The Russell 2000 is a way to measure a particular slice of the stock market. Understanding the slice matters more than remembering an index number.
A headline saying “small caps rallied” usually compresses several questions: which index, what period, how are its members selected, and how much influence does each member have? This guide connects those questions to the Russell 1000, Russell 2000 and Russell 3000. You will build a small fictional index, check its return, and finish a record you can reuse when reading a real factsheet.
In this guideJump to a section
Choose a useful first pass
For the core lesson, follow these stops, then check your explanation and keep one record. You can finish there today. The full guide remains a reference; sections labeled optional go further.
- Name the size segment: the Russell family.
- Check who qualifies: eligibility and banding.
- Follow the weights: the weighted-return exercise.
- Check the return convention: compatible comparisons.
One broad universe, two complementary size segments
FTSE Russell’s Russell 3000 overview describes a broad US equity benchmark designed to cover approximately 98% of investable US market capitalization. The Russell 1000 and Russell 2000 divide that parent universe into larger and smaller company segments. Those are size benchmarks, not exchanges, account types or lists of recommended purchases.
| Index | Segment | How to read the name |
|---|---|---|
| Russell 3000 | Broad US equity universe | Approximately 3,000 companies in the size-selection framework |
| Russell 1000 | Larger-company segment | Approximately the largest 1,000 within that framework |
| Russell 2000 | Smaller-company segment | Approximately the next 2,000; the small-cap complement to the Russell 1000 |
The member counts do not tell you the segments’ economic weights. Two thousand smaller businesses can have less combined market value than a much smaller number of giant businesses. Likewise, combining equal dollars in two funds tracking these segments would not automatically reproduce the Russell 3000’s market-cap weights. First identify the exposure being combined and its weighting.
“Small” is relative to this eligible universe and its changing size boundary. It is not a permanent promise that every member sits within a fixed dollar range. A company can grow while remaining in the index between reviews. Always attach a date to a published capitalization range, and distinguish the range at the selection date from values measured later.
Eligibility comes before size ranking
The August 2026 construction methodology, version 7.2, sets country-assignment, exchange, security-type, size, price, float and voting requirements, including exceptions. A US exchange listing alone does not establish US classification. Membership is ranked using total company capitalization; index weighting then uses available, float-adjusted capitalization. These are different calculations.
Existing-member banding near the Russell 1000/2000 boundary can retain a company’s segment despite a modest ranking change. Consequently, “ranks 1,001–3,000” is a useful starting picture, not a rule that reproduces every actual constituent. Multiple eligible share classes can also make security counts differ from company counts. Use “approximately” and inspect the factsheet’s count label.
Those details explain why copying a generic website’s largest-company table is not a reliable reconstruction of the index. The table might use a different date, country definition, share count or eligibility screen. A discrepancy calls for checking definitions, not immediately deciding that one source is wrong. The provider’s actual constituent information remains the reference for membership.
Build a six-company model before handling thousands
This original exercise is deliberately smaller than the real Russell family. Assume six fictional companies all pass eligibility, each has one share class, and no banding or corporate actions apply. Total company market values are $40 billion, $30 billion, $10 billion, $8 billion, $6 billion and $4 billion. Our invented rule places the largest two in a “large” group and the remaining four in a “small” group.
The lower group contains Cedar, Harbor, Meadow and Pine at $10 billion, $8 billion, $6 billion and $4 billion. These are invented dollar values, not current Russell cutoffs. Under the toy rule, Pine qualifies despite being smaller than Cedar. Neither a low share price nor a familiar company name was used to decide membership.
Now imagine Cedar’s total value rises to $31 billion while everything else stays unchanged. At a fresh application of our simplified ranking, Cedar would enter the top two and the $30 billion company would move into the lower group. That illustrates how a size benchmark can change membership. It does not predict an actual Russell migration, because the real methodology contains the eligibility, timing and banding rules just described.
The number of names is not the weight of each name
Free-float adjustments remove shares treated as unavailable to public investors under the provider’s rules. Multiplying price by the available shares gives the value used for weighting. Our float and share-count guide explains why that differs from simply multiplying price by all outstanding shares.
For the four-company small group, assume the following available values. Divide each by the group’s $16 billion total. This is an original teaching model, not actual index holdings or free-float data.
| Company | Total value | Available value | Starting weight |
|---|---|---|---|
| Cedar | $10 billion | $8 billion | 50% |
| Harbor | $8 billion | $4 billion | 25% |
| Meadow | $6 billion | $3 billion | 18.75% |
| Pine | $4 billion | $1 billion | 6.25% |
| Total | $28 billion | $16 billion | 100% |
Suppose their prices then change by +6%, −4%, −2% and −8%, respectively. Hold shares and membership fixed, and assume no dividends or other adjustments. The portfolio-style one-period calculation is:
(50% × 6%) + (25% × −4%) + (18.75% × −2%) + (6.25% × −8%) = +1.125%.
The four contributions are +3, −1, −0.375 and −0.5 percentage points. Three of four stocks fell, yet the weighted measure rose. Counting advancing stocks answers a different question from measuring the value-weighted return. The equally weighted average would be −2%; using that average would silently change the exercise’s weighting rule.
Check the answer a second way: $8 billion becomes $8.48 billion, $4 billion becomes $3.84 billion, $3 billion becomes $2.94 billion and $1 billion becomes $0.92 billion. The total rises from $16 billion to $16.18 billion. Dividing 16.18 by 16 and subtracting one again gives 1.125%.
This simplified check follows the weighting principle explained in FTSE Russell’s index calculation explainer. Real index calculations maintain continuity through membership and corporate-action adjustments. Adding a company is not, by itself, an investment gain. An index level such as 2,000 is also not a share price or a dollar valuation of all its businesses.
The 2026 schedule changed: June and December
As checked November 1, 2026, the Russell US size indexes use semiannual reconstitution in June and December. An older explanation saying the process happens only once a year is incomplete for 2026. The current provider calendar and December schedule announcement take precedence over stale overview wording.
For the upcoming December review, rank day was October 30. Preliminary additions and deletions are scheduled to begin November 13. Changes are scheduled for after the December 11 US close, with the revised indexes operating at the December 14 open. Preliminary information is not a final constituent list. Recheck the official calendar before using these dates.
Reconstitution revisits membership and size segmentation. Rebalancing concerns the weights and shares used to implement an index; the terms sometimes appear together because a membership review can also change weights. Eligible IPO additions and ongoing corporate-action maintenance mean the indexes do not remain completely frozen between the two reviews.
December is not an identical rerun of every June procedure. In particular, the provider’s October 26 explanation says growth/value style probabilities are updated for additions and migrations, while companies remaining in the same parent index retain their June probabilities. Read the specific size or style index’s rules instead of assuming that every product with “Russell” in its name changes the same way.
A fund following a changing benchmark may need to adjust holdings. That can create trading needs around implementation, but it does not establish a guaranteed gain from buying a candidate beforehand. Other participants can anticipate changes, prices can move for unrelated reasons, and a tracking fund has implementation choices and costs. Treat a membership announcement as a documented index event, not a company earnings result.
What a small-cap label can and cannot tell you
A broad collection of smaller companies spreads exposure across many businesses, but count alone does not describe concentration. Inspect constituent weights, industries and the date of the data. A sector can matter even when no single name dominates. The diversification lesson shows how to look through a basket instead of treating its label as the whole risk assessment.
Rate sensitivity is a question to investigate, not a fixed law that all small caps rise when rates fall. Consider a fictional company with debt refinancing next year: a higher refinancing rate could increase its costs. A second company with substantial cash and no near-term refinancing has a different exposure. Customer demand, pricing, profitability and the reason rates changed can also matter. These competing mechanisms prevent a single macro headline from explaining every constituent.
Likewise, index inclusion is not a profitability certificate. When a factsheet presents a valuation statistic, read the qualifier. For example, a P/E statistic that excludes negative earnings describes a calculation with exclusions; it does not prove that every member earns a profit. Record how the provider treats missing or negative values before comparing ratios across benchmarks.
The official Russell 2000 factsheet provides dated characteristics and return information. Its figures can change after this lesson’s review. Use a saved, dated copy for a reproducible comparison; do not combine an old sector table with a new return and label them one contemporaneous snapshot.
Match the return before comparing performance
A price-return index and a total-return index can describe the same constituent universe but answer different performance questions. Total return includes the provider’s dividend-reinvestment treatment; price return does not include that same regular-dividend reinvestment. FTSE Russell’s education centre explains the distinction. Also match the currency, start and end dates, and any tax convention.
For a fictional one-period holding, start at $100, end at a $102 share price and receive a $1 cash dividend at period-end. Price return is 2%; adding the stated distribution gives 3% for this simplified period before costs and taxes. This example omits reinvestment timing. It demonstrates why comparing a price-only chart with a distribution-inclusive result can mislead.
An ETF is a separate investment product that may track an index. Its fees, trading price, distributions and implementation can affect an investor’s realized result. Do not compare an ETF’s share price directly with an index’s point level. Compare compatible percentage returns, state whether distributions are included, and distinguish a fund’s market-price return from its net-asset-value return.
Optional: Rebuild the broad basket using capital, not company counts
The Russell 1000 and Russell 2000 are complementary size segments within the Russell 3000. That does not give them equal influence on the combined benchmark. To understand the aggregation, use each segment’s share of the broad basket’s applicable index capitalization at the start of the period, with consistent returns and membership treatment.
Consider an invented universe whose large-company segment has $90 billion of index capitalization and whose small-company segment has $10 billion. The broad basket’s weights are 90% and 10%. Suppose the large segment has a price return of +4% and the small segment has a price return of −6%, with no distributions, intervening membership changes or corporate actions. Their ending values are $93.6 billion and $9.4 billion. The total rises from $100 billion to $103 billion: a 3% gain.
The weighted calculation agrees: (0.90 × 4%) + (0.10 × −6%) = 3%. A simple average of the two segment returns gives −1%, which answers a different question. It describes an initially equal-dollar allocation to the two segments, not this broad capitalization-weighted basket.
A fictional 50/50 portfolio would therefore trail the broad basket by 4 percentage points during this interval, despite both using the same underlying segment returns. That gap comes from the starting allocation. It does not identify a stock-selection error. The 90/10 figures are teaching assumptions, not current Russell weights or an allocation recommendation.
After these returns, the large segment represents 93.6 ÷ 103, or about 90.87%, of the broad basket. A second-period calculation should use the applicable new weights rather than silently reuse 90/10. If you intentionally reset a blended benchmark to fixed weights, state its reset schedule. For real index replication, use the provider’s index-share data and maintenance conventions, not this simplified two-segment model.
Optional: Do not rebuild the past from today’s members
Small-company membership changes as businesses grow, shrink, merge or cease to qualify. A current constituent file describes the current basket. It does not by itself identify the holdings and weights that produced a return years earlier. Looking only at present members can omit companies that mattered during the period you are trying to explain.
Use another fictional example, separate from the earlier six-company universe. At the start of a year, a two-company small-cap basket holds Fir and Elm, each with $50 of index value. Fir gains 100% and finishes at $100; Elm loses 40% and finishes at $30. With unchanged holdings during that year, the basket ends at $130 against its original $100: +30%.
Assume a scheduled review after that year removes Fir and admits Pine. Pine had gained 10% during the year, but it was not a member during that period. A researcher opens the new constituent file, averages Elm’s −40% and Pine’s +10%, and reports −15% as the old basket’s return. That calculation uses the wrong historical membership. It also substitutes a simple average for a verified historical weight calculation.
The example is not a model of the Russell admission rules; it isolates the data mistake. Keep dated constituent files, historical weights and effective dates when making an attribution. If those records are unavailable, use the provider’s published historical index series for its measured return and identify the missing inputs before claiming to explain individual contributions.
Optional: Check what a valuation summary excludes
A reported price-to-earnings ratio can depend strongly on how losses are handled. Imagine two fictional businesses, each valued at $100. One earns $10 over the stated year; the other loses $5. Using aggregate value divided by aggregate earnings gives $200 ÷ $5 = 40 times earnings.
If an illustrative summary instead excludes the loss-making company from both numerator and denominator, the ratio becomes $100 ÷ $10 = 10 times. The underlying businesses did not change. The included sample changed. These are deliberately specified classroom calculations, not claims about the exact method in any particular Russell factsheet.
Before comparing an index’s valuation with another benchmark, identify the earnings period, forecast versus reported earnings, loss treatment and aggregation method. If aggregate earnings are zero or negative, an ordinary positive P/E interpretation fails; do not replace a missing ratio with zero and call the basket inexpensive.
Practice: what can the three checks establish?
The broad-basket calculation establishes a 3% return under the stated 90/10 assumptions. The current-member file does not establish the earlier basket’s performance. A 10-times ratio for profitable firms does not describe all the companies in the two-business example. In each case, the useful next question is the same: which observations and weights are actually included in the denominator?
Build a seven-line index-reading record
Open a provider factsheet and its linked methodology, then write these seven lines. Leave a field unanswered when the source does not supply it. The aim is a note another reader could reconstruct.
- Identity: exact index name, provider, source URL and factsheet date. Specify size, growth/value or another variant.
- Universe: what market segment is represented, and how country and security eligibility are defined.
- Selection versus weighting: record the membership rule separately from float-adjusted weights. Note banding without inventing a fixed dollar cutoff.
- Composition: count label, largest weights and industry exposures, all from the same dated source.
- Performance: comparison period, currency and price/total-return convention. Keep fund returns separate from index returns.
- Maintenance: last review, next scheduled review and whether the information is preliminary or effective.
- One unresolved question: identify a definition, exposure or calculation that needs another primary source.
For our fictional small group, a completed calculation line reads: “Four companies; $16 billion of available capitalization; starting weights 50%, 25%, 18.75% and 6.25%; one-period price return +1.125%; shares fixed, no dividends or corporate actions.” An open question is how the result would change under a different membership or weighting rule. None of that establishes which real security to buy.
Check your understanding: four claims to repair
“Two thousand names means equal weights” confuses count with weighting. “Small cap always means the same dollar ceiling” ignores changing boundaries. “Three of four stocks fell, so the index fell” ignores contributions. “The Russell is reviewed only in June” misses the 2026 move to June and December. A sound correction names the rule and date behind the answer.
Keep the completed record with the research notebook, then return to the market-index learning route to compare a differently constructed benchmark. Read ETF basics before treating an index description as a fund description.
Finish with one explanation you can check
Your finish line: Explain the size segment without confusing company counts and weights.
Would equal dollars in a Russell 1000 fund and a Russell 2000 fund automatically recreate the Russell 3000?
No. The two segments contain different amounts of float-adjusted market value. A 50/50 dollar mix is a chosen allocation, not automatically the broad index’s market-value weighting. Fund expenses and implementation also affect results.
Keep one completed shared index-reading worksheet (.txt). The copyable blank record and completed four-stock example are available on the S&P 500 lesson.
Next: compare the same stocks under different weights. Use one shared record to separate index return, breadth and the weighting rule.
What changed: November 2, 2026: added the core route, completion check and shared worksheet links. The November 1 methodology checks above retain their original scope. Prepared with AI assistance; this is not an independent professional review.
