smooting data series

[Martin Desjardins]

Hi,

I would like to analyse the four moments of differents distributions of annual return but I don't have enough data. Could I use a simple moving average of monthly returns to "create" data ? I only want to have a rough estimate of the distribution moments, especially the skewness and kurtosis.
I have not find any "simple" methods to annualize the third and fourth moments of a distribution.

Thanks in advance,

 

[Ken Abbott]

I think that would be hard to justify. What data, exactly?

 

[Martin Desjardins]

I think that would be hard to justify. What data, exactly?

stock index returns , with index like MSCI Infra and MSCI Real Estates there's no more than like 8 years of data.

 

[euroazn]

No no no no no no no
This is incredibly bad for a billion reasons, but the biggest thing it introduces is all kinds of nasty autocorrelations.

You can not do this

 

[Ken Abbott]

No no no no no no no
This is incredibly bad for a billion reasons, but the biggest thing it introduces is all kinds of nasty autocorrelations.

You can not do this

What he said.

 

[Martin Desjardins]

Thanks guys

 

[vertigo]

If that doesn't work, you can pour some hot chocolate over the data. That will make it very smooth.

 

[Ken Abbott]

If that doesn't work, you can pour some hot chocolate over the data. That will make it very smooth.

Even better: Truffle Butter

 

[PepeQuant]

oh man...try cubic/quadratic smooth spline.

 

[Daniel Duffy]

oh man...try cubic/quadratic smooth spline.

why not quintic?