Implied Volatility autocorrelation

[Rudolf13100]

Quote from MasterAtWork:

Correct.That's my own opinion.

There is no such a thing. Correlation factor is just a past historical statistic, that means, no one would bet his life on it. The correlation between the two indexes is only a flow correlation. It's a very weak one.
Thus, if for any reasons people don't think Dax as a proxy of SPX but prefer Mib, the relationship between (volatility) indexes won't last anymore. And your global position is only made by both. Hence, your prior aim to bet on the spread to reduce the risk is now a much more dangerous one.

I understand that the bet would be that the historical correlation between the two will last. Before entering a position we would have to understand why such a correlation has been observed in the past and whether the fundamental circumstances explaining the high correlation have changed.
How do we call such as strategy?

 

[MasterAtWork]

Hi Rudolf13100

"I understand that the bet would be that the historical correlation between the two will last."

Yes, that is. (IMHO)


"Before entering a position we would have to understand why such a correlation has been observed in the past and whether the fundamental circumstances explaining the high correlation have changed."

It's not quite simple because the changes can exhibit very quickly. Political decisions, some liquidity matters, ...and the correlation disappears.
The fact that the Dax and the SPX are correlated is, IMO, that international investors act like DAX is representative enough of european markets. Historic crises and recoveries broke several times the correlation.

Cheers

 

[dmo]

Quote from Rudolf13100:

In essence it sounds similar to trading a currency pair, say the eur/dollars. We would sell the euro and buy the dollar following the rationale that historically the euro is high relatively to the dollar. Am I missing something?

If we didn't want to take the currency risk we could hedge it by taking a position on the curreency future or the forex, no?

I actually think that spreading S&P IV against DAX IV at extremes is a better bet than doing so with a currency pair. There's nothing really that holds a currency relationship together; it can just keep marching apart steadily for years and years.

However, S&P IV and DAX IV are, at their core, the same thing - human emotion but in two different places. The relationship between those places - the US and Germany - is close enough that it seems unlikely there would be complacency in one (low IV) and panic in the other (high IV) for long. To the contrary, the two seem quite linked, with a clear natural tendency to come together.

Human emotion is very contagious. If you see someone crying, it's hard not to feel sad. If you see someone laughing, you smile too. Panic? That's probably the most contagious emotion of all. So the VDAX/VIX correlation is not a fluke - there's a powerful human tie holding them together.

 

[MasterAtWork]

Hi Dmo,

Quote from dmo:



Human emotion is very contagious. If you see someone crying, it's hard not to feel sad. If you see someone laughing, you smile too. Panic? That's probably the most contagious emotion of all. So the VDAX/VIX correlation is not a fluke - there's a powerful human tie holding them together.

I understand what you mean, but I respectfully disagree.

Because every spread strategies are based on time space, you could be finally right and blown up.
Spreads are built on the same time. Do you remember for sure early 2000's dotcom companies. A lot of them never show a single earning. It made sens to short some and buy old blue ships against. If you had done that strategy you were blown up by margin calls, even you would be, some years after, definitly right.

 

[dmo]

Quote from MasterAtWork:

Hi Dmo,



I understand what you mean, but I respectfully disagree.

Because every spread strategies are based on time space, you could be finally right and blown up.
Spreads are built on the same time. Do you remember for sure early 2000's dotcom companies. A lot of them never show a single earning. It made sens to short some and buy old blue ships against. If you had done that strategy you were blown up by margin calls, even you would be, some years after, definitly right.

Well like I said, ya gotta base your trading on something, and this relationship strikes me as potentially the germ of something worthwhile. Of course, until I put together a few trades based upon it that actually put money in my pocket, that opinion isn't worth the electrons it's written with. Even then, the next trade could lose me all my profits and then some.

That's trading for you!

 

[Rudolf13100]

Quote from MasterAtWork:

Hi Dmo,



I understand what you mean, but I respectfully disagree.

Because every spread strategies are based on time space, you could be finally right and blown up.
Spreads are built on the same time. Do you remember for sure early 2000's dotcom companies. A lot of them never show a single earning. It made sens to short some and buy old blue ships against. If you had done that strategy you were blown up by margin calls, even you would be, some years after, definitly right.

Now, in a portfolio consisting of, say, 10 such spread positions that are not correlated to each others... Diversification could enormously decrease overall risk, no?
I don't know however if the number of potential spreads is large enough to manage a portfolio consisting uniquely of spread positions, do you?

 

[dmo]

Quote from Rudolf13100:

Now, in a portfolio consisting of, say, 10 such spread positions that are not correlated to each others... Diversification could enormously decrease overall risk, no?
I don't know however if the number of potential spreads is large enough to manage a portfolio consisting uniquely of spread positions, do you?

You might find 10 good spreads not correlated to each other, but it's unlikely they would all be at extremes and therefore playable at the same time.

 

[Rudolf13100]

Quote from dmo:

I actually think that spreading S&P IV against DAX IV at extremes is a better bet than doing so with a currency pair. There's nothing really that holds a currency relationship together; it can just keep marching apart steadily for years and years.

Regarding this point, I think it could be extremely interesting to graphically represent the distribution of the spread between the Vdax and the Vix and the distribution of a major currency pair. Then, we would compare both distributions (their respective standard deviation, skewness, kurtosis...).
That way we would see what has historically been a better bet.

Can you guys suggest an efficient way of doing that?