When Everything Trades As One: Goldman Declares War On ETFs, Says "May Generate Negative Alpha"

Tyler Durden's picture

Overnight, Goldman's Robert Boroujerdi released a report whose conclusion we have been warning about for the past 3 years, which also happens to be its title: "ETFs: An Imperfect Hedge?"  Goldman's findings in a nutshell: "The rise of investor usage of ETFs as hedges continues. In a bid to gain quick exposure to evolving markets, avoid single stock M&A risk or take sector views, we believe the use of “blunt force” hedging via ETFs may impair portfolio returns and potentially create negative alpha." Read that again: not zero alpha, i.e., same returns as market, but negative alpha. In other words, the great cottage industry that has been the basis for so many riches for the likes of BlackRock, and that has ensnared so many gullible retail investors, is essentially a guaranteed money losing get rich quick scheme?

Who da thunk it.

What is more curious, are Goldman's observations on historical cross industry correlations because they show that in the grand scheme of things, virtually everything trades as one!

To wit:

Given the rise in trading activity in ETFs, we compare and contrast where correlations sit for individual sectors against (1) one another, (2) versus history, and (3) the market as a whole.

To begin, we frame the basis of our discussion by providing a historical analysis of daily and weekly pair-wise correlation of ETFs versus one another, summarized in Exhibits 10 and 11 below. These matrices yield insights on the growing impact of price movements in certain sectors versus each other. They also highlight the unpredictable impact that “blunt force” ETF usage can have on portfolio construction and returns.

Indeed, of the 36 sector ETF pairs we examined, correlations on a 3-year basis are higher than 70% for 31 of the instances, emphasizing the importance for portfolio managers to choose sectors wisely when hedging at specific points in time.

Although correlations for the individual components of the S&P 500 are now broadly in line with their 1-year and 5-year medians (1-month correlations now stand at 53% versus 50% and 44% on a 1- and 5-year median) we see higher levels of correlations on a sector-to-sector basis (using ETFs as a proxy). To give some context, the average sector-pair correlation is 72% when measured on a 3-month basis.

In Exhibit 10, we show a 3-month daily correlation snapshot along with a guide to the most and least correlated pairs.

To gauge correlations on a longer horizon, we include a 3-year view in Exhibit 11. We note that these figures are calculated on a weekly basis.

Current (3m) cross-sector correlations and hedging takeaways…

  • Perhaps not surprisingly, we see among the least amount of dispersion between Industrials and Energy as well as Materials and Industrials. Surprising, however, is that less obviously paired ETFs exhibit strikingly high correlations. For example, we highlight: Financials & Industrials, Tech & Discretionary and Materials & Discretionary.
  • Utilities is the sector least correlated to others and, as such, may provide a more optimal hedge if an investor has a specific sector view on the group. We do note that Staples are the second least correlated.

We provide a full listing of 3-month and 3-year sector-pair correlations in the Appendix for those who wish to dig deeper.

… and how they’ve moved: now vs history (3 months vs 3-years)

  • Tech (XLK) and Industrials (XLI) show pockets of falling correlations to other sectors, albeit the figures remain high. In addition, the correlation between the Tech and Industrials fell to 77% on a 3-month basis from 89% on a 3-year basis. Among the notable examples of falling correlations, those between Tech and Energy fell to 72% from 83% over the same two periods.
  • Healthcare (XLV) has become more coupled to other sectors. While 32 of the examined 36 sector pairs showed decreases in correlations, all of the four that showed increases included Healthcare. As an example, correlations between Healthcare and Financials rose to 76% from 73%, versus an average decrease in correlations of 8% among all 36 pairs.

How sectors are correlated to the broader market

  • Utilities (XLU), Staples (XLP) and Tech (XLK) are the sector ETFs least correlated to the S&P 500, with current 3-month correlations of 47% each.
  • Energy (XLE) is the only sector with an increase in correlations to the S&P 500, having risen to 62% on a 3-month basis versus 56% on a 3-year. In contrast, Tech (XLK) and Discretionary (XLY) saw the greatest declines in correlations. They fell to 47% from 60% and 50% from 58%, respectively, on a 3-month and 3-year basis.

* * *

So... HFT is basically done now that GETCO liquidated a third of their sales force, and the focus is now shifting on ETFs which will be the next to fall under the knife. Which leads to the question: will anyone be left trading anything in the stock market in a few years time?

full report can be found here