MVIDENT Procedure

@MVIdent creates a Tiao-Box cross correlation matrix with +,- and . symbols for significant positive, negative and insignificant cross correlations respectively. This can be helpful in developing small dynamic models of a set of series.

@MVIdent( options ) start end

# list of series

Parameters

start, end

range of use. By default, the common range of the listed series.

Options

LAGS=number of lags [12]

Example

This does a five lag/lead analysis of the autocorrelations and crosscorrelations of a pair of return series (for IBM stock and the general SP500).

* Tsay, Analysis of Financial Time Series, 3rd edition

* Example 8.1 from pp 393-395

open data m-ibmsp2608.txt

calendar(m) 1926:1

data(format=prn,org=columns) 1926:01 2008:12 date ibm sp

* Transform to log return percentages

set ibm = 100.0*log(ibm+1)

set sp = 100.0*log(sp+1)

@MVIdent(lags=5)

# ibm sp

Sample Output

The top gives the estimated correlations. The top row is the correlation between IBM (the first series) with lags of either IBM or SP, and the bottom row is SP with lags of IBM or SP. So .098 (the second number in the first row under "Lag 1") is the correlation between IBM(t) and SP(t-1). The bottom shows the sign of any significant correlations using a \(1/\sqrt T \) approximation to the standard error. "." indicates an insignificant correlation.

Lag 1 | Lag 2 | Lag 3 | Lag 4 | Lag 5 |

0.044 0.098 0.001 -0.078 -0.014 -0.058 -0.028 -0.031 0.024 0.081

0.037 0.084 0.015 -0.018 -0.052 -0.095 0.038 0.027 0.005 0.089

. + . - . . . . . +

. + . . . - . . . +