The S&P 500 Index is so familiar that we assume we understand it. We think of it as a simple metric, neutral and benign. Like any Index (so we assume) it tracks the changes in the values of a selection of assets, trading freely in the market. It measures, but does not create or distort, those values.Â
This view is wrong. Indexes are no longer â€œjust indexes.â€ They increasingly determine the flows of investment funds around the world. The explicit role that index providers play in shaping these flows has come to light in a number of contexts, such as the long struggle waged by Chinese authorities to win the approval of index-provider MSCI to include Chinese A-shares (Chinaâ€™s domestic stock market) in MSCIâ€™s Emerging Markets Index, a powerful global benchmark. Index-tracking funds, of course, have to buy whatever assets the index providers decide to include in the index (whether or not they might otherwise want to own them). These examples illustrate the power of what we might call the allocation effect that certain indexes have come to exercise over the market. This power is overt, obvious, and well-recognized. The composition of important indexes like the S&P 500 directly determines the allocation of trillions of investment dollars today, and indirectly influences trillions more. Â
Less well understood are the more â€œorganicâ€ effects of indexing â€” the way indexes are designed, adjusted, â€œrebalancedâ€ and â€œreconstituted.â€ These go beyond the allocation effect to shape trading patterns and pricing movements â€“ and yes, sometimes to alter the values of the assets and the markets they reference. These effects are becoming more powerful. With the rise of passive index-based investing â€“Â now by some measures encompassing half of the investment universe â€“Â the indexersâ€™ influence extends deep into the market. When the index changes, money moves.Â
Are there â€œfeaturesâ€ (â€œflawsâ€) in this indexing model that could give rise to market disruptions and even create systemic financial risk?Â
The Tesla Episode
On November 16, S&P announced that Tesla would be added to the S&P 500 â€“ the largest company by far ever to join the worldâ€™s most important equity benchmark.Â
It became, in effect, a â€œstress testâ€ of the S&P 500.
MORE FOR YOU
The change took effect, as scheduled, before the market opened on Monday December 21.
On November 16, the company was worth $408 Bn. In the five weeks following the announcement, Teslaâ€™s share price rose 70%, adding $272 Bn to the companyâ€™s market value â€“Â including a spike of $40 Bn on Friday Dec 18 alone (the last trading day before the change). These short-term gains surpassed the entire market capitalization of every other automaker in the world. Â
Indeed, Tesla was suddenly worth more than the next 9 leading incumbents combined.
This market frenzy will go down as a â€œsign of the timesâ€ â€“Â but just what does it signify? Â
It is obvious that this surge was connected, somehow, with the the prospect of Teslaâ€™s admission to the S&P 500. But what was cause, and what was effect? Â
The news media have mostly made it a story about Tesla, as a company and a business. This is understandable. First of all, they tend to accept the view that the S&P 500 is a just a neutral metric. And Tesla is an incredibly exciting story, indeed a raft of stories: a celebrity CEO, a visionary business model, a deeply flawed business model, fantastic technologies, ridiculous valuations, politics (climate change), politics (taxpayer-financed, â€œmarket-distortingâ€ electric vehicle subsidies), politics (personal â€“ will Elon Musk leave Cali for Texas?)â€¦ space travel, exploding rocketsâ€¦ colonies on Marsâ€¦Â Â
It is understandable â€“ but wrong. Teslaâ€™s price surge between Nov 16 and Dec 21 was not actually about Tesla as a business. The Tesla bubble is a passing phenomenon, one way or another. The real story here is much more important, indeed fundamental, and its significance is enduring. It focuses on the distinctly unexciting topic of index construction â€“ and specifically the technical features (and perhaps flaws) of the worldâ€™s most important stock market index: the S&P 500.Â And how those features (and flaws) may have engendered and exacerbated the frenzy. How they actually could create or distort the values they are intended to track.Â
To understand what happened with Tesla, we have to start with some basics â€“ how the S&P 500 index is constructed. That is the focus of this article. In Part 2 (next), we will look at how the Tesla â€œstress testâ€ unfolded, and how market distortions can arise from certain aspects of the indexing structure.
The S&P 500: What Is It Really?Â
Everyone assumes they know what the S&P 500 is. But few really do.Â
What It Is Not
I used to tell students that the S&P 500 was a market-cap-weighted average of the 500 largest public companies in the U.S.Â
I was wrong in every respect.Â
- The S&P 500 is not an average
- It does not include the 500 largest public companies â€“Â there are hundreds of large-cap â€œexceptionsâ€ (companies with more than $10 Bn in market value each) that are not part of the Index
- It is not cap-weightedÂ
- It does not even have 500 components
So, let us begin with a close look at what the S&P 500 actually is.Â
Some Assembly Required
Isnâ€™t the S&P 500 based on the list of the 500 largest public companies in the U.S., ranked by market value?Â
Well, no â€” not exactly. In fact, not at all. There are â€œcriteriaâ€ to consider, conditions to be met (not all of which are well publicized). One such condition which has been mentioned prominently in connection with Teslaâ€™s potential membership is the requirement for 4 consecutive profitable quarters. Tesla achieved this in the 2nd quarter of 2020, as reported in July. The company fully expected to be admitted to the index â€” but was passed over. (On that news, Tesla shares dropped 21% in a single day â€“ a sign of things to come.)
Adding a company to the index is not automatic. There are many large companies that are not in the index. Zoom Video Communications, Inc. (to name one you may have heard of lately) has had 6 profitable quarters, with a market cap ofÂ over $100 Bn. It is not a member. Â
A less obvious criterion is sector representation. According to S&P Global:
- â€œOne way we can grade the S&P 500â€™s performance is through sector representation. Since the index has a fixed count of 500 companies, not every eligible company can be added to the index. One of the factors the Index Committee looks at when considering changes to the index is sector composition.â€
In other words, it is not just size (market cap) or profitability that count. It also matters what the company does. Or even its â€œbalance with respect to the market.â€ (Whatever that means.)Â
This underscores the judgment factor which permeates the indexing process. Decisions about S&P 500 membership are not strictly rule-driven. Nor are they entirely transparent. Changes to the index are made by a â€œsecretive committeeâ€ (membersâ€™ identities are not publicly disclosed, except for the chairman). Their approach is flexible. The former chairman has written â€œa rigid rule book wonâ€™t work.â€ In the aftermath of Teslaâ€™s failure to make the index in September, a baffled Bloomberg Business reporter wrote that â€œfew organizations in finance are as mysterious as the committee that oversees the S&P 500.â€
Even the common assumption that market capitalization is the chief criterion for inclusion is incorrect. The S&P 500 company that was bumped off the index to make room for Tesla â€“ Apartment Investment Management Co. (Aimco) â€“ has a market cap of about $800 million. There are over 200 public companies in the U.S with market capitalizations at least 10 times as large, which are not part of the S&P 500 index.
And by the way, as a detail â€“ the index actually includes 505 stock listings, not 500.Â
- â€œThe S&P 500 does include 500 companies. But those companies are represented by 505 securities due to multiple share classes of Google parent Alphabet, Discovery, News Corp, 21st Century Fox, and Under Armour.â€
Well, isnâ€™t the weight of each company in the index proportional to its market capitalization, so we can at least describe the S&P 500 as a cap-weighted index?Â
No, not quite â€“Â the S&P 500 is actually a float-weighted index. The weighting of each company only takes account of the value of the shares that are in the â€œpublic floatâ€ â€“ shares that are freely tradable by ordinary investors. This can exclude a big slice of the value of some companies. There are at least 10 categories of shareholdings that are omitted from the calculation of the index weight (the float weight), including shares held by officers and directors, venture capitalists, â€œstrategic investors,â€ shares owned by other public companies, and sovereign wealth funds. Many types of equity instruments are also excluded, such as restricted shares (unregistered shares), treasury stock, stock options, warrants, preferred stock, convertible stock and stock rights â€“ all excluded from the â€œpublic float.â€Â
S&P implemented float-weighting in 2004. They are a bit cagey about the rationale. On the one hand, they have argued that float-weighting lowers costs (how?) and â€œimproves investability.â€ On the other, they claim that â€œthe difference is very small, barely visible.â€ The former chairman of S&Pâ€™s Index committee has acknowledged that this â€œdifference-that-makes-no-differenceâ€ is somewhat paradoxical. â€œDoes all this mean that float adjustment was unnecessary?â€ His answer is No, apparently because the impact of the change is so very small.
- â€œThe difference in index levels [this was written in July 2013] is 13 index points or 77 basis points.Â The weight of the ten largest stocks in the float adjusted S&P 500 is 18.09%, slightly less than the ten largest stocks in the non-adjusted index which were 18.14%.â€
The float factor of the overall market is 96-98% â€“ so the discount is only a couple percent (which I guess is â€œbarely visibleâ€).Â But for certain stocks, the difference is substantial. The float factor for Facebook is 84%, for Amazon 85%, GoogleÂ 90%.Â The Tesla Float factor is 80.1%. Â
One could argue that the S&P methodology undervalues Tesla as a component of the index by 20%. Not trivial.Â
The â€œMarket Averageâ€ That Isnâ€™t
Ok â€” so, float-adjusted â€“ but then isnâ€™t the index simply an average of the (float-adjusted) market values of those 500 companies? Like the Dow Jones Industrial â€œAverageâ€? Â
No. The Dow Jones Industrial Average is not an average, and neither is the S&P 500.Â
Here is how the S&P 500 index weights are calculated:Â
Unpacking this formula is a lengthy undertaking, documented extensively by S&P Global. There are several pages devoted just to the calculation of the â€œinvestable weight factorâ€ â€“ IWFj (â€œthe float factorâ€).
Then there is the â€œdivisor.â€
The formulaâ€™s mysteriously-labeled denominator is critical. It exists to keep the index stable in the face of â€œcorporate actionsâ€ â€“ like share buybacks, stock dividends, rights offerings, spin-offsâ€¦ that affect the share count, and hence the index weight, and to smooth over changes in the index composition. The goal is to avoid â€œgapsâ€ in the index when it is either â€œrebalanced,â€ or â€œreconstituted.â€Â
Rebalancing is a regular housekeeping measure.
- â€œDuring a rebalance, S&P 500 constituentsâ€™ individual weights are adjusted to reflect their latest share counts and float. Company share counts are constantly changing as they issue stock and perform buybacks, so the S&P 500 is rebalanced every quarter to adjust each companyâ€™s weighting based on its latest share count and float.â€Â
In other words, the â€œfloat factorâ€ or IWF is adjusted for every company once a quarter.
â€œReconstitutionâ€ happens when companies are added or dropped from the index, which may occur at any time:
- â€œWhile the Index Committee can also reconstitute the index during a rebalance by adding or removing companies, these changes to index membership can be made at any timeâ€”they donâ€™t have to happen during a rebalance and a rebalance doesnâ€™t have to include a reconstitution.â€Â
The divisor is critical to the indexâ€™s stability. Tesla is 100 times more valuable than Aimco. When Tesla replaced Aimco in the index, the total value of the S&P 500 index members would have â€œgapped upâ€ significantly, if the index were a simple average. The divisor is the little trick that prevents this gapping from taking place. (The Dow uses a divisor as well, for the same reason.)Â The divisor is what makes the S&P 500 something far more than an average of cap-weighted components.Â
(In any case, if you ever wondered â€“ as I once did, naively â€“ why anyone with access to a Bloomberg terminal couldnâ€™t construct their own index like the S&P 500â€¦ Now you have some idea.) Â
Pay No Attention to that Index Behind the CurtainÂ
Normally, you arenâ€™t supposed to be thinking about any of this, of course. Not that it is secret â€“Â S&P Global publishes its methodology in very clear languageÂ â€“ but it is not really intended for the general audience. The ethos of the S&P 500 index is simplicity, and the absenceÂ of drama. Like a thermometer. Your iPhone tells you it is 42Â° F outside and you take it for granted, and put on a warmer coat. You donâ€™t think much about how it was calculated, or when, or whether the thermometer was accurate, the humidity, the solar index,â€¦ itâ€™s just plain 42Â°F. The S&P 500 projects, officially, the same bland, untroubled aspect. A simple metric, that gives us the value of â€œthe market.â€Â It has to be seen as stable and accurate.
[It is an awesome responsibility. Weâ€™ve seen what happens when a metric for an important financial standard is seen to be unstable or inaccurate. Libor (the London Interbank Offered Rate) â€“ once a global standard for short-term interest rates â€“Â was crippled by scandal following the 2008 financial crisis. It was found that the method for calculating Libor was being manipulated, falsified, and as a result it will lose its pre-eminent role as a benchmark for trillions of dollars of derivatives and other financial instruments. Replacing Libor is proving to be very disruptive and very costly. If anyone came ever to feel that the S&P 500 was not an accurate measure of the market, the consequences would be catastrophicâ€¦]
The weather is changeable, but we count on the thermometer.Â The S&P 500 is well-managed (thatâ€™s what S&P gets paid, very handsomely, by its licensees to do). Its weaknesses (if that is what they are) are well-hidden. And heaven forbid that anyone would ever imagine the index itself might actually cause disruptive market movementsâ€¦ any more than that the thermometer could cause the temperature to rise.
Back to Tesla
Tesla sideswiped this applecart.Â
The immediate problem was of course Teslaâ€™s scale. At $650 Bn by the time it was added, it was by far the largest company by market value to ever join the index (3 times bigger than Berkshire Hathaway when it joined the index in 2010, and almost 6 times the market value of Google when it joined in 2006). Teslaâ€™s value was more than twice the combined value of all the other 15 companies added to the S&P 500 this year. Tesla is worth about as much as the entire energy sector, Exxon, Chevron and the whole club. Tesla accounted for about 1.7% of the index value (even with its 20% discount for float-weighting). It weighed in at about 13% of its sector â€“ Consumer Discretionary.
Tesla was so big that the normal small-scale effects that occur when a new company joins an index were magnified to an unprecedented degree. It exposed the indexing machinery clanking and grinding away behind the curtain. The index came under stress, greater than it has perhaps ever had to endure. It is difficult to claim that it was possible to maintain â€œbalance with respect to the market.â€ Certain features (flaws) of its operation and especially of its interaction with the market became suddenly quite a bit more visible.Â
To understand how the seemingly simple and well-practiced addition of one new company to the other 499 has broken the faÃ§ade of market-neutrality that the S&P 500 has labored to create, in the next Part we will look to three aspects of this extraordinary episode:
- the effects of Tesla upon the S&P 500 â€“Â in particular, the transmission of the companyâ€™s obvious over-valuation, excess volatility, and increased skew to arguably distort the index Â
- the effects of the Index in transmitting and amplifying Teslaâ€™s troubling characteristics and projecting them onto a much larger segment of the financial system
- problems arising from the structure and operation of the Index itself, which may create systemic risks for the proper functioning the financial marketsÂ