Thanks.. I thought they might be the same thing. And I agree, concatenating is definately a bad idea. It seems like it would be a natural extension to the estimation method to somehow simultaenously estimate one model for several observations over the same time period. I know this doesn't make sense in the traditional framework, but it might be possible to create a model that assumes intraday the relationships are stable, and they only diverge in the long rong during the close/next day open.SIC and BIC are the same thing: Schwartz information criterion is sometimes called Bayesian information criteria. The thinh about concatenating the series is that at the point where you go from day t to day t+1, and if the 2 contiguous points are not really adjacent time periods (one series ends at 4pm and the other starts at 7am), you will have a strange return, that may overwhelm the other returns(which may be hourly,1mn, etc returns)

This looks interesting. Cointegration analysis in the presence of structural breaks in the deterministic trendWhen analysing macroeconomic data it is often of relevance to allow for structural breaks in the statistical analysis. In particular, cointegration analysis in the presence of structural breaks could be of interest. We propose a cointegration model with piecewise linear trend and known break points. Within this model it is possible to test cointegration rank, restrictions on the cointegrating vector as well as restrictions on the slopes of the broken linear trend.It certainly seems like intraday data could be treated with this.. the known breakpoints are the start/end of the day. Aplicable? Does anyone have experience with modifying existing tests to use this?

BIC and SIC do not have same formulas. There is one difference in the formulas where you use LN() in SIC and 2*LOG() in BIC at some point and results are very close to each other, that might be the reason why people think that they are same. QuoteI've observed that some series are only together very tightly in the short term and only loosly over the long term. Can cointegration still be used in this case?You can use any type of series as long as your results indicate co integration. Some series will be too tightly co integrated that deviations won’t be practical enough to make money off of them unless you can make money off of spreads. QuoteAlso, re: Stochastic structural breaks. I don't think that is needed here, when estimating intraday conintegration the breaks are deterministic and known.. at the beginning an end of each day Even though you can get away without using it you should still check using stochastic structural break because some breaks occur during the day in intraday data even though they are not obvious and you won’t catch them by assuming breaks only at the opening and closing. There is a benefit to assume that breaks will be random.

I can't seem to get that paper, could someone post it or send it to me? thanks

here you goQuoteOriginally posted by: JamesH83I can't seem to get that paper, could someone post it or send it to me? thanks