That's the question posed by Howard Gold in this MarketWatch article:

Don't put all your eggs in one basket.

The truth behind that cliché is that some investments zig when others zag, so by diversifying you reduce your potential losses and lower your risk. That's why Nobel Prize-winning economist Harry Markowitz called diversification the only free lunch.

But what if all the baskets were floating in the same direction, and the only one that wasn't was filled with rotten eggs?

That's the problem many investors face now. Risky assets, like stocks, commodities, and real estate are moving in lockstep. Only long-dated bonds are swimming against the tide, and nobody wants them because of fears the Federal Reserve will eventually stop its massive bond buying and raise interest rates.

So, the choice appears to be throwing even more money into stocks, which are nearly five years into a bull market, or buying bonds, which we know will go down in price. Or keeping more in cash (with its negative real return) or stuffing money in the mattress. Or, God forbid, buying leveraged inverse ETFs as a "hedge." (emphasis added)

Mr. Gold then posts two charts, which I don't feel like posting because they are basically worthless, showing the correlation of various asset classes from 2003-2007 and from 2008 to 2012. What these charts show is that correlation among most assets has risen in the latter period with one exception - 20 year US Treasury bonds. The problem with both charts is that they aren't based on enough data to be useful. Correlations over a four year period tell you absolutely nothing about how to allocate your long term investments. They tell you even less about what might happen in the short term.

For someone considering how they should diversify their assets, the phrase I bolded in the quote above is the very reason asset allocation doesn't work for most people. Mr. Gold believes long term bonds are in a secular bear market and he is entitled to his opinion, but with all due respect he doesn't know any such thing. He is making a value judgment about an asset class and eliminating it as an allocation choice. Asset allocation is about developing a long term plan using a defined set of asset classes based on your risk tolerance. The only reason to change your long term allocation is that your circumstances have changed. You've gotten older or you've inherited a bunch of money or something else that affects your ability to reach your goals while also being able to sleep at night. What you don't do is make an assumption about the future and abandon a plan that is based on decades - rather than four years - of data.

This is how so many investors end up buying high and selling low. They base their long term investing plan on short term market movements and when it doesn't work out, they abandon it for the latest new, new thing. In other words, asset allocation isn't working because most people are doing it wrong. 

The fact that Howard Gold and a whole bunch of other people think long term bonds can do nothing but go down is, in my book, a very, very good reason to think they might just be a good place to park some cash right now. I also think US stocks - or at least the S&P 500 as a whole - are probably overvalued relative to history. But I wouldn't eliminate stocks from my asset allocation choices. Having an asset allocation plan is mostly about forcing you to do things that make you feel uncomfortable. It meant buying stocks in early 2009 (assuming you rebalance yearly) and it probably means buying long term bonds now. 

BTW, one thing Howard gets very right is the part about leveraged inverse ETFs. Stay away, stay very far away. These things should have warning labels.