Equal weighted ETFs

[LearningMachine]

Has anyone looked into equal-weighted ETFs, e.g. RSP, compared to market-weighted ETFs?

[Jurgis]

I was very much pro equal-weighted ETFs about 5 years ago or so. Based on the faulty logic that market-weighted indexes get top heavy with overpriced stocks. Now I'm totally on the other side of the argument. It's the Internetz age of winner-takes-most. You want to hold the best companies available, you don't want to overweight (or equal weight) shitcos. Do you really want to hold a fund that sells FAANMG every time these great companies go up and buys crapco every time it goes down?

Plus currently most of the biggest market cap companies are reasonably priced (apart from TSLA maybe). I might have different opinion if FAANMG were all hugely overpriced and mid-small caps were cheap. I don't see that much.

Also once you are buying equal-weighted, you are no longer buying the market returns. You are factor investing even if you don't call it factor investing. So maybe then you are better off explicitly factor investing?

There are also structural issues with equal-weighted funds. They cannot be big or they have to hold huge stock percentages of the smallest stocks. Or they have to use synthetic positions.

YMMV. I doubt you'll outperform market-weighted much if at all. But if for some reason it makes you sleep better at night, it's also likely not going to be a horrible investment. You may not underperform a lot either.

[LearningMachine]

23 hours ago, Jurgis said:

I was very much pro equal-weighted ETFs about 5 years ago or so. Based on the faulty logic that market-weighted indexes get top heavy with overpriced stocks. Now I'm totally on the other side of the argument. It's the Internetz age of winner-takes-most. You want to hold the best companies available, you don't want to overweight (or equal weight) shitcos. Do you really want to hold a fund that sells FAANMG every time these great companies go up and buys crapco every time it goes down?

Plus currently most of the biggest market cap companies are reasonably priced (apart from TSLA maybe). I might have different opinion if FAANMG were all hugely overpriced and mid-small caps were cheap. I don't see that much.

Also once you are buying equal-weighted, you are no longer buying the market returns. You are factor investing even if you don't call it factor investing. So maybe then you are better off explicitly factor investing?

There are also structural issues with equal-weighted funds. They cannot be big or they have to hold huge stock percentages of the smallest stocks. Or they have to use synthetic positions.

YMMV. I doubt you'll outperform market-weighted much if at all. But if for some reason it makes you sleep better at night, it's also likely not going to be a horrible investment. You may not underperform a lot either.

Thanks @Jurgis for sharing.  Gives me more to think about. 

[LearningMachine]

On 5/16/2021 at 6:38 PM, Jurgis said:

I was very much pro equal-weighted ETFs about 5 years ago or so. Based on the faulty logic that market-weighted indexes get top heavy with overpriced stocks. Now I'm totally on the other side of the argument. It's the Internetz age of winner-takes-most. You want to hold the best companies available, you don't want to overweight (or equal weight) shitcos. Do you really want to hold a fund that sells FAANMG every time these great companies go up and buys crapco every time it goes down?

Plus currently most of the biggest market cap companies are reasonably priced (apart from TSLA maybe). I might have different opinion if FAANMG were all hugely overpriced and mid-small caps were cheap. I don't see that much.

Also once you are buying equal-weighted, you are no longer buying the market returns. You are factor investing even if you don't call it factor investing. So maybe then you are better off explicitly factor investing?

There are also structural issues with equal-weighted funds. They cannot be big or they have to hold huge stock percentages of the smallest stocks. Or they have to use synthetic positions.

YMMV. I doubt you'll outperform market-weighted much if at all. But if for some reason it makes you sleep better at night, it's also likely not going to be a horrible investment. You may not underperform a lot either.

@Jurgis, thinking more, whether equal-weighted or market-weighted index or large-cap-focused index will perform better can depend on whether scale results in higher "extraction-power" or not.  

In some areas, scale gives you a lot of power.  For example, having market share above certain percentage allows retailers to negotiate lower prices with suppliers.  Having a big percentage of other humans on a network already also provides advantage.   For such companies, focusing on large-cap or market-weighted indexes could work well. 

In some areas, you can get "extraction-power" without scale, e.g. in biotech, even a small company can get huge pricing power upon approval of a drug.  In such cases, scale can actually dilute the impact of a drug approval.   Another such area could be patent trolls, where even a small company holding a patent that can hold a big company hostage can provide the small company with a lot of "extraction-power".   Here, focusing on smaller-cap or equal-weighted indexes could work well, while being careful to not spread your bets across too many dot-com equivalents.   Patent-trolls are unique and there aren't too many of them traded publicly. Wonder if biotech industry is able to avoid dot-com-like startups because barrier to entry for a biotech company might be somewhat higher than a dot-com?   However, looks like only a small percentage of clinical trials are successful.  So, a company specializing in selecting the right companies and taking them through clinical trials might be the way to go here - having history of such successful investments is key though - however such a company is not cheap in the first place.

Thoughts? 

Edited May 18, 2021 by LearningMachine

[Jurgis]

@LearningMachine: You are right that equal-weighted indexes have benefits:

1. EW overweights smaller caps more. Smaller caps theoretically may grow faster, etc.
2. If stock rises without change in fundamentals, EW sells it, which is selling expensive stock. If stock drops without change in fundamentals, EW buys it - buying cheap stock.
3. EW is more diversified - theoretically - since no company can become big part of the fund.

To be fair, there are negatives, as I mentioned:

1. While underweighting large dead-wood bureaucratic companies, EW also underweights large growing monopolists.
2. EW does not capture the business growth (vs stock price increase), so it might be selling stock that had price increase but valuation decrease. Same with buying stock that may have had price decrease, but valuation deterioration.
3. The diversification is real, but frozen into the initial sector breakdown. This is a bit handled by external stock list changes.

So ultimately, I don't know if there is a right choice. Maybe the right choice is factor investing. But that comes with its own set of issues.

[Jurgis]

Biotechs: nah, there's a lot of biotech startups that are very dot-com'y. In fact, biotechs are mostly very binary: either you get 10-100x or you get zero. Sometimes you get zero after a very long time. Sometimes you get 10-100x after long time.

Yes, the right thing would be to buy a fund that has a great DD insight into biotechs. I am not aware of such mutual fund or ETF that does better than XBI. I think someone on this site has in the past mentioned a hedge fund run by someone in Boston (?) that is really good at doing deep DD. I don't remember the name of the fund or the person. . And I'm not sure I'd invest.

No, it was not Seth Klarman's fund - I know that he has a team that does deep DD on biotechs and buys some, but I don't know if his results are good. And in any case you mostly can't get into Baupost and Baupost's biotech results are swamped by non-biotech results.

I think you asked me about XBI PE on the other thread. I haven't looked. And in any case, PE is not the way to evaluate biotech company or fund. You should use something like Price/TAM*ApprovalChance. Maybe not TAM, but DAM - drug addressable market. The issue really is ApprovalChance. (Well, there's also issue of competition/how well drug sells, but that is probably easier to predict.)

[IceCreamMan]

Company ABC is a typical company. Sometimes its stock is overvalued, sometimes it's undervalued. When it's overvalued, the market cap-weighted ETF holds more of it. When it's undervalued, the market cap-weighted ETF holds less of it. The valuation-driven mean reversion causes a drag on the market cap-weighted ETF's returns. On the other hand, the equal-weighted ETF sells shares when they've appreciated relatively fast and buys shares when they've appreciated relatively slowly. (Win for the equal-weight).

Company XYZ is a company with a durable competitive advantage that earns superior returns over time and its stock price appreciates at an above-average rate over a long period of time. The market cap-weighted ETF allows the weighting of XYZ to increase over time, while the equal-weighed ETF does not. Win for the cap-weight. Sort of a Bayesian weighting approach that says "if the stock has appreciated, maybe it's a compounder, so let's increase the weighting."

Overall, the performance of the two ETFs seems to depend on the relative prevalence of companies like ABC vs. XYZ and the strength of the two factors-- valuation & ROIC. In other words, for a basket of stocks with similar ROIC but wide dispersion in the tendency to get over- or under-valued, I'd expect equal-weighted to outperform. For a basket of stocks with wide dispersion in ROIC, I'd expect market cap-weighted to outperform.

Does that make sense?

[LearningMachine]

24 minutes ago, IceCreamMan said:

Company ABC is a typical company. Sometimes its stock is overvalued, sometimes it's undervalued. When it's overvalued, the market cap-weighted ETF holds more of it. When it's undervalued, the market cap-weighted ETF holds less of it. The valuation-driven mean reversion causes a drag on the market cap-weighted ETF's returns. On the other hand, the equal-weighted ETF sells shares when they've appreciated relatively fast and buys shares when they've appreciated relatively slowly. (Win for the equal-weight).

Company XYZ is a company with a durable competitive advantage that earns superior returns over time and its stock price appreciates at an above-average rate over a long period of time. The market cap-weighted ETF allows the weighting of XYZ to increase over time, while the equal-weighed ETF does not. Win for the cap-weight. Sort of a Bayesian weighting approach that says "if the stock has appreciated, maybe it's a compounder, so let's increase the weighting."

Overall, the performance of the two ETFs seems to depend on the relative prevalence of companies like ABC vs. XYZ and the strength of the two factors-- valuation & ROIC. In other words, for a basket of stocks with similar ROIC but wide dispersion in the tendency to get over- or under-valued, I'd expect equal-weighted to outperform. For a basket of stocks with wide dispersion in ROIC, I'd expect market cap-weighted to outperform.

Does that make sense?

Well, the higher ROIC business might not be big to begin with, and it may not stay high ROIC forever.  The business might start out as small and high ROIC, when market-cap based index weighs it lower.  When the business becomes big and low ROIC, the market-cap might will actually weigh it higher. 

I wonder if market-cap weighted index is better when bigger businesses have higher ROIC, but not better if bigger businesses have lower ROIC. 

Edited May 19, 2021 by LearningMachine

[Munger_Disciple]

In a taxable account traditional market cap weighted index fund makes more sense as there is less leakage due to taxes. I think equal weighted index acts like a "value" factor as opposed to the "growth" factor of mkt cap weighted index. Equal weighted index (in a tax deferred account) makes sense if you think value will outperform growth in the future. 

[LearningMachine]

8 minutes ago, Munger_Disciple said:

In a taxable account traditional market cap weighted index fund makes more sense as there is less leakage due to taxes. I think equal weighted index acts like a "value" factor as opposed to the "growth" factor of mkt cap weighted index. Equal weighted index (in a tax deferred account) makes sense if you think value will outperform growth in the future. 

If bought as an ETF that you are going to hold forever, both equal-weighted and market-weighted ETFs generally have no leakage because portfolio-rebalancing and redemptions are done using "in-kind" transactions.  Please see https://www.invesco.com/us/en/insights/the-tax-benefits-of-etfs.html.  This is the beauty of holding an ETF that you're willing to hold forever.   With stocks you're going to have to eventually pay taxes if you cannot hold them forever, and capital gains tax-rate might go up - so, after-tax returns with all that trading of stocks might end up being close to sitting on an ETF not performing as well but that you can sit on forever.

Vanguard index funds that are paired with an ETF also generally avoid leakage due to taxes by leveraging the tax-benefits of paired ETF by doing "heartbeat" trades.  Please see https://www.bloomberg.com/graphics/2019-vanguard-mutual-fund-tax-dodge/.

[Jurgis]

10 hours ago, LearningMachine said:

I wonder if market-cap weighted index is better when bigger businesses have higher ROIC, but not better if bigger businesses have lower ROIC. 

It's not just ROIC - although perhaps you can only consider ROIC if you are going to hold forever.

At least short term, valuations matter too, like @IceCreamMan indicated in their post (which I'm gonna +1 here )

Assume an index of two companies A and B. A has ROIC 20%, B has ROIC 10%. A market cap is 10 times the size of B.

Two situations:

1. A trades at 100 P/E, B trades at 1 P/E.
2. A trades at 10 P/E. B trades at 10 P/E.

You never know (TM), but likely short term in the first situation EW will outperform, while in the second situation MW will outperform. Long term, the situation gets more based on ROIC and a bit less based on valuation. Though in extreme cases even very high ROIC business may underperform for very long time if it starts at extreme valuation.

Edited May 19, 2021 by Jurgis

[Munger_Disciple]

9 hours ago, LearningMachine said:

If bought as an ETF that you are going to hold forever, both equal-weighted and market-weighted ETFs generally have no leakage because portfolio-rebalancing and redemptions are done using "in-kind" transactions.  Please see https://www.invesco.com/us/en/insights/the-tax-benefits-of-etfs.html.  This is the beauty of holding an ETF that you're willing to hold forever.   With stocks you're going to have to eventually pay taxes if you cannot hold them forever, and capital gains tax-rate might go up - so, after-tax returns with all that trading of stocks might end up being close to sitting on an ETF not performing as well but that you can sit on forever.

Vanguard index funds that are paired with an ETF also generally avoid leakage due to taxes by leveraging the tax-benefits of paired ETF by doing "heartbeat" trades.  Please see https://www.bloomberg.com/graphics/2019-vanguard-mutual-fund-tax-dodge/.

You are right. I was thinking about mutual funds. ETFs have tax benefits over mutual funds; so tax leakage is not an issue for a rebalancing ETF like  RSP. 

[LearningMachine]

46 minutes ago, Jurgis said:

It's not just ROIC - although perhaps you can only consider ROIC if you are going to hold forever.

At least short term, valuations matter too, like @IceCreamMan indicated in their post (which I'm gonna +1 here )

Assume an index of two companies A and B. A has ROIC 20%, B has ROIC 10%. A market cap is 10 times the size of B.

Two situations:

1. A trades at 100 P/E, B trades at 1 P/E.
2. A trades at 10 P/E. B trades at 10 P/E.

You never know (TM), but likely short term in the first situation EW will outperform, while in the second situation MW will outperform. Long term, the situation gets more based on ROIC and a bit less based on valuation. Though in extreme cases even very high ROIC business may underperform for very long time if it starts at extreme valuation.

Totally agree, valuations matter too. Agreed that for situation #1, equal-weighted ETF would out-perform.

Beyond two situations above, we can also have situation #3:  

A has ROIC 10%, B has ROIC 20%. A is a very mature, huge business, and done growing while B is getting started. A's market-cap is 10x the size of B.
A trades at 10 P/E. B trades at 10 P/E.

Along with Situation #1 above, for situation #3 also, you wouldn't want to overweight A at 10x B, and equal-weighted ETF would outperform.

Edited May 19, 2021 by LearningMachine

[IceCreamMan]

18 hours ago, LearningMachine said:

@Jurgis, thinking more, whether equal-weighted or market-weighted index or large-cap-focused index will perform better can depend on whether scale results in higher "extraction-power" or not.  

In some areas, scale gives you a lot of power.  For example, having market share above certain percentage allows retailers to negotiate lower prices with suppliers.  Having a big percentage of other humans on a network already also provides advantage.   For such companies, focusing on large-cap or market-weighted indexes could work well. 

 

11 hours ago, LearningMachine said:

I wonder if market-cap weighted index is better when bigger businesses have higher ROIC, but not better if bigger businesses have lower ROIC. 

This is an interesting hypothesis. If this were true, wouldn't we see some empirical data showing that large caps outperform small caps in some markets/indexes?

Let's keep in mind that "large market cap" is not synonymous with "large company." When you buy a market-cap weighted index, you are sizing each position in proportion to how high the stock's valuation is... by definition.

There may be some "fundamentally weighted" indexes out there that you could use to more accurately overweight large companies.

[LearningMachine]

3 hours ago, IceCreamMan said:

This is an interesting hypothesis. If this were true, wouldn't we see some empirical data showing that large caps outperform small caps in some markets/indexes?

Thanks @IceCreamMan.  I am not saying large-market-cap will always out-perform small-cap.  On the contrary, it can be hard for large-cap to grow after certain-size.  That said, in some industries, scale can let large-cap extract more than small-cap can extract, and at the very least keep their "extraction-power" while some of the small-caps can burn out. 

For example, below are a couple of charts of a few ETFs, that includes comparing equal-weighted (smaller-cap-heavy) and market-cap-weighted (larger-cap-heavy) sector-focused ETFs

• Scale provides more extraction-power with Information Technology:
  • VGT (market-cap weighted): Even though it is multi-cap, it is heavily weighted towards large-cap because of being market-cap weighted. https://institutional.vanguard.com/iippdf/pdfs/FS958R.pdf
  • RYT (equal-weighted): Even though it is S&P500 only, it is heavily weighted towards smaller-cap because of equal-weighting.
• Scale provides more extraction-power with Consumer Discretionary (brands):
  • VCR (market-cap weighted): Includes high-flying Amazon, Tesla, Home Depot, McDonald's which are very high P/E, and thus may not do that well going forward.  https://institutional.vanguard.com/iippdf/pdfs/FS954R.pdf
  • RCD (equal-weighted): Probably hard to build a brand and "extraction-power" for smaller cap here. https://www.invesco.com/us-rest/contentdetail?contentId=1e0c7c23dbd92610VgnVCM1000006e36b50aRCRD
• Scale does not provide more extraction-power in BioTech because even a small-cap can get huge pricing power upon approval:
  • IBB (market-cap weighted): https://www.ishares.com/us/literature/fact-sheet/ibb-ishares-nasdaq-biotechnology-etf-fund-fact-sheet-en-us.pdf
  • XBI (equal-weighted): https://www.ssga.com/library-content/products/factsheets/etfs/us/factsheet-us-en-xbi.pdf.  Need to be careful here of dotcom-style runaway excitement where any company calling itself biotech is getting funding, and printing shares.  

Edited May 19, 2021 by LearningMachine