed, e.g. “mypage\log(x)”.

You may specify more than one higher-frequency series, and those series may be of different frequencies from different pages. However, we caution you that using more than one high frequency regressor oftens leads to multicollinearity issues and, in the case of the non-linear weighting, increases the complexity of estimation dramatically. An alternative approach suggested by Andreou, et al. (2013) would be to estimate several univariate models and then use forecast combination to produce a final forecast.

Date TimingWhen specifying your high frequency variable, care should must be taken to ensure that you refer to the correct observations from the higher frequency page.To illustrate, let’s assume our dependent variable, Y, is quarterly, and our regressor, X, is monthly. We would like to use 4 lags (months) of X to explain each quarter of Y. EViews will use the 4 months up to, and including, the last month of the corresponding quarter. Quarter 1 will thus be explained by March, February, January and December. Quarter 2 will be explained by June, May, April and March.If you wish to use different sets of months, you can use the lag operator when specifying the regressor. In our example, if we want Quarter 1 to be explained by January, December, November and October, and Quarter 2 to be explained by April, March, February and January, we would specify the regressor as “monthlypage\x(-2)”; i.e., using the second lagged values of X.

All of the MIDAS estimation methods require a value for
, the number of high frequency lags to be included in the low frequency regression equation.

All of the MIDAS estimation methods require a value for
, the number of high frequency lags to be included in the low frequency regression equation.Just below the Higher frequency regressors edit field are radio buttons that control the number of lags. You may provide a fixed number of lags by selecting the appropriate radio button and entering a value, or you can elect to determine the number of lags using minimal sum-of-squared residuals as the selection criterion. If you select the latter radio button, you will prompted to enter a value for the maximum number of lags. Note that automatic selection is only available for the Almon and Step weighting methods.