When an Establishment Is New but the Firm Is Not
The BLS Business Employment Dynamics (BDM) does not define a new business by asking whether a new legal company has been incorporated. It observes employment at operating units across two March reference dates. The question here is what that lifecycle measure can tell us—and what it cannot—about openings, job flows, and the age of the firms behind them. That distinction lets us use BDM data without turning every newly active location into an independent startup.
The key is to keep three objects separate. An establishment is the local economic unit where goods or services are produced. A firm is the legal business that may own one establishment or several. Employment is the number of jobs associated with those units. A change in one object can occur without a matching change in the others. An established firm can add a location; a location can lose jobs without closing; and a firm can be young even when its first establishment is not classified in the same way as a later branch.
Opening: positive employment is a threshold, not a biography
BLS defines an opening as a unit with positive employment in March of the current year after reporting zero employment in March of the previous year. That is a precise and useful rule. It identifies a transition from no recorded employment to positive recorded employment at an operating unit. It does not attempt to describe every event that produced the transition.
The rule therefore answers a narrower question than “How many new companies were founded?” It tells us how many units crossed an employment threshold between two reference dates. The unit might belong to a newly formed firm. It might be a new branch of a long-established chain. It might be a work site whose reporting relationship has become visible in the frame. The BDM definition allows these possibilities to coexist; it does not supply a universal conversion from openings to independent firm births.
That is not a defect in the measure. Local employment is an important object in its own right. A new clinic, warehouse, restaurant location, or care-service work site can change where jobs are available and where services are delivered even if the parent firm has existed for decades. If the research question concerns job creation by place, treating the establishment as the unit is appropriate. If the question concerns entrepreneurial entry, the establishment count is a starting point, not the conclusion.
The word “opening” also describes a comparison, not necessarily a single moment. The current observation is positive and the prior March observation is zero. The unit may have hired workers shortly before the current reference date or may have operated for part of the interval under a form that was not yet represented in the same way. An annual category tells us what changed between the observations; it does not provide a day-by-day history of the unit.
This is why a headline can overstate what the measure establishes. “Openings rose” is a statement about newly active establishments under a defined frame. “More startups were founded” adds a claim about independent firms and their legal or organizational identity. To bridge those statements, a researcher must identify the parent firm and determine whether the unit is part of an existing multi-establishment organization.
Gross flows: what changes between two March observations
BDM becomes more informative when openings are read together with expansions, contractions, and closings. An expansion is a unit with positive employment in both the previous and current March and a net increase over the interval. A contraction remains active in both observations but has a net employment decrease. A closing has positive employment in the previous March and no employment in the current March.
These categories divide changes at existing and new units into gross job gains and gross job losses. Gross job gains are the jobs added at opening or expanding units. Gross job losses are the jobs lost at closing or contracting units. Net employment change is the difference between those two totals. The categories are related, but they answer different questions.
Suppose a group of establishments adds 100 jobs at openings and expansions while losing 80 jobs at contractions and closings. Net employment rises by 20 jobs, but the gross figures reveal 180 jobs of movement. The example is arithmetic rather than a claim about any particular year. Its point is that net change is a balance, while gross flows describe the activity that produced the balance. A small net increase can accompany substantial creation and destruction at the establishment level.
The distinction matters for the startup question in two directions. First, a rise in gross gains can reflect expansions by existing establishments as well as openings. It is not a pure measure of firm entry. Second, a rise in openings can coexist with a large number of closings. The economy may be churning through locations rather than adding a corresponding number of durable independent businesses. Reading only the net change would conceal that difference; reading only openings would conceal it in another way.
The March-to-March design also controls what persistence means. A unit that appears with positive employment in the current March is an opening relative to the previous March, but its later status depends on what happens in subsequent comparisons. It may continue and expand, continue while contracting, or disappear. The classification is attached to the transition being measured, not to a permanent identity such as “successful startup” or “failed startup.”
The lifecycle can be represented as a set of transitions rather than as a simple success story:
The diagram is deliberately modest. BDM’s categories are defined by adjacent observations, so the arrows are a way to organize the possible statuses across comparisons, not a claim that every unit passes through them in one fixed order. An establishment that is active in one interval can expand in the next and contract later. A unit’s local employment history and its parent firm’s organizational history must be tracked separately.
That separation is especially important when people use gross job flows to describe dynamism. Dynamism can mean high rates of entry and exit among locations, rapid expansion within surviving establishments, or reallocation of workers across an established firm’s network. BDM can illuminate the first two at the establishment level. It does not, without firm-level interpretation, tell us how much movement came from independent entrepreneurs versus organizations scaling an existing model.
Age belongs to a unit—and to a different unit when the label is “firm age”
BLS defines establishment age as the difference between the reference period and the first time an establishment reported positive employment. Firm age is classified using the age of the oldest establishment within the firm. These rules encode two different histories. Establishment age asks when a local unit became active. Firm age asks how long the organization’s oldest observed establishment has been active.
Imagine a firm that has operated one location for twelve years and opens a second location this year. The second establishment can fall into the youngest establishment-age group even though the firm belongs to an older age group. If a reader calls the second location a “twelve-year-old firm,” they have transferred the firm’s age to the wrong object. If they call it a new independent startup, they have made the opposite error. The useful description is that an older firm added a young establishment.
The oldest-establishment rule also makes firm age an organizational measure rather than a simple average of locations. It gives the firm a stable anchor in its earliest observed establishment, allowing researchers to distinguish a young organization with a young network from an old organization that is still adding new units. The sources do not say that every firm is perfectly observed, so the rule should not be treated as a complete biography of ownership. It is the classification BLS uses for the firm-age view.
This is where the BDM’s alternative measures become valuable. Gross flows are measured at the establishment level, while information about establishments and firms can be used to report those flows by establishment age, firm age, establishment size, and firm size. The same employment event can therefore be described from different perspectives. A job added at a new branch is part of the branch’s young-establishment experience, but it can also be part of an older firm’s expansion.
The difference is not merely semantic. Consider two economies with the same number of openings. In the first, each opening belongs to a newly formed single-establishment firm. In the second, every opening is a branch added by established multi-unit firms. Establishment-level openings are identical, but the implications for entrepreneurship, concentration, organizational scaling, and competitive entry are not. A statistic that does not preserve the firm relationship cannot distinguish the two cases by itself.
Size classes and denominators can change the apparent meaning of “young”
Age is often combined with size, which adds useful detail but also more opportunities for category error. BLS offers establishment and firm characteristics separately. Establishment size is the employment reported for one establishment. Firm size is the total employment reported across establishments under the same employer identification number. A small establishment can therefore belong to a large firm, while a large establishment can be the only unit of a smaller firm.
BLS also uses different sizing methods. Base size is employment in March of the previous year. Average size is the average of employment in the previous and current March. Openings are classified by employment in the reference year because they had zero employment in the prior March. These choices affect how a unit is placed in an age-by-size table. They are not interchangeable versions of the same number.
The practical question is always, “Size of what?” A table about establishments with one to four employees describes local operating units. A table about firms in the same nominal size range describes total employment under the owner’s identifier. The two tables can tell opposite stories about the prevalence of small businesses because a large firm can contain many small establishments.
The youngest age group has an additional denominator issue. BLS notes that rates for units less than one year old can be 200% or 0%. The numerator is the relevant opening, expansion, contraction, or closing component. A unit less than one year old is, by definition, an opening, so the opening numerator can be positive while the other component numerators are zero. The denominator is the average of the prior and current employment levels; the prior level is zero. The resulting edge values follow the definition and formula.
A reader who sees 200% without checking the denominator might infer an implausibly large general growth rate. The more accurate interpretation is that the youngest category has a special base-period structure. This is an example of a broader rule: a rate is not portable across categories unless its numerator, denominator, reference period, and population are understood.
The same caution applies to comparisons across age definitions. “Young establishments” and “young firms” are not two casual descriptions of the same population. One follows the local unit’s first positive employment; the other follows the oldest establishment in the firm. If the question is independent business formation, firm age and ownership structure are indispensable. If the question is where new jobs are appearing, establishment age may be the more direct lens.
What the lifecycle measure lets us say
The BDM lifecycle gives a disciplined vocabulary for business dynamism. It can tell us that a unit moved from zero to positive employment between March observations, that an active unit added or lost employment, or that a previously active unit no longer reported employment. It can separate gross job gains and losses from their net balance. Those are substantive facts about the movement of work across operating units.
It does not, on its own, tell us that every opening was an independent company, that every young establishment represents new entrepreneurship, or that a high gross-gain number means the economy has become more competitive. Those conclusions require the firm relationships and frame boundaries that the lifecycle categories alone do not settle.
The most reliable reading therefore keeps two statements side by side: “A new establishment appeared under this definition,” and “The organizational identity behind it is still a separate question.” Once that distinction is in place, the next issue becomes unavoidable. If an establishment and its parent firm depend on how records are assembled and linked, why can the BLS and Census systems see different networks of establishments in the first place?