EconPlot › Explainers › How concentrated an S&P 500 fund really is

Is an S&P 500 fund really 500 even bets?

Updated 2026-09-06 · how we source every figure

No — one holding is 8.23 percent, and the ten largest are 37.955 percent

The bets are wildly unequal in size. The SPDR S&P 500 ETF Trust is an exchange-traded fund, run by State Street Global Advisors, that holds the Standard & Poor's 500 in full, so the holdings file it publishes is the index's own list of weights. As of 2 September 2026, read on 3 September 2026, the ten largest positions were:

• NVIDIA, 8.23 percent
• Apple, 7.22 percent
• Microsoft, 5.59 percent
• Amazon, 3.78 percent
• Alphabet Class A shares, 2.99 percent
• Broadcom, 2.63 percent
• Alphabet Class C shares, 2.39 percent
• Meta, 1.97 percent
• Micron, 1.63 percent
• Tesla, 1.52 percent

Those ten came to 37.955 percent of the fund on that date.

A weight is simply the share of the fund's money sitting in one company. NVIDIA's 8.23 percent means 8.23 percent of every dollar in the fund was tracking that single company, against 1.52 percent tracking Tesla — the same fund, the same day, more than five times the exposure to one name as to the other.

The file also has 505 lines rather than 500, because a few companies are in the index with more than one class of stock. Alphabet is the visible case above: its Class A and Class C shares appear as two separate holdings. Since the file is sorted by size, everything not shown above is smaller than the tenth entry, so each of the remaining 495 lines (505 lines minus the 10 listed) was under 1.52 percent of the fund.

The slices are unequal because the index weights by market value, not by company

An S&P 500 fund does not divide money between companies; it divides money in proportion to what the companies are worth on the stock market. S&P Dow Jones Indices — Standard & Poor's Dow Jones Indices, the company that maintains the index — sets each member's weight by float-adjusted market capitalisation: the share price multiplied by the number of shares actually available for public investors to buy, leaving out large blocks that are not freely traded. That rule is written down in the published U.S. Indices Methodology of S&P Dow Jones Indices.

Follow the consequence in steps.

First, a company worth ten times as much on the stock market as another gets ten times the weight, no matter how many people it employs, how much it sells, or which companies it is standing next to in the list. Stock-market size is the only input.

Second, the fund copies that weighting rule. The SPDR S&P 500 ETF Trust replicates the index in full, so the weights the fund publishes are the index's weights, not a manager's opinion about which companies deserve more money.

Third — and this is the part that is easy to miss — the weights move on their own. Because a weight is defined by market value, a company whose share price rises faster than the rest of the index takes up a larger share of the index the next day without anyone buying anything. No trade is required and no committee decides. Concentration can build up purely as a by-product of share prices moving, which is also why any statement about how concentrated the index is has to carry a date.

So the honest description of an S&P 500 fund is not 500 even bets, and never was. It is roughly 500 companies held in proportion to their stock-market value, which on 2 September 2026 meant a range running from 8.23 percent at the top down to fractions of a percent at the bottom.

Twenty-three of the 505 holdings are half the money

Sorting the 2 September 2026 weights from largest to smallest and adding them up as you go, the running total first passes half of the fund at the twenty-third holding. Those weights are from the same holdings file the top-ten figures come from, and anyone with that file can repeat the addition.

What that means for a person holding the fund: half of every dollar is tracking twenty-three companies, and the other half is spread across all the rest of the list. The first ten of those twenty-three were already 37.955 percent of the fund, so only a handful of further names carries the running total across the halfway mark.

The comparison that makes this concrete is what an evenly split fund would look like. If the money were divided equally across the file, reaching half the fund would take half the names on it — hundreds of holdings, since the file has 505 lines. Instead it takes twenty-three. That gap is the entire effect of weighting by market value.

None of this is a defect in how the fund is run. It is what the fund promises to do: track the index, including the index's shape.

The law's diversification test draws its line at 5 percent per holding, and three holdings are over it

United States fund law defines the word 'diversified', and the definition is a numerical test that a holding of 8.23 percent fails. The rule sits in the Investment Company Act of 1940, the federal statute that governs mutual funds and exchange-traded funds. To be a 'diversified company' under that statute, at least 75 percent of a fund's total asset value must sit in cash, government securities, shares of other investment companies, and other securities limited to no more than 5 percent of total assets in any one issuer — the company whose shares the fund is holding — and no more than 10 percent of that issuer's voting shares.

Apply the 5 percent half of that test to the holdings file as of 2 September 2026. NVIDIA at 8.23 percent, Apple at 7.22 percent and Microsoft at 5.59 percent were each above the line. Those three together were 21.04 percent of the fund, and none of that 21.04 percent may be counted toward the 75 percent the statute asks for. In the plainest terms available: the three biggest bets in the fund are, in law, too big to count as diversification.

Now the other side of the same arithmetic, which is the part a headline usually drops. Everything else — 78.91 percent of the fund — sat in positions of 5 percent or less each. The test asks for 75 percent. Reading the published file line by line, the index clears the requirement by 3.91 percentage points (78.91 minus 75). So on that reading the S&P 500 still satisfies the diversification requirement. The index satisfies the requirement while its three largest holdings are individually disqualified from helping, which is a stranger fact than either 'it passes' or 'it fails' on its own.

Why a fund can hold an 8.23 percent position and still be called diversified

Because the fund's remaining money is small-position money: 78.91 percent of it sat in holdings of 5 percent or less on 2 September 2026, against the 75 percent the Investment Company Act of 1940 requires. One thing is therefore keeping the label in place today, and a second stands behind it. The arithmetic is what holds now — the 3.91-percentage-point margin above, measured on that day's file.

The second thing matters only if that margin goes: the regulator has said it will not enforce the limit in this situation. On 24 June 2019 the staff of the United States Securities and Exchange Commission, the federal agency that regulates funds, issued a no-action letter in response to a request from the law firm Stradley Ronon Stevens & Young LLP on behalf of index fund sponsors. The staff said it would not recommend enforcement action against an index-based fund that exceeds the diversification limits 'to the extent necessary to approximate the composition of the fund's target broad-based index'. Two conditions attach: the fund must update its registration statement to state that it 'may become non-diversified ... solely as a result of a change in relative market capitalization or index weighting', and it must notify shareholders.

The staff's stated reason is the same phenomenon this page has been measuring. Certain constituents of large-capitalization United States equity growth broad-based indices, the letter says, 'had grown to represent more than 5% of such indices and, in some cases, in the aggregate, more than 25%'.

It is worth being precise about what that letter is. It is a staff position on whether the staff would recommend enforcement — not an amendment to the statute, not a court ruling, and not a finding that any particular fund is diversified. The statute's 5 percent line is untouched. What changed in 2019 is that a fund which drifts over the line by faithfully tracking its index was told it would not be pursued for doing so, provided the fund states that it may become non-diversified in exactly that way.

What this page does not establish

• Every weight here is one day's snapshot, 2 September 2026, and weights move with share prices without anyone trading. The 3.91-percentage-point margin over the statutory 75 percent is measured on that single day and moves as those prices move; this page holds no evidence about how far or how fast it has moved before. A reader who treats 78.91 percent as a standing property of the index rather than one Wednesday's measurement will be wrong at some point without anything having gone wrong.
• The figures are one fund's published holdings — the SPDR S&P 500 ETF Trust, which replicates the index in full. Other funds tracking the same index hold the same companies but publish their own files, and differences in cash, timing and share-class lines mean their weights will not match these to the decimal. The concentration story carries across; the exact numbers do not.
• The statutory 5 percent test counts issuers, and the published file counts lines. Alphabet's two share classes appear separately at 2.99 and 2.39 percent; added together that is 5.38 percent, above the line, even though neither line alone is. This matters for the margin on this page: if those two lines are combined at the issuer level, Alphabet's money would also fall outside the qualifying pool, leaving about 73.53 percent (78.91 minus 2.99 minus 2.39) — under the 75 percent threshold rather than 3.91 points over it. The 78.91 percent and the margin quoted here are line-level figures, and whether the test is applied to Alphabet's classes separately or together is a legal question a holdings file cannot answer.
• Passing the statute's test is not a statement that risk is spread out. The test only asks how big each position is; it says nothing about the companies being in the same industry, selling to the same customers, or having share prices that rise and fall at the same time. A fund could hold two dozen separate positions of under 5 percent each in companies from a single industry and satisfy the test completely, which is why 'diversified' on fund paperwork is a legal label rather than a description of what the holdings do together in a downturn.
• The June 2019 letter is a staff no-action position, not a change to the statute. It states what the Securities and Exchange Commission's staff would not recommend, it is conditional on the fund disclosing that it may become non-diversified, and staff positions can be narrowed or withdrawn. Reading it as 'index funds are exempt from the diversification rule' overstates the letter's reach and ignores its conditions.

Questions people also ask

Sources

The evidence behind every claim above, as recorded when the piece was made.

index fund holds about 500

The S&P 500 holds 500 constituent companies but slightly more share classes; the SPY holdings file read on 2026-09-03 lists 505 lines, because companies such as Alphabet appear as two classes. The script says 'about five hundred'.

market value weighting

S&P 500 constituents are weighted by float-adjusted market capitalisation (S&P Dow Jones Indices, 'S&P U.S. Indices Methodology'), which is why the slices are unequal; SPY replicates the index in full, so its published weights are the index's weights.

top ten 38 percent

SPDR S&P 500 ETF Trust published daily holdings, as of 02-Sep-2026, retrieved 2026-09-03 by scripts/fetch_index_weights.py into data/filings/spy-weights.json: the ten largest weights sum to 37.955 percent, narrated and shown as about 38 percent. The ten were NVIDIA 8.23, Apple 7.22, Microsoft 5.59, Amazon 3.78, Alphabet A 2.99, Broadcom 2.63, Alphabet C 2.39, Meta 1.97, Micron 1.63, Tesla 1.52.

twenty three names are half

Same file: running the weights in descending order, the cumulative total first crosses 50 percent at the 23rd holding. Recorded as names_for_half_the_index in data/filings/spy-weights.json.

date anchored

Concentration is only true on a date, so the holdings date (2 September 2026) is stated on screen and in the narration's 'early September 2026'.

editorial

Explains, does not advise: no view on whether concentration is good or bad, no suggestion to buy, sell, diversify or switch funds, and no named holding is presented as an opportunity.

house accent

Accent set to the channel yellow #f2b705 used by the rest of the queue (was #2f6fed); the accent colours the chart line, the badge and the stamp, so an odd one out reads as a different channel while scrolling.

hook rewrite

Both reviewers failed hook-impact 2026-09-03: 'you own a slice of about five hundred companies' is the ordinary premise, not the shock, and 38 percent was deferred to scene 4. The hook now carries the stake and the figure in one sentence. DeepSeek also flagged scene 3 as a near-verbatim restatement of its own on-screen highlight, so the narration was rewritten to say something the card does not.

doccard retired

Owner, 2026-09-03: 'why we need doc card? when we already have another card like this'. The doccard scene type rendered a headline on a bare paper card with no artwork at all, which is newscard minus the picture; it was the only text-only type in the renderer and every project had unused card art sitting beside it. DocCard was deleted from remotion/src/MarketShort.tsx and these scenes are newscard now. NewsCard took over its two extras: the kicker renders above the headline, and a highlight that is a whole sentence not lifted from the headline renders as a caption bar over the artwork instead of painting headline words yellow. Not one word of narration or on-screen text changed, so the script fingerprint and its validation still stand.

deepseek undelivered derivation

DeepSeek review 2026-09-03, a finding Claude passed. The close was 'So half of your five hundred company fund is twenty three companies.' The 'so' claimed to follow from the body, but the body establishes only that the ten biggest are about 38 percent; nothing in the video gets the viewer from ten-at-38-percent to twenty-three-at-half. This is the same defect the owner caught in the recession script the same day, in a different video. The close now walks the one step it was skipping: ten is 38 percent, twenty-three is half. Both figures come from the SPY holdings file that scripts/fetch_index_weights.py records (it stores the top-5/10/25 share and how many names make half the index), so no new claim enters — only the connection between two facts the video already owns.

parallel members in the close

Both reviewers, 2026-09-04, on the line written earlier the same day to fix the missing derivation: 'Ten of them are thirty-eight percent. It takes just twenty-three to reach half your fund.' The derivation was right but the two members were built differently — [count] are [share] against [it takes count] to reach [share] — so the comparison the whole video exists to make had to be re-parsed. Both members are now [count] of them are [share]. The on-screen pair had the same fault and is fixed with it: scene 4 reads 'The ten biggest are 38% of the index' and scene 5 now 'Twenty-three of them are half the index', where it previously inverted to 'Half of it is twenty-three companies'. No figure changed; both still come from the SPY holdings file.

sentence caption removed

Owner, 2026-09-05, on the staged drafts: 'the label ... should be removed. Way too many words and too little time to read ... they look like a caption, and is redundant with the caption almost. can we just please remove this type of text.' The caption bar was added on 2026-09-03 so the retired doccard's second sentence had somewhere to live; that was the wrong instinct. A short already burns in word-by-word captions, so a second block of prose over the artwork competes with the text the viewer is reading, in a scene lasting a few seconds. Removed from the renderer and from these scenes: scene 2: Bigger company, bigger slice of your money.; scene 4: Not five hundred. Twenty-three.

diversification test 1940 act

Owner, 2026-09-05: 'I remember somewhere sending a notification that says sp500 index no longer satisfy the diversification requirement set by a law maybe in 1970s or something. maybe we can cover that.' Verified and corrected in one respect - the law is the Investment Company Act of 1940, section 5(b)(1), not the 1970s, and the index has NOT stopped satisfying it. The statutory '75-5-10' test: to be a 'diversified company' at least 75 percent of total asset value must be in cash, government securities, other investment companies and other securities limited to no more than 5 percent of total assets per issuer and no more than 10 percent of any issuer's voting securities. Measured on the holdings file we already hold, data/filings/spy-weights.json, SPY as published by State Street as of 02-Sep-2026, 505 holdings: NVDA 8.2269, AAPL 7.2241 and MSFT 5.5860 percent each exceed 5 percent, together 21.04 percent; the remainder, 78.91 percent, sits in positions of 5 percent or less against the 75 percent the test requires - a margin of 3.91 points, which the narration rounds to 'four points'. The narration says the index PASSES, because it does. Anyone can recompute this from the same file by summing the weights at or below 5 percent.

sec no action 2019

SEC staff no-action letter dated 24 June 2019, issued in response to a request from Stradley Ronon Stevens & Young LLP on behalf of index fund sponsors. The staff said it would not recommend enforcement action against an index-based fund that exceeds the diversification limits 'to the extent necessary to approximate the composition of the fund's target broad-based index', provided the fund updates its registration statement to state that it 'may become non-diversified ... solely as a result of a change in relative market capitalization or index weighting' and notifies shareholders. The SEC's stated reason was that certain constituents of large-capitalization US equity growth broad-based indices 'had grown to represent more than 5% of such indices and, in some cases, in the aggregate, more than 25%'. The narration says only what the letter says: the regulator would not act if the index pushed a fund past the line. It does not claim any particular fund became non-diversified.

length extended from 21s

The video ran 21.0 seconds after the on-screen text cuts, under the 25-second floor, and it ended on a fact with no consequence ('twenty-three of them are half'). Two scenes were added rather than padding the existing ones. Number budget: years do not count against the three-number cap (scripts/audit_numbers.py), so 2019 is free and the additions spend only 'three quarters' and 'five percent'. Scene 3's balance-scale artwork was also replaced under validations/scene-artwork.md - the sentence is about unequal slices, so it now shows a pie cut into wildly unequal wedges. The closing line first read 'The S&P 500 passes by four points', which audit_numbers caught: '500' is a figure and scene 2 already narrates it, so the name of the index cannot be spoken twice. It says 'The index' instead. Three findings from the first pass on the extended script were fixed rather than accepted. LENGTH: the old closing line repeated 'thirty-eight percent', already narrated in scene 2 and shown three times on screen, so it now says only 'Twenty-three of them are half.' PUNCHLINE: the video ended on the reassuring no-action clause, which softened its own reversal; the measured line ('passes by four points') moved up into the rule scene and the video now ends on 'The index outgrew the definition.' SCENE-ARTWORK: the new scene-3 picture was a second pie chart, and the cover is already a pie; it is now a skyline where three towers dwarf a row of low buildings, which is what weighting by market value looks like. Second pass, same day. LENGTH and SCENE-ARTWORK both pointed at the same scene: 'Those ten are the names you already know' restated the 38 percent its own card and stamp already showed, and its picture (ten towers crowding shopfronts) repeated the device its neighbour had just used. It was cut rather than rewritten. PUNCHLINE, the more important one: 'The index outgrew the definition' was not earned - the video shows the index PASSING the test, so the line asserted something the evidence contradicts. It is gone. The close is now what the record actually supports: three companies are individually over the 5 percent limit, the index clears the test by four points, and in 2019 the regulator said it would not act if an index pushed a fund past the line. The rule bending is the reversal, and it is quoted from the letter rather than inferred.

rule stated in one part

Owner, 2026-09-05: 'it says "3 quarter of something cannot exceed 5%" I'm not getting it.' Fair - the statutory test has two halves (at least 75 percent of assets, held in positions of 5 percent or less each) and the script tried to say both in one breath, which parses as neither. The narration now states only the half the video needs: the 5 percent line, and what crossing it does. The 75 percent half is precise and belongs in the record and the description, not in a sentence heard once at speed. Nothing is overstated by the shorter form: a holding above 5 percent genuinely cannot be counted toward the diversification requirement of Investment Company Act section 5(b)(1), which is exactly what the line says.

punchline earned

Owner, 2026-09-05: 'why it doesn't have a punch line?' The previous ending - 'In 2019 the regulator told index funds it would not act if an index pushed them past that line' - is a fact, not a punch: nothing turns over on it and nobody repeats it. The reversal that was sitting in the data unused: the three largest holdings are each above the 5 percent line, so under the statute they cannot be counted toward the fund being diversified. The biggest bets in a diversified fund are, in law, too big to count as diversification. Measured from data/filings/spy-weights.json as of 02-Sep-2026: NVDA 8.2269, AAPL 7.2241, MSFT 5.5860 percent. The 2019 SEC relief moved to an on-screen stamp and the description - it is context for why this is tolerated, not the thing worth repeating. Second pass: the closing comparison was still a bare assertion - the viewer heard that three holdings were 'over that line' without hearing a single weight, so neither side of the comparison was ever spoken. The hook was rewritten to carry it instead of the ten-company 38 percent figure: it now names Nvidia at about eight percent in the first sentence, so by the time the five percent line arrives the viewer has already heard the number that breaks it, and the punch line lands on a comparison they can do themselves. This also keeps the video inside the three-number cap: eight percent, five percent, and the index's own name. The scene 6 stamp 'SEC RELIEF: 2019' was deleted - it introduced an institution and a year the narration never mentions, propping up the close with a new fact.

on screen text swept

validations/on-screen-brevity.md sweep, 2026-09-05. The dominant defect was one habit: the headline restating the scene's own narration, so the card read aloud what the burned-in captions were already spelling out. A headline is the label a viewer takes in at a glance — the angle, not the sentence. Also fixed: highlights appearing nowhere in their headline (they paint nothing), stamps repeating a figure the headline already carries, over-long headlines, and stamps over five words. 8% and 5% were each told three times in their scene; scenes 3, 4, 5 and 6 restated their narration. Scene 5's headline now names the concept the closing line depends on, which the assumed-knowledge check separately flagged as never explained.

yellow belongs to type

Owner, 2026-09-06: 'Pull the yellow prompt.' Every image prompt asked for 'one marker-yellow highlight strip'. In black and white that strip WAS the accent and it worked; in colour the generator started using yellow as a compositional block, so the artwork's yellow competed with the yellow our typography uses to point at a word — two things claiming the same signal on one frame. Yellow now belongs to the type alone. The collage keeps its torn-paper colour blocks, which no longer include a deliberate yellow marker.

no marks in artwork

Owner, 2026-09-06: 'costco card usually have costco on it. is it a fair use?' Naming a company in the narration is ordinary editorial use and this channel depends on it; rendering its mark is a different act, and a generated mark is always a wrong mark - an approximation that misrepresents the logo rather than reporting on it, on top of platform brand complaints that are decided without adjudicating anything. The Costco card was already specified as 'a plain white plastic membership card'. The real exposure was incidental branding nobody had thought of as a brand shot: supermarket shelves, storefronts, packaged goods, and a container port whose boxes carry shipping-line marks. 42 of 63 prompts in the library lacked the clause; every image and video prompt now ends 'no text, no logos, no brand names'.

house style 2026 09 06

Applied the house changes settled on the egg-prices and costco pilots: image prompts moved from newspaper halftone black and white to full-colour magazine collage (saturation 41 -> 84, colourfulness 48 -> 74, against Economics Explained's 91 / 54); no marker-yellow in any prompt, so yellow belongs to the typography alone; and scenes after the stakes card carry the narration captions plus at most one stamp holding a figure, which is what the field's inner slides actually do. The colour-era prompt constraints - pin real colours, state how people are framed, state relative scale - are enforced by validations/scene-artwork.md and repaired by the heal loop below.

healed

Healed 2026-09-06 against acronyms, ambiguity, announced-structure, assumed-knowledge, logical-connectives, on-screen-brevity, parallel-structure, plain-sentences, pronouns, punchline, simplicity and scene-artwork: the title dropped its unsupported verdict and now glosses the initialism inside sixty characters while the narration spells out Standard and Poor's; every pronoun was replaced with its noun; the 'twenty-three of them are half' scene was cut, which took the script back to three figures (eight percent, five percent, three holdings) and removed the bare 'half' with no stated whole; diversification is now glossed before the rule, the five percent line is stated as a share of the fund, the count of three is announced before Apple and Microsoft are named so the punch line introduces no new fact, the cover and stakes cards stopped restating their own narration, and the borrowed regulator-facade and balance-scale artwork was replaced by pictures of each scene's own subject plus two b-roll clips so the beat after the stakes card moves; deliberately left: the central-claim-evidence request for Apple's and Microsoft's individual weights, because narrating them would take the script to five figures against the three-figure cap, so the two are carried by the published-holdings attribution the narration names and the weights stay in the description and the record.

number budget kept at three

The cloud build failed on audit_numbers: '4 distinct numbers, cap is 3'. The fourth was the digit in the stamp 'OVER THE LINE: 3', which the narration already says in words - 'three companies are over that line' - so the stamp was spending a slot in the number budget to repeat something nobody needed twice. The stamp now reads 'OVER THE LINE' and the budget is 500, 8 percent and 5 percent. Worth noting the failure came from the MECHANICAL audit, before any model read the script, and it is the cheapest kind of finding in the pipeline.

silent outro card

Owner, 2026-09-06: 'Can we add a call to action slide at the end ... for users to follow my econplot videos to understand what moves your wallet.' Implemented as the one shape this channel's own rules allow. validations/punchline.md: 'A silent end card (no words) after the punch line is allowed - it is a held frame, not more video', so the punch line still lands last in the NARRATION and nothing follows it aloud. validations/editorial-register.md bans comment-bait specifically ('comment X and I'll send you', 'link in bio for') and says a plain editorial ending is fine - a channel name and what the channel is about is neither bait nor financial advice, which is what STRATEGY.md's 'no call to action' was written to exclude. validations/scene-artwork.md permits shared artwork for 'a channel-level element (a logo, an outro card)', so one card serves every video rather than each generating its own. The line is topic-tied where the topic earns it - 'what moves your mortgage' on the bond video - and the house line otherwise. 1.8 seconds, no narration, no audio.

Related

• What actually sets my mortgage rate?