The question of “whether this market is fragmented or controlled by a few players” arises in almost every analytical project – and almost always ends in a dispute over data. HHI market concentration and the CR4 and CR5 ratios make it possible to turn intuition into a number, but only when it is clear which market definition and which estimate of market shares they are based on. Below is a practical guide: how to calculate these metrics, where to obtain data when information is incomplete, and how to verify the result so that it does not measure something other than what was intended.
What does HHI market concentration indicate, and when is it worth calculating?
HHI market concentration is a measure describing the extent to which sales (or another volume-based variable) are concentrated in the hands of the largest entities. It is not an assessment of the market – it is a description of its structure and a starting point for further analysis of competition, pricing power, or the dynamics of market entry and exit.
The three most commonly used metrics differ in how much information they retain about the distribution of market shares:
- CR4 and CR5 – the combined market shares of the four or five largest entities. Simple, easy to interpret, and requiring data only on market leaders.
- HHI (Herfindahl-Hirschman Index) – the sum of the squared percentage shares of all market participants. Squaring the shares means that large players have a disproportionately greater impact on the result than small ones.
- Effective number of competitors – the inverse of HHI calculated using fractional shares; a supplementary interpretation that translates the metric into “how many equally sized players would correspond to this structure.”
The practical difference is significant. Two markets may have an identical CR4 but a completely different HHI: in the first, four players each hold one-quarter of the market; in the second, one has a clear majority, while the other three each have only a few percent. CR4 does not distinguish between these situations, whereas HHI does. This is why, in research practice, both metrics are reported together rather than used interchangeably.
When does measuring HHI market concentration make sense? When the research question concerns market structure rather than market size alone: when monitoring changes in the balance of power over time, comparing countries or segments, analyzing market fragmentation in categories dominated by local suppliers, or describing the effects of consolidation. The metric itself does not answer what this means for a particular company – it answers what the distribution looks like.
How can HHI market concentration be calculated using incomplete data?
The textbook formula is straightforward. The difficulty lies in two areas: defining the market and estimating the shares of entities that do not publish data. The workflow in a research project is as follows:
- Define the relevant market. There are three dimensions: product (what is a substitute and what is not), geographic (country, region, or the customer’s travel area), and value chain level (manufacturer, distributor, or point of sale). This decision determines the result more strongly than data quality. The same set of companies will produce a high HHI in a narrowly defined market and a low one in a broadly defined market.
- Select the measurement variable. Revenue, sales volume, number of subscribers, installed capacity, number of locations, or floor space. A metric calculated based on value may differ substantially from one calculated based on volume if players operate in different price ranges.
- Build a list of market participants. Business registers, financial statement databases, membership lists of industry organizations, trade fair directories, concession and permit registers, and customs data for import markets.
- Estimate market shares. For reporting entities – financial data, taking into account the corporate group structure and the allocation of revenue to the relevant segment. For the remaining entities – triangulation: expert interviews with channel participants, employment and productivity data, the number of sales outlets multiplied by estimated turnover per outlet, and data from suppliers of raw materials or components.
- Close the market. The sum of shares must add up to the entire market. The remainder (“other”) requires an assumption about its distribution – and that assumption must be described explicitly, as it affects HHI.
- Calculate the metrics and test sensitivity. HHI, CR4, and CR5 across several market-definition variants and several scenarios for estimating the shares of non-reporting entities.
The most common technical issue is the so-called market tail. When there are many small players, it is not possible to estimate each one individually. HHI mathematics helps here: individual shares below one percentage point make only a small contribution to the metric because their squares are small. However, the combined contribution of many such entities may still be significant. It is therefore worth estimating market leaders and total market size reliably, while treating the remainder as dispersed – with an explicit assumption about the average share in the tail and an assessment of how changes to it affect the result.
HHI is particularly sensitive to errors in estimating the shares of the largest players because their shares are squared. Therefore, in analytical projects, the priority is to verify the top of the ranking rather than refine estimates for the tail.
In fieldwork practice, triangulation often means comparing at least three independent sources. An example from a B2B category: reporting data indicate a company’s revenue, interviews with distributors suggest its position in the channel, and import volume data make it possible to verify whether the physical scale is consistent with the stated value. A discrepancy between these sources is not always an error – it may mean that part of the company’s activity falls outside the defined relevant market or that its revenue includes related services.
Which errors most often distort the measurement of market structure?
Concentration metrics are easy to calculate and just as easy to calculate incorrectly. Below are the pitfalls most commonly encountered in market structure analyses:
- Inconsistent market definition across periods. Comparing HHI from two years when the segment definition or data source changed in the meantime produces an apparent change in concentration. Time series require all years to be recalculated according to the same definition.
- Ignoring ownership structure. Companies controlled by the same corporate group should generally be treated as a single market participant for concentration purposes. Treating them separately may systematically underestimate HHI, CR4, and CR5.
- Mixing value chain levels. Combining manufacturers with importers and distributors in a single market share table leads to double-counting the same sales.
- Ignoring geography within a country. Local service markets may be fragmented nationally while highly concentrated at the city or county level. In such cases, national HHI describes a structure in which customers do not actually make a choice.
- Treating thresholds as verdicts. Interpretive thresholds used in regulatory and analytical practice are conventions, not laws of nature. When estimates are subject to uncertainty, a result close to a threshold should be reported as a range rather than as a single value.
- No information about data coverage. A metric calculated on a sample covering part of the market, without adjustment for the missing portion, is not comparable with a metric calculated for a closed market.
A separate limitation is substantive in nature: HHI market concentration is a static measure. It contains no information about barriers to entry, the rate of player turnover, buyers’ bargaining power, parallel imports, or whether market shares are stable or change from year to year. Two markets with the same HHI may have entirely different dynamics. This is why research reports combine concentration metrics with a measure of changes in market shares over time and with the number of entries and exits during the period analyzed – only then is market structure described in a way that is analytically useful.
Alternatives and supplements worth considering include the Lorenz curve and Gini coefficient (which describe inequality in the distribution of shares), the entropy index (which is more sensitive to changes in the middle of the distribution), and, when analyzing market fragmentation, simply the number of entities exceeding a specified scale threshold. None of these replaces HHI, but each shows a different aspect of the same distribution.
How can a completed measurement be verified? A checklist
Before using the result for further work, it is worth going through a set of control questions. Every “I don’t know” on this list indicates an area in which the metric may be misleading:
- Is the definition of the product market, geographic market, and value chain level stated in one sentence, without exceptions or caveats?
- Were shares calculated based on value or volume – and is this choice aligned with the research question?
- Have capital-linked entities been aggregated?
- Do the shares add up to the entire market, and does market size come from an independent estimate?
- How many leading players have a share confirmed by at least two independent sources?
- How do HHI, CR4, and CR5 change when assumptions about the shares of the largest entities are varied within a reasonable range?
- Does the report include the data reference date – given that concentration is always a snapshot of a specific point in time?
- Is a measure of changes in market shares over time provided alongside the metric?
Measuring HHI market concentration according to this sequence produces a replicable result – and in this case, replicability is more important than precision to the second decimal place. A metric whose assumptions are explicitly described can be compared across periods and markets. A metric presented without methodology is a meaningless number.
Frequently asked questions
How is HHI calculated?
HHI is the sum of the squared market shares of all market participants, expressed in percentage points. For a market with players holding shares of 40%, 30%, 20%, and 10%, the result is 1,600 + 900 + 400 + 100, or 2,600. Alternatively, it can be calculated using fractional shares – in that case, the result ranges from a value close to zero to one; both variants should be clearly labeled because their scales differ by four orders of magnitude.
How does HHI differ from the CR4 ratio?
CR4 adds up the shares of the four largest entities and ignores everything beyond them, as well as the distribution within those four. HHI includes all market participants and, by squaring shares, distinguishes a market with one dominant player from a market with several equally sized players. CR4 and CR5 are easier to communicate and less data-intensive, while HHI is more sensitive to asymmetry – which is why they are reported together.
What data are needed to measure market concentration?
Accurately calculating HHI requires the shares of all market participants, aggregated at the corporate group level. For a measurement based on incomplete data, the minimum requirement is an independent estimate of the size of the relevant market and the shares of the largest entities, supplemented by an explicit assumption about the distribution of the remaining part of the market. Sources include financial statements, registers and databases of business entities, customs and production data, industry organization statistics, and expert interviews with sales channel participants. For small players in the market tail, an aggregate estimate may be used together with an assumption about the distribution, because the impact of a single small share on HHI is limited.
Ask about market structure and concentration analysis
If you need to calculate HHI, CR4, and CR5 in a market where most players do not publish data, Hume’s Institute designs the measurement process from defining the relevant market to triangulating market shares. Contact us to discuss the scope of data and the estimation methodology.