Market returns / need to hold best shares / holding shares until they die

[Lootman]

The S&P biotech equal weighted fund XBI seems to exclude Amgen, Gilead and Moderna, some of the biggest biotechs. It’s a misnomer to call a global cap weighted tracker with 3500 stocks ‘cap weighting the market’, but it seems a much bigger misrepresentation to call a fund equal weighted if it omits three of the biggest in the business.

No, XBI does not exclude AMGN, GILD etc. It simply holds them in equal weight to the other smaller names. You can see this here:

https://stockanalysis.com/etf/xbi/holdings/

[Lootman]

The S&P biotech equal weighted fund XBI seems to exclude Amgen, Gilead and Moderna, some of the biggest biotechs. It’s a misnomer to call a global cap weighted tracker with 3500 stocks ‘cap weighting the market’, but it seems a much bigger misrepresentation to call a fund equal weighted if it omits three of the biggest in the business.

No, XBI does not exclude AMGN, GILD, MRNA etc. It simply holds them in equal weight to the other smaller names. You can see this here:

https://stockanalysis.com/etf/xbi/holdings/

[JohnW]

Thanks, I should be more careful with numbers. I looked at https://www.etf.com/XBI and saw those missing from the biggest holdings; classic error.
Risking it again, there might be 250 biotech stocks on US markets, with 150 in XBI. Selectively ‘equal weighted’.

[Bubblesofearth]

It’s starting to sound like a stock picking active management approach to investing if 60 stocks will represent 10,000, rather than a ‘buy the market but do it equally’ approach. And the further it deviates from that the less the attractive theory is operating.
The S&P biotech equal weighted fund XBI seems to exclude Amgen, Gilead and Moderna, some of the biggest biotechs. It’s a misnomer to call a global cap weighted tracker with 3500 stocks ‘cap weighting the market’, but it seems a much bigger misrepresentation to call a fund equal weighted if it omits three of the biggest in the business.

Active management implies regular buy and sell decisions which is the opposite of buy and leave alone.

You are not 'buying the market' by investing in a portfolio of shares, you are trying to take risk and reward levels to close to those that would pertain to an equal weight portfolio of every traded share. It is of necessity a compromise. Lots of studies have shown that (volatility) risk is reduced to close to that of the market once you get beyond about 10-15 shares. What I'm suggesting is that you can get close to (equal weight) market reward once you get to 50-60 shares. The latter number is greater than the former because of the already discussed skewness of share performance.

BoE

[hiriskpaul]

The main "analysis" there relies on a claim that seems to go against what most people say - that capitalization-weighted index trackers outperform most funds. The claim on that board, without a source that I can see (presumably discussed there so much they all take it as read) is

"Remember that conventional market-wide cap-weight indexes are big outliers in terms of [bad] performance: almost any other portfolio weighting method you try will usually do better, most obviously equal weight."

Now, I haven't seen comparisons with equal weight indexes, but active funds are said, in pretty much all analyses of developed markets, to underperform cap-weighted index trackers.

There are backtests that show many systematic investing strategies outperform cap-weighted. This paper is one I have always found amusing

https://www.researchaffiliates.com/cont ... monkey.pdf

Following the reverse of many proposed strategies do as well if not better than the original strategies and dart throwing monkey portfolios also outperform. With results like these, it is astonishing how badly actively managed funds do in comparison, with the majority not even able to outperform a cap weighted index, let alone dart throwing monkeys.

[EthicsGradient]

The main "analysis" there relies on a claim that seems to go against what most people say - that capitalization-weighted index trackers outperform most funds. The claim on that board, without a source that I can see (presumably discussed there so much they all take it as read) is

"Remember that conventional market-wide cap-weight indexes are big outliers in terms of [bad] performance: almost any other portfolio weighting method you try will usually do better, most obviously equal weight."

Now, I haven't seen comparisons with equal weight indexes, but active funds are said, in pretty much all analyses of developed markets, to underperform cap-weighted index trackers.

There are backtests that show many systematic investing strategies outperform cap-weighted. This paper is one I have always found amusing

https://www.researchaffiliates.com/cont ... monkey.pdf

Following the reverse of many proposed strategies do as well if not better than the original strategies and dart throwing monkey portfolios also outperform. With results like these, it is astonishing how badly actively managed funds do in comparison, with the majority not even able to outperform a cap weighted index, let alone dart throwing monkeys.

Though as that admits, they didn't take trading costs into consideration. We saw that in practice, a cap-weighted S&P 500 tracker usually outperformed an equal-weighted one, though people claim that equal weighting "is better"; a cap-weighted tracker probably involves the least churn for a collective investment, so maybe that is just what gives them the real-life edge.

[hiriskpaul]

There are backtests that show many systematic investing strategies outperform cap-weighted. This paper is one I have always found amusing

https://www.researchaffiliates.com/cont ... monkey.pdf

Following the reverse of many proposed strategies do as well if not better than the original strategies and dart throwing monkey portfolios also outperform. With results like these, it is astonishing how badly actively managed funds do in comparison, with the majority not even able to outperform a cap weighted index, let alone dart throwing monkeys.

Though as that admits, they didn't take trading costs into consideration. We saw that in practice, a cap-weighted S&P 500 tracker usually outperformed an equal-weighted one, though people claim that equal weighting "is better"; a cap-weighted tracker probably involves the least churn for a collective investment, so maybe that is just what gives them the real-life edge.

Agreed. There is also the issue that the paper was written in 2013 and the period investigated was very good for size, value and momentum factors. I would not be at all surprised if cap-weighting would have performed much better compared to the other strategies over the last 10 years when investing in size, value and momentum factors would have not helped returns anywhere near as much and may have even dragged returns.

[hiriskpaul]

Agreed. There is also the issue that the paper was written in 2013 and the period investigated was very good for size, value and momentum factors. I would not be at all surprised if cap-weighting would have performed much better compared to the other strategies over the last 10 years when investing in size, value and momentum factors would have not helped returns anywhere near as much and may have even dragged returns.

To illustrate, 10 year annualised returns from iShares Us listed S&P 500, momentum, value and size ETFs:

S&P 500   12.16%
Momentum  12.05%
Size      10.47%
Value      8.62%

[vand]

Well this is the whole theory behind passive investing, isn't it? Passive as in a market cap weighted global index.

- Most shares underperform the index, due to the competitive nature of a capitalist system where the best companies grow at the expense of their competitors
- Therefore the odds are stacked against anyone trying to pick out the winners from the companies that will eventually be outcompeted
- By holding a basket of everything you ensure that you are holding all the winners, all the losers, and everything in between
- The small number of winners will more than compensate for the large number of losers over time

The last decade has been a textbook example of this, where we have had the incredible growth bigTech driving most of the gains in the market overall. Unless you were concentrated in this small handful of companies then you will have underperformed the market, without question. It has probably never been more difficult to beat the market than over the last decade.

The theory is sound from a mathematical and economic sense imo, and there is obvious empirical evidence when you look at the miserable performance of active funds.

The other thing to point out with a passive index strategy is that it is also, by definition, a trend following strategy in disguise - you remain increasingly long on the components of the index that have the most positive outperformance, and lessen your exposure to components that are reducing in weighting. That's not the case with an equal-weighted index.

[Bubblesofearth]

Well this is the whole theory behind passive investing, isn't it? Passive as in a market cap weighted global index.

- Most shares underperform the index, due to the competitive nature of a capitalist system where the best companies grow at the expense of their competitors
- Therefore the odds are stacked against anyone trying to pick out the winners from the companies that will eventually be outcompeted
- By holding a basket of everything you ensure that you are holding all the winners, all the losers, and everything in between
- The small number of winners will more than compensate for the large number of losers over time

The last decade has been a textbook example of this, where we have had the incredible growth bigTech driving most of the gains in the market overall. Unless you were concentrated in this small handful of companies then you will have underperformed the market, without question. It has probably never been more difficult to beat the market than over the last decade.

The theory is sound from a mathematical and economic sense imo, and there is obvious empirical evidence when you look at the miserable performance of active funds.

The other thing to point out with a passive index strategy is that it is also, by definition, a trend following strategy in disguise - you remain increasingly long on the components of the index that have the most positive outperformance, and lessen your exposure to components that are reducing in weighting. That's not the case with an equal-weighted index.

Big winners tend to start small. It's unusual for a very large company to multi-bag. This means that holdings of companies that will go on to be big winners will typically be small in cap-weighted indices and their trackers.

An initial equal weight portfolio doesn't need to contain all available shares. As long as it contains sufficient to give a high probability of capturing a representative number, relative to the market, then it should outperform the cap-weighted index.

Note, I'm not talking about active funds or equal weight trackers but rather bespoke portfolios of shares that individuals hereabouts can set up. The HYP portfolios are examples of these but they do not need to be restricted by yield or geography.

BoE

[vand]

Big winners tend to start small. It's unusual for a very large company to multi-bag. This means that holdings of companies that will go on to be big winners will typically be small in cap-weighted indices and their trackers.

An initial equal weight portfolio doesn't need to contain all available shares. As long as it contains sufficient to give a high probability of capturing a representative number, relative to the market, then it should outperform the cap-weighted index.

Then.. why doesn't it beat market cap weighted?

The problem with an equal weighted - how many shares are you holding? The global indexes contain about 3000-4000 publically traded securities in developed markets. Say you own 1000 in an equal weighted index. That means each company gets 0.1% weighting. An individual stock can thousand-bag (as some have done) and you are not able to capture any of that because you are continually selling and rebalancing in the EW index.

If EW is better then an EW index would outperform cap-weighted across most timeframes... and you know what? it just doesn't.

[tjh290633]

Then.. why doesn't it beat market cap weighted?

Over the years I have consitently beaten the FTSE100, using an equal weight approach. Here are the results sine 1987:

.            Income Units              Accumulation   Rebased   Rebased   April   Rebased
Year to      Unit Value     Div/Unit   Unit Value     FT30      FT100     RPI     RPI    
21-Apr-87            1.00       0.00           1.00      1.00      1.00   1.018      1.00
05-Apr-88            0.91       2.86           0.94      0.92      0.91   1.058      1.04
05-Apr-89            1.18       2.72           1.28      1.10      1.05   1.143      1.12
05-Apr-90            1.21       4.24           1.40      1.13      1.14   1.251      1.23
05-Apr-91            1.34       5.42           1.69      1.28      1.26   1.331      1.31
05-Apr-92            1.30       7.52           1.75      1.24      1.26   1.388      1.36
05-Apr-93            1.51       6.91           2.13      1.44      1.46   1.406      1.38
05-Apr-94            1.70       6.27           2.50      1.65      1.65   1.442      1.42
05-Apr-95            1.66       7.48           2.55      1.57      1.62   1.490      1.46
05-Apr-96            1.95       7.38           3.13      1.80      1.90   1.526      1.50
05-Apr-97            2.16       8.40           3.62      1.85      2.21   1.563      1.54
05-Apr-98            3.31      10.00           5.72      2.45      3.05   1.626      1.60
05-Apr-99            3.44       8.46           6.12      2.47      3.21   1.652      1.62
05-Apr-00            3.32      11.33           6.13      2.42      3.35   1.701      1.67
05-Apr-01            3.29      12.42           6.32      2.05      2.89   1.731      1.70
05-Apr-02            3.37      13.02           6.76      1.65      2.69   1.757      1.73
05-Apr-03            2.29      12.10           4.85      0.85      1.85   1.812      1.78
05-Apr-04            2.92      13.38           6.56      1.22      2.25   1.857      1.82
05-Apr-05            3.46      13.06           8.10      1.33      2.51   1.916      1.88
05-Apr-06            4.30      17.42          10.57      1.68      3.06   1.965      1.93
05-Apr-07            4.91      19.42          12.63      1.90      3.31   2.054      2.02
05-Apr-08            4.14      24.32          11.21      1.58      2.93   2.140      2.10
05-Apr-09            2.28      21.17           6.46      0.87      2.01   2.115      2.08
05-Apr-10            3.69      11.06          10.86      1.33      2.91   2.228      2.19
05-Apr-11            4.16      16.71          12.76      1.43      3.03   2.344      2.30
05-Apr-12            4.40      17.73          14.19      1.33      2.96   2.408      2.37
05-Apr-13            5.27      21.83          17.01      1.54      3.29   2.476      2.43
05-Apr-14            5.34      23.05          18.88      1.75      3.38   2.557      2.51
05-Apr-15            5.91      24.98          21.84      1.91      3.47   2.580      2.53
05-Apr-16            5.92      22.67          21.72      1.79      3.17   2.614      2.57
05-Apr-17            6.62      26.21          25.47      2.10      3.76   2.706      2.66
05-Apr-18            6.12      33.19          24.66      1.79      3.62   2.797      2.75
05-Apr-19            6.35      31.25          27.04      1.95      3.82   2.856      2.81
05-Apr-20            4.60      31.57          26.64      1.44      2.77   2.926      2.87
05-Apr-21            5.99      22.54          29.07      1.76      3.46   2.960      2.89
05-Apr-22            6.70      26.76          34.29      1.69      3.86   3.346      3.29
05-Apr-23            6.22      40.34          34.18      1.73      3.93   3.645      3.58

Not every year, I grant you, but over the years, certainly. The dividend income has thoroughly outpaced the RPI.

TJH

[1nvest]

Then.. why doesn't it beat market cap weighted?

Over the years I have consitently beaten the FTSE100, using an equal weight approach. Here are the results sine 1987:

.            Income Units              Accumulation   Rebased   Rebased   April   Rebased
Year to      Unit Value     Div/Unit   Unit Value     FT30      FT100     RPI     RPI    
21-Apr-87            1.00       0.00           1.00      1.00      1.00   1.018      1.00
05-Apr-88            0.91       2.86           0.94      0.92      0.91   1.058      1.04
05-Apr-89            1.18       2.72           1.28      1.10      1.05   1.143      1.12
05-Apr-90            1.21       4.24           1.40      1.13      1.14   1.251      1.23
05-Apr-91            1.34       5.42           1.69      1.28      1.26   1.331      1.31
05-Apr-92            1.30       7.52           1.75      1.24      1.26   1.388      1.36
05-Apr-93            1.51       6.91           2.13      1.44      1.46   1.406      1.38
05-Apr-94            1.70       6.27           2.50      1.65      1.65   1.442      1.42
05-Apr-95            1.66       7.48           2.55      1.57      1.62   1.490      1.46
05-Apr-96            1.95       7.38           3.13      1.80      1.90   1.526      1.50
05-Apr-97            2.16       8.40           3.62      1.85      2.21   1.563      1.54
05-Apr-98            3.31      10.00           5.72      2.45      3.05   1.626      1.60
05-Apr-99            3.44       8.46           6.12      2.47      3.21   1.652      1.62
05-Apr-00            3.32      11.33           6.13      2.42      3.35   1.701      1.67
05-Apr-01            3.29      12.42           6.32      2.05      2.89   1.731      1.70
05-Apr-02            3.37      13.02           6.76      1.65      2.69   1.757      1.73
05-Apr-03            2.29      12.10           4.85      0.85      1.85   1.812      1.78
05-Apr-04            2.92      13.38           6.56      1.22      2.25   1.857      1.82
05-Apr-05            3.46      13.06           8.10      1.33      2.51   1.916      1.88
05-Apr-06            4.30      17.42          10.57      1.68      3.06   1.965      1.93
05-Apr-07            4.91      19.42          12.63      1.90      3.31   2.054      2.02
05-Apr-08            4.14      24.32          11.21      1.58      2.93   2.140      2.10
05-Apr-09            2.28      21.17           6.46      0.87      2.01   2.115      2.08
05-Apr-10            3.69      11.06          10.86      1.33      2.91   2.228      2.19
05-Apr-11            4.16      16.71          12.76      1.43      3.03   2.344      2.30
05-Apr-12            4.40      17.73          14.19      1.33      2.96   2.408      2.37
05-Apr-13            5.27      21.83          17.01      1.54      3.29   2.476      2.43
05-Apr-14            5.34      23.05          18.88      1.75      3.38   2.557      2.51
05-Apr-15            5.91      24.98          21.84      1.91      3.47   2.580      2.53
05-Apr-16            5.92      22.67          21.72      1.79      3.17   2.614      2.57
05-Apr-17            6.62      26.21          25.47      2.10      3.76   2.706      2.66
05-Apr-18            6.12      33.19          24.66      1.79      3.62   2.797      2.75
05-Apr-19            6.35      31.25          27.04      1.95      3.82   2.856      2.81
05-Apr-20            4.60      31.57          26.64      1.44      2.77   2.926      2.87
05-Apr-21            5.99      22.54          29.07      1.76      3.46   2.960      2.89
05-Apr-22            6.70      26.76          34.29      1.69      3.86   3.346      3.29
05-Apr-23            6.22      40.34          34.18      1.73      3.93   3.645      3.58

Not every year, I grant you, but over the years, certainly. The dividend income has thoroughly outpaced the RPI.

TJH

Here are the end of March yearly S&P500 accumulation returns for VFINX (S&P500) adjusted to GB Pounds

Code: Select all

To April   VFINX £
1987   1
1988   0.78
1989   1.03
1990   1.25
1991   1.34
1992   1.50
1993   1.98
1994   2.04
1995   2.16
1996   3.02
1997   3.36
1998   4.87
1999   6.00
2000   7.17
2001   6.31
2002   6.28
2003   4.26
2004   4.94
2005   5.13
2006   6.21
2007   6.13
2008   5.76
2009   4.95
2010   6.98
2011   7.63
2012   8.31
2013   9.94
2014   11.02
2015   13.93
2016   14.62
2017   19.63
2018   19.97
2019   23.50
2020   22.85
2021   32.22
2022   39.02
2023   38.23


notable is that compared pretty closely to TJH HYP Accum from 1987 up to April 2001, but then TJH HYP accum pulled ahead into a distinct lead peaking in 2007 where accumulation units were over twice that of VFINX£, but subsequently has lagged to where the total returns have declined back down to (below) that of VFINX£

Data sourced from PV for VFINX and FRED for £/$ end of month

[hiriskpaul]

Investing by equal weight is another way of increasing value and smaller cap shares in the portfolio, so would be expected to do well when value and small caps do well. As I illustrated above, value and small cap investing has not done well compared to the market and growth/large caps over the last decade. Neither have equal weight ETFs.

[Bubblesofearth]

Then.. why doesn't it beat market cap weighted?

The problem with an equal weighted - how many shares are you holding? The global indexes contain about 3000-4000 publically traded securities in developed markets. Say you own 1000 in an equal weighted index. That means each company gets 0.1% weighting. An individual stock can thousand-bag (as some have done) and you are not able to capture any of that because you are continually selling and rebalancing in the EW index.

If EW is better then an EW index would outperform cap-weighted across most timeframes... and you know what? it just doesn't.

I'm not talking about EW indices or rebalancing. I'm talking about an initial EW portfolio of shares that is then left alone. A portfolio of maybe 50-60 shares. You don't need to capture all future big winners, just a representative sample.

If buying individual shares isn't your bag then an alternative approach would be to buy ETF's or IT's for different geographies and to EW these on purchase then leave alone.

BoE

[EthicsGradient]

Then.. why doesn't it beat market cap weighted?

Over the years I have consitently beaten the FTSE100, using an equal weight approach. Here are the results sine 1987:
...
Not every year, I grant you, but over the years, certainly. The dividend income has thoroughly outpaced the RPI.

TJH

When you say "equal weight", do you mean an "equal-weight-at-start-and then-leave" approach that others here are suggesting, or an "equal-weight-and-rebalance" approach that a tracker could take? What was the scenario - a lump sum at the start, dividing equally between a certain number of shares (and how many), or did you have extra investments, beyond the dividend income, to make, over time - in which case how did you allocate it? Was there a way of adding companies new to the LSE over the years, or has this all been with the companies and their successors from 1987?

[tjh290633]

Over the years I have consitently beaten the FTSE100, using an equal weight approach. Here are the results sine 1987:
...
Not every year, I grant you, but over the years, certainly. The dividend income has thoroughly outpaced the RPI.

TJH

When you say "equal weight", do you mean an "equal-weight-at-start-and then-leave" approach that others here are suggesting, or an "equal-weight-and-rebalance" approach that a tracker could take? What was the scenario - a lump sum at the start, dividing equally between a certain number of shares (and how many), or did you have extra investments, beyond the dividend income, to make, over time - in which case how did you allocate it? Was there a way of adding companies new to the LSE over the years, or has this all been with the companies and their successors from 1987?

By equal weight I mean that new holdings have been bought at median weight at the time of purchase. Some have been bought in a couple of instalments to bring them up to that level.

The scenario is that the portfolio started as a PEP in 1987, investing the maximum permitted for the first four years, and reinvesting the dividends. A couple more subscriptions were made in the late 1990s, then a parallel ISA was opened in 1999 with no duplication of holdings. It took four annual subscriptions to bring the holdings up to the PEP median weight and the two were amalgamated when permitted. Rebalancing began in 1997 when both LLOY and the then Zeneca both rose above 10% and I decided to set a weight limit.

I also transferred in a single company PEP and another unit trust/OEIC PEP. One in specie, the other as cash. The cash was allocated to a new holding, as part of readjustments because of holdings being taken over. One or two more subscriptions were added, again leading to new holdings. I keep a watchlist of possible new holdings, but corporate actions and disposals have been the reason for most new additions. Some, like S32 and HLN, were new to the market.

I could write a book about the history of the portfolio, and some of the lessons learnt along the way. Far too long to post here.

TJH

[Lootman]

Big winners tend to start small. It's unusual for a very large company to multi-bag.

A notable recent exception is Nvidia. Even at its 2022 low of about $105, it would have been the biggest share in the UK had it been British.

More recently it has been bouncing around either side of $400 - a near quadrupling in something like 9 months! Its market cap is now around $1 trillion.

I started buying it last year after it had halved in price, and continued to buy all the way down to $110, where I started to lose my nerve, but did not sell. Pretty happy with that now.

[1nvest]

Big winners tend to start small. It's unusual for a very large company to multi-bag.

A notable recent exception is Nvidia. Even at its 2022 low of about $105, it would have been the biggest share in the UK had it been British.

More recently it has been bouncing around either side of $400 - a near quadrupling in something like 9 months! Its market cap is now around $1 trillion.

I started buying it last year after it had halved in price, and continued to buy all the way down to $110, where I started to lose my nerve, but did not sell. Pretty happy with that now.

Which just indicates what a Ponzi stocks can be. Present valuation of all shares in issue (market cap) is directed by what price present buyers/sellers of a relatively small number of shares agree at any one time. A trillion market cap value isn't actual capital, its just a figure, suggested by IF you could sell all of the shares at the present agreed share price.

The longer you buy into a Ponzi, the more you're inclined to be disappointed. Rather than initial weighting a number of stocks equally and perpetually holding and seeing one stock rise to become the relative giant holding, its wiser to take profits and more broadly diversify such capital/gains.

From my own observation, HYP1 non rebalanced and TJH HYP (rebalanced) have broadly yielded similar total returns, but where HYP1 has tended to become overweight in certain individual stocks, idiosyncratic risk. The same/similar reward with less risk is the better risk-adjusted reward.

[1nvest]

Investing by equal weight is another way of increasing value and smaller cap shares in the portfolio, so would be expected to do well when value and small caps do well. As I illustrated above, value and small cap investing has not done well compared to the market and growth/large caps over the last decade. Neither have equal weight ETFs.

But over that period the US stock index (cap weighted) has become tilted, a few great stocks such as Tesla that gained over +700% in 2020 alone. At times those great gainers can suddenly drop. 2000 peak dot-com and high dot com weightings, enduring the dot com crash. Financials up to 2008 increasingly more heavily weighted enduring the financial crisis hits ...etc.

A stock universe of 100, one of which will be the next years best performing stock. With equal weight you take a neutral stance, invest £1 in each stock equally. With cap weighted you let bets ride, last years winner might have £1.50 riding on that, whilst another might have just 50p riding on it after a bad year. Often last years losers become the next years winners and vice-versa.