Author Topic: Semper Augustus letter  (Read 20979 times)


  • Lifetime Member
  • Hero Member
  • *****
  • Posts: 2583
Re: Semper Augustus letter
« Reply #60 on: March 09, 2018, 08:00:51 AM »
I got what you said. Perhaps I was not clear in my post. But overall the answer is that purely passive (market cap index) investing cannot cause prices to go from 50/50 to 75/25. They cannot cause relative market cap change.

Right, I was saying the active investors cause it in that case (or just random perturbation in a smaller case).

In market with active investors, the active investors cause the price changes. It depends on what you mean with "reinforce" and "stabilize momentum", but you may be right that indexing supports or enlarges the influence of active investors. In a sense that $1 actively invested in a company with $99 index investment could drive up the price disproportionately, because indexers are not selling at any price and if they sell, they sell proportionately all stocks in index.

Edit: Also active investor selling active positions and buying index causes the index skew towards the stocks they did not own. But it's their selling that's causing the skew. If they sold and went to cash they would still cause the same skew.

Perhaps this is the effect I mean--momentum clearly already happens, but if the passive indices reinforce or amplify that momentum, then it could potentially cause issues both during the bull market and the reversal after, since I'm assuming it would work similarly in reverse.


  • Full Member
  • ***
  • Posts: 118
Re: Semper Augustus letter
« Reply #61 on: March 11, 2018, 05:53:20 AM »
Good points.

That is the big question, how does this work in reverse? My understanding: Companies with the largest proportion of shares held by active owners, where the active owners are willing to sell shares in a downturn, will have an outsized decrease in share price?


  • Full Member
  • ***
  • Posts: 218
Re: Semper Augustus letter
« Reply #62 on: March 13, 2018, 05:43:35 AM »
I've finally got round to reading the indexing section in detail.

I think I'll need to re-read it, having thought about it, to see what is and is not being suggested in it.

It reads like a good narrative, with good reasoning and feels like it has a lot of either validity or truthiness. I certainly don't feel I'm as equipped as the author to know from the top of my head or from long experience and study how all these things work.

I can certainly envisage the instinctive actions of a large number of retail investors acting on instinct. The group I'm thinking of would be the type to start investing in equities only after a sustained period of 'consistent good performance' has been demonstrated, such as the last 2-5 years (i.e. certainly not buying low, more likely to buy near the top) and who tend to panic and sell if the market seems to be going down (i.e. a tendency to sell low). They have no concept of value being different from price, only a number that mysteriously moves up and down and shows 20% gains in each of the last couple of years.

What I am still puzzled about is whether index rebalancing has any multiplicative or reinforcing effect on 'price momentum' or even a countervailing effect on it at time when there is no net inflow or outflow as various indexing investors add funds and withdraw funds over a period.

After that, sure if there are net inflows it will tend to boost each company in proportion to their weighting.

I believe that the S&P500 index figure represents a fixed fraction of the market cap (or free-float adjusted market cap) of the companies involved.
The S&P500's total market cap at 31st Dec 2017 was $23,938,148.8 mn, float-adjusted: $22,900,164.8 mn, Index value = 2673.61

On that date, for example, AAPL was priced at $169.23 and had Mkt Cap of $858,675.6 mn (approx - I've assumed no change in share count in the last couple of months, but that doesn't change the gist of what I'm working out).

AAPL should make up 858675.6/23938148.8 of the index = 3.587% by Market Cap on 31st Dec 2017 and would still represent 3.587% of the S&P500 index value of 2673.61, meaning 95.90 index points. I we assume 1 'point' is worth $1, than means for every purchase of 1 unit of S&P500 at $2,673.61, $95.90 of that was APPL, so the number of shares of Apple purchased was $95.90/$169.23 = 0.5667 shares of AAPL.

With 5,074 mn shares in issue, that's 1/8,563,000,000 ths of the shares outstanding in AAPL represented in the index.

It should be that every other firm in the index also has 1/8,563,000,000 ths of its market cap (or perhaps that fraction of its free float market cap) represented at present. So a firm XYZ Corp valued at exactly 1/100th of AAPL's market cap on 31st December would represent 0.03587% of the S&P on that date or $0.9590.

If AAPL happened to do a 2-for-1 stock split on 1st Jan 2018, it would represent 1.1334 shares, still worth $95.90 (as the AAPL price would be $84.615) - no change in index weighting.

If it didn't split, and rose to $181.72 (close on 12th March 2018) while the index rose to 2783.02, that 0.5667 shares would be worth $102.98 out of every $2,783.02 unit of index fund (at 12th March). This is now 3.700% of the index, but didn't involve index funds buying more shares in AAPL, it just reflects its rise in stock price having increased faster than the rest of the index increased.

If rebalancing were carried out today, it would only be a reflection of changes in the number of shares in issue.

For example if AAPL were to buy back and retire 10% of its stock this quarter, effectively the index funds would have to sell 10% of their AAPL holdings to rebalance exactly. This is extreme, and no company is likely to buy back that much - maybe 2-4% in a quarter on rare occasions, and index funds could take their time rebalancing and accept some tracking error.

So having got that straight, what happens when net inflows into index funds are occurring?

Today, for every net $2,783.02 coming in, the fund will be buying typically 0.5667 shares of AAPL give or take some tracking error. This is 1/8,563,000,000 ths of its market cap.

Likewise, XYZ Corp shares would be bought at 1/8,563,000,000 ths of its market cap (i.e. they'd buy 1/8,563,000,000 ths of the shares outstanding).

Relatively, the amount of buying demand on both index constituents is the same proportion of its shares outstanding. If APPL happened to fall 10% relative to the S&P500 (still at 2783.02), and the cash inflows for the index funds were the same, they'd still buy typically 0.5667 shares of AAPL for every net $2,783.02 coming in. If nobody repurchased their own shares, it would still be 0.5667 shares.

Likewise is XYZ Corp fell 50% the S&P500 would barely budge, yet it would still have index funds buying the same 1/8,563,000,000 ths of the shares outstanding for every net inflow of $2,783.02 into these funds.

If the S&P500 fell consistently or very sharply for a few months, especially with a serious geopolitical or economic event as a 'reason', what I would imagine is that most index funds would see net outflows of capital and would then switch to being on the selling side. For every $2,783.02 of net outflow today, they'd have to sell 1/8,563,000,000 ths of the shares outstanding (or free float) in every stock in the index, give or take tracking error, meaning 0.5667 shares of AAPL worth $102.98 and almost a dollar's worth of XYZ Corp etc.

Now, I imagine the know-nothing retail investors herding into and out of index funds based on emotions will also be accompanied by retail investors herding into and out of active managed funds too (with the exception of those few Value Funds that successfully manage to discourage this adverse behaviour among their partners, either by persuasion or by penalties for withdrawals without sufficient notice).

It's likely that the active funds will also have to sell many of their positions regardless in order to fund the net redemptions, though they might be selective and strategic about which positions they sell in ways that index funds will not.

I can certainly see how the net flows of capital will shift the balance of supply and demand and the herding behaviour would, for a time, reinforce the price action that caused the herding - a positive feedback loop (positive feedback loops in 'control theory' being unstable, causing overshoots and wild swings, whereas negative feedback loops tend to cause stable more gradual response to a sudden stimulus). Negative feedback loops are more 'positive' emotionally, while positive feedback loops can often produce emotionally 'negative' outcomes.

I think it has always been this way.

It seems you need to force out most of the emotional actors from the market before only the more rational actors are left and the self-reinforcement over downward 'momentum' can correct. The longer the boom, the more irrational actors are drawn in and the higher the market will peak before it busts, and the deeper the bust will go before they are driven out.

Buffett's words may encourage many more people to buy index funds when the 'going is good' than the number that will be persuaded by his words to have the emotional detachment to stick to investing regularly even when the market has been in decline and looked scary. Only the latter group will reap the full rewards of investing in the wide range of businesses represented by the index. Those who bail out when fear abounds will tend to capture most of the falls and miss most of the rises.

What I'm not seeing is how a rise in indexing is really any different to any rise in retail fund investing (e.g. mostly actively managed in the past booms). In the past, the range of active funds was wide, and although there was herding, there were many popular approaches including momentum-based and sector-focused that paid relatively scant attention to intrinsic value, especially as retail investors piled in towards the peak of a boom. In aggregate, I think the net inflow of funds still caused increased demand in almost all stocks causing prices to tend to rise, and when there was a net outflow of funds, that caused increased supply of almost all stocks, causing prices to fall.

I'm thinking that it could be the active funds where they wish to advertise that you're taking part in the 'performance' of sexy well-known stocks like Apple, Facebook, Amazon, Google and the likes (and even Berkshire!) that may be over-weighting these  stocks to increase their appeal to market their funds to bright-eyed retail investors who want a piece of that recent 'performance' as "it's sure to continue in future" in their minds. It can be subtle things like that which will sway them into picking specific funds (and the fund's market departments know it), and perhaps its that which would drive the relatively higher demand for these market darlings who have recently 'performed' so 'well' (or as we'd put it, whose price has become less attractive in relation to their intrinsic value).

Equally, if you're trying to manage large funds actively, if a lot of additional money is flowing into your fund, it really forces you to look at investing in those stocks with the biggest market caps so you don't drastically distort the supply and demand of the smaller names by making up a great proportion of their daily volume.

I admit I'm struggling to see how it's indexing rather than just the general flow of capital into funds of all kinds, that is driving the concentration of gains into a narrow range of large-caps.

Assuming most stocks of all sizes have similar percentage turnover of their shares in issue during a year (notable exceptions like Berkshire being the rare counter-example), and ignoring companies newly entering or leaving the index (especially large-cap entrances like BRK.B a few years ago), I cannot see index funds being responsible for the momentum multiplying effect causing the largest caps to experience the largest gains. To my mind, it seems more plausible that our culprit is the majority of non-value active managers possibly aiming to attract the most Assets Under Management as higher priority than long-term performance, that are most likely to focus on the 'big names' and large caps as all this new money comes in as we near the peak of the boom. To me it seems like that's the more likely mechanism for this concentration that precedes so many of these crashes.

But I'm willing to be persuaded, and would be glad to be shown if I'm wrong in any of my assumptions. I try to remain a true skeptic - willing to change my beliefs on the basis of good evidence.
« Last Edit: March 14, 2018, 06:55:39 AM by Dynamic »


  • Hero Member
  • *****
  • Posts: 637
Re: Semper Augustus letter
« Reply #63 on: Today at 06:40:57 AM »
"I admit I'm struggling to see how it's indexing rather than just the general flow of capital into funds of all kinds, that is driving the concentration of gains into a narrow range of large-caps."
I actually agree with that statement.
My assessment is that mostly the shift to index funds can be explained by the move away from mutual funds and even if index funds are passively managed, the underlying investor population, in essence, has a passive mindset.
I submit though that there may be pockets of ETFs where this does not apply: specialized, leveraged and synthetic ETFs. These funds may attract a momentum crowd and liquidity issues with precipitated attempts at price discovery have not been tested (remember how that worked out with packaged real estate subprimes securities).
IMO the infatuation with indexing is simply part of the larger picture and is based on momentum (may work in both directions as markets don't usually follow a straight line).
Isaac Newton would have said: "what goes up must come down" but markets tend to go up and his investment record is not impeccable.

I thought you would be the type to be interested in the following:

If pressed for time,
-the first link shows a nice graph (exhibit 4, page 6).
-the second link refers to well done specific studies evaluating the relevant underlying questions.

Apparently, according to Mr. Bogle, the father of indexing, as long as 25% of funds are actively managed, we're probably OK.
I wonder if that number has a margin of safety.