Personally I also agree with Stefano (something like rescale or rescale itself). Of course in this case - strictly speaking - the demeaning option is not rescaling. I agree with Jack that maybe recursive demeaning goes a long way (let alone rolling ;) ). But... _I wouldn't endow a function with too many options, I find that very R-ish,_ _but maybe it's just me. Anyway, recursive demeaning (that I have never_ _seen used anywhere, I'd be curious to see a real-life use case) is easy_ _enough to do in hansl:_ _<hansl>_ _X = mnormal(20,3)_ _M = X - cum(X) ./ seq(1, rows(X))'_ _</hansl>_ Jack (and anyone interested in), have a look in section 3.2, p.585 Michael W. McCracken & Serena Ng (2016) FRED-MD: A Monthly Database for Macroeconomic Research, Journal of Business & Economic Statistics, 34:4, 574-589, DOI: 10.1080/07350015.2015.1086655 This (recursive standardization) is a feature that I would like to see in the staticfactor package and at a later stage in the DFM package (just to replicate some of those results as in McCracken). Of course, I am not sure (because i have not given a thought) about the theoretical arguments that would support Principal Components estimation or other estimators on recursively demeaned and/or standardized data (moment conditions, consistency, convergence rates). Apparently, I had - slightly changed - the source code in staticfactor package quite a while ago to replicate the papers diffusion index. Yiannis