That means one set of exposures if “almost perfectly offset” by exposures in the opposing ETF. As he writes:
“…for every dollar invested in ETFs that provides a positive factor exposure, there is also a dollar invested in other ETFs that (implicitly) provide a similar-sized negative exposure towards the same factor.”
That takes care of the overcrowding problem he says.
But it creates another issue – turns out the only thing that’s left is a whole lot of beta.
As Blitz explains, “despite a large variation in exposures across funds, all that remains when everything is added up is plain market beta exposure.”
For this he offers an example – ETFs that provide targeted exposure to low volatility stocks are capturing similar a similar premium captured by some sector or high dividend ETFs. In the case of ETFs that are looking for the same stocks, there are an equally large number of ETFs looking for stocks that are looking for negative exposure to those factors.
Taken in aggregate, they neutralize the exposures of low volatility funds.
So while the trades themselves might not be crowded, the net exposure to some of these factors is “indistinguishable from zero” he writes.
You can download the paper here.