Positive Autocorrelation and hedge fund illiqudity

[Liming]

Dear David,

I'm struggling :-/ between the concept of positive autocorrelation and hedge fund liquidity. Appreciate that you can give some helping hand on it.

1) I understand that a positive autocorrelation, which is characterized by a positive and significant beta coefficient in fund return regression on its own lagged value, implies that the fund returns exhibit a trend. A upward return movement will be followed by another upward return movement and conversely for downward return movement.

2) I also read from some readings that hedge fund illiqudity risk can be represented by a positive first-order autocorrelation as well.

So to me, it seems a hedge fund which is perfectly liquid but simply exhibiting upward trend and a hedge fund which is illiquid will both have positive autocorrelation. Is this correct? and if yes, it's really quite confusing to me because
first of all, I can't link positive autocorrelation to liquidity, understanding that simply uptrend fund also have positive autocorrelation. I also don't understand why less frequent market to market or less liquid price will contribute to positive autocorrelation.
second of all, my first intuition tells me that 1) and 2) shouldn't coexist because in the case when an analayst is analyzing a fund return that has positive autocorrelation, it will become a totally subjective exercise left to the analyst to decide whether it is due to uptrend or to poor liquidity.

Thank you very much!

Cheers
Liming
8/10/2009

 

[Hardy Noman]

Hi Shakti / David (@David Harper, CFA, FRM, CIPM , ShaktiRathore)

I have the same doubt,
can you please clarify.

thank you!

 

[ShaktiRathore]

Hi,
The concepts is simple its this(according to my understanding):
The illiquidity of hedge funds requires us to price them frequently as these funds does not trade in intermittent periods. As a result they do not have market available prices , so that hedge fund managers price these hedge funds so that the returns are smoothed out so as to reduce the volatility of the fund and thus increase the Sharpe ratio which is the main measure of performance for the hedge funds. As returns are intently smoothed out and are not the actual market value returns these causes the returns to show less deviations around the mean.
That is if returns are negative at some point of time and positive in others e.g. 15%,-2%,15%,-7%,-1%(mean=20/5=4%, stdDevition=10.29%) etc this tends to show negative autocorrelation and high volatility of returns around the mean ,the manager to inflate performance and lower volatility would show the returns like 1%,2%,4%,6,7(mean same=20/5=4%,stdDev=2.54%) which exhibits a continues improvement in performance and less standard deviation around the mean and a positive serial correlation, This shows that a negative and positive returns together causes volatility to increase as the are on opposite sides of means but the positive returns above produced a lesser volatility in way producing positive serial correlation.
So that overall fund mean return over the period are same but the smoothing of returns due to pricing of fund by managers themselves due to unavailability of true market prices causes the manager to reduce volatility to inflate performance thus making all returns positive , avoiding neagative returns (prohibiting or decreasing the negative autocorrelation) and in this way producing positive serial correlation. In this way the positive autocorrelation also serves as best measure of funds illiquidity.
thanks

 

[Hardy Noman]

Hi Shakti,

Thanks for the explanation.

I understand that Hedge Fund manager has an intention to smooth out returns....But there could be a hedge fund manager who is invested in fairly liquid investments and performs really well (with a consistent or increasing returns). In this case the positive autocorrelation could falsely conclude that the hedge fund is illiquid. while the fact is the Hedge fund assets are not only liquid, but also the manager performed well.

Im not sure, but is my example sort of a Type 2 error (in the autocorrelation - liquidity model) ?? or is the model flawed in a way??

Thanks a million!
Cheers.

 

[ShaktiRathore]

At the first place i shall tell you that hedge funds are usually and mostly illiquid. Like investment in private equity/ venture capital,real estate and certain other instruments which are illiquid. And i have talked and explained from this point of view.
If the assets are liquid which is rarely the case i think than serial positive correlation can hardly predict illiquidity but we can always predict illiquidity by measuring the autoserial correlation at a given level of significance. If we are confident that the auto correlation is significant(above some threshold value) than we can infer some illiquidity but if the investments are already liquid (i think if we know the composition of hedge fund)why to measure the illiquidity at first place.
thanks

 

[Hardy Noman]

Yes...i get it now.

sorry for the silly question !

thank you very much for your time! ShaktiRathore

 

[sharman.jamie]

If anyone has found this thread and needs more info, the pages in Chapter 13 Asset Management (Chapter 4 P69 Risk Management and Investment Management If you bought the GARP books) Is really helpful