Hacker News Reader

Comment by tosseraccount | Hacker News Reader

"everything is correlated almost to a degree of 1"

Nonsense. Evidence : https://www.portfoliovisualizer.com/asset-class-correlations ( 2009-2015 time period )

Example: VTI and TLT (total stocks vs. total bonds) has -0.34 correlation.

Is there similar data for longer periods of time? Seems like it be interesting to see that correlation in periods that include some recessions or am I completely off the mark?

Figure 2 illustrates five-year correlations between monthly U.S. stock and U.S. bond total returns over five-year intervals since 1926 (17 distinct, nonoverlapping periods). While the long-term average correlation between these two asset classes has been 0.25, the figure shows that correlations over shorter windows vary widely from this average, with a range of 0.72 for the five years ended 1975 to –0.54 for the five years ended 2005.

http://www.vanguard.com/pdf/s130.pdf

Lets not be silly here -- of course not everything ends up with a correlation of exactly 1 during crashes. It is an exaggeration demonstrating a general trend.

In 2008, bonds AND stocks fell because of all-out panic where individuals dumped everything and tried to go into cash. The only thing not falling were US Treasuries since that is where all the capital ended up.

In 2015, we have a different issue -- a lot of the market is controlled by algos which, upon seeing instability, sell and go to cash. This causes a valley for all asset classes except the safe harbours (US Treasuries...maybe gold.)

Markets are controlled by algorithms? Has anyone done a regression analysis on "algo" trading determining market prices? Is it really significant?

What percentage of the market would you estimate to be controlled/managed by algorithms?

Sure - for the 2009-2015 time period.

But the claim was not "everything is correlated almost to a degree of 1". The actual claim was "when things go wrong everything is correlated almost to a degree of 1". Look at the data from 2008, not 2009-2015.

Do you have a link for 2008?

He said "... bad advice as they are now positively correlated."

No. I heard it as a common, repeated observation about 2008, but I don't have a link to actual data.

Yet in your previous comment you were urging people to look at the data... perhaps you should follow your own advice?

Well, my point was that tosseraccount's evidence that the claim was wrong did not in fact apply. I was not actually asserting that the claim was true (though I believe it is), merely that the evidence cited against it was irrelevant because it was the wrong time period.

If you look hard enough, you will find specific periods with any values of the correlation. Portfolio is about long-term strategy. That is not to say that I disagree that asset classes became more correlated lately.

https://www.portfoliovisualizer.com/asset-correlations?s=y&n...

Correlation = -0.49 for 2008.

I see. So in 2008, what seems to have happened is a flight to safety, and therefore bonds and the stock market were very much not correlated.

I stand corrected.

You can't use a treasury ETF as a proxy for the whole bond market.

The quote was about corporate bonds and TLT is a US treasury ETF. I hold TLT in my portfolio as it one of the only ETFs I can find that tends not to move in tandem with absolutely everything else.

Reply on news.ycombinator.com