ETF XLK: the concentrated portfolio reduces its appeal
The SPDR technology sector, XLK, is made up of S&P 500 stocks classified as technology. This ETF is quite popular, with $48.5 billion in assets, although much smaller than QQQ, which has $194 billion in assets. It might be tempting to assume that QQQ and XLK are somewhat interchangeable, but it’s important to recognize the differences. It starts with looking at the top holdings. While QQQ and XLK both have Apple (NASDAQ:AAPL) and Microsoft (NASDAQ: MSFT) as their 2 main holdings, the wallets are otherwise not very similar. Amazon (NASDAQ:AMZN)QQQ’s 3rd largest holding is in the SPDR consumer discretionary sector, XLY, rather than XLK, for example.
Given the differences between the largest holdings, it should come as no surprise that the returns are quite different. XLK has returned 23.15% over the past 12 months and the 3-year annualized return is 36.4%, compared to QQQ’s 13% and 30.95% for those two periods, respectively.
Another way to compare these two funds is to look at a Fama-French factor analysis. From this point of view, QQQ and XLK are very similar. Both have betas slightly greater than 1 and both have factor tilts that favor large caps (negative loads on the SMB factor) and both are tilted to favor growth over value (negative loads on HML). XLK has a stronger large cap tilt, which is not surprising given the huge allocations to AAPL and MSFT.
A third way to compare QQQ and XLK is to look at their correlations with the S&P 500 (SPY) and, for context, the Vanguard Information Technology ETF (VGT). VGT holds securities very similar to XLK, but includes companies with smaller market caps. XLK holds 70% of its assets in what Morningstar classifies as giant stocks, compared to 61% for VGT. The correlations of QQQ, XLK and VGT with the S&P 500 are almost identical. The correlations of XLK and VGT to QQQ are equal, at 98%. XLK and VGT have a 100% correlation with each other. These results use 3-year return data.
While a 98% correlation between XLK and QQQ might indicate that these funds are reasonable substitutes for each other, the discrepancy in returns over the past year underscores that holdings details can have a substantial impact on performance.
I last wrote about XLK on Feb 13, 2021, almost a year ago, and gave XLK a bullish/buy rating as an indicator for tech stocks. Since then, XLK has achieved a total return of 17.4%. The bullish view was based on the market’s implied outlook, a probabilistic forecast that mirrors the consensus view in the options market.
The price of an option on a stock or ETF reflects the market’s consensus estimate of the likelihood that the price of the stock or ETF will rise above (call option) or fall below (put option) a specific level (the strike price of the option stock) by the time the option expires. By analyzing call and put option prices at a range of strike prices, all with the same expiration date, it is possible to calculate a probabilistic price return forecast that reconciles all option prices. This is called the implied market outlook and represents the consensus opinion implicit in option pricing. This technique is widely studied and I use my own implementation. To calculate the implied market outlook, options traded on the underlying stock or ETF must be relatively active. Although options trading on XLK is small relative to QQQ, there is enough trading to make the implied market outlook reasonable.
I calculated the implied market outlook for XLK and QQQ through mid-2022 and early 2023.
Implied market outlook
I have calculated the implied market outlook for XLK and QQQ for the 4.4 month period from now until June 17, 2022 and for the 11.5 month period from now until January 20, 2022 using options that expire on those dates.
The standard presentation of the implied market outlook is as a probability distribution of price return, with probability on the vertical axis and return on the horizontal.
The overall shape of the distribution is what we would expect from a diversified portfolio (as opposed to an individual stock) with a well-defined peak at a positive piercing return and a long negative tail (i.e. a negative asymmetry). The maximum probability corresponds to a price return of +5.7% for the 4.4 month period. The calculated volatility for this distribution is 17.1% (28% annualized).
To facilitate direct comparison of positive and negative return probabilities, I rotate the negative return side of the distribution around the vertical axis (see graph below).
It is useful to compare the implied market outlook for QQQ and XLK for the 4.4 month period.
The outlook for QQQ has the maximum probability of a return of +7.8% and the volatility of this distribution is 17.1% (28% annualized volatility). I look at the relationship between the maximum return of the probability and the volatility of the period. For QQQ, the ratio is 45.6% (7.8%/17.1%) versus 33% (5.7%/17.1%) for XLK. Based on this metric, QQQ provides a more attractive risk/reward proposition. This is also evident in the larger gap between the probabilities of positive and negative returns (the vertical distance between the solid blue line and the dashed red line).
Looking at the 11.5-month outlook for XLK and QQQ (calculated using options that expire on January 20, 2023), the results are consistent with a view through mid-2022.
The maximum probabilities correspond to returns of +7.5% for XLK and +10.5% for QQQ. The expected volatilities calculated from these outlooks are respectively 25.8% and 25.7% for XLK and QQQ. The ratio of maximum probability return to volatility is 29% for XLK and 41% for QQQ. QQQ looks significantly more attractive than XLK for the 11.5 month outlook.
A key difference between XLK and QQQ is that XLK is much more concentrated (AAPL and MSFT make up 47% of the portfolio). This, in turn, means that XLK is more exposed to company-specific risks associated with AAPL and MSFT.
XLK owns S&P 500 technology stocks. QQQ tracks the NASDAQ 100 and is not explicitly a technology fund/index. The big questions are (1) whether XLK is attractive on its own merits and (2) whether investors are likely to be better served by gaining technology exposure using QQQ? The market’s implied outlook suggests that the overall risk-reward proposition for QQQ is better than for XLK, although this may be entirely due to XLK’s higher level of concentration. I am reducing my rating on XLK to neutral as the implied market outlook through mid-2022 and early 2023 suggests that the specific portfolio held by this fund is not very attractive relative to QQQ. XLK has significantly outperformed QQQ over the past year and 3 years, but the market’s implied outlook indicates that this outperformance is unlikely to persist.