HRM Flexibility: The Flexible Firm Model, Its Four Forms

HRM flexibility describes the ways an organisation varies the quantity, cost, skill and deployment of labour in response to changing demand. The standard framework is John Atkinson's flexible firm model, published in 1984, which arranges these into a structure of a core workforce and several peripheral groups.

The model is descriptive and normative at once, which is the source of most of the argument about it. Atkinson described what he observed firms beginning to do; the model was widely read as a prescription for what firms should do, and evaluating it requires holding the two readings apart.

The four forms of flexibility

Functional flexibility. The capacity to move employees between tasks and roles as requirements change. It depends on broad skills, training investment and the absence of demarcation, and it is associated with the core workforce because it requires employees who will be retained long enough for the investment to return.

Numerical flexibility. The capacity to vary the number of workers, or the hours worked, in line with demand. Achieved through temporary contracts, agency labour, part-time working, zero-hours arrangements, overtime and short-time working.

Financial flexibility. The capacity to vary labour cost, through performance-related pay, local and not national bargaining, variable bonuses and shift premia. In principle it allows pay to track productivity or ability to pay; in practice it is the least developed of the four in most organisations.

Distancing. Replacing employment with commercial contracts — subcontracting, outsourcing, self-employment, franchising. Strictly this is not a form of flexibility in the labour force but a decision to have a smaller one, which is why some accounts treat it separately.

The core and the periphery

The structural claim is that firms divide the workforce into groups treated differently.

The core consists of employees with firm-specific skills that are difficult to buy in. They receive security, training, career progression and better terms, and supply functional flexibility in return. The employer's investment in them is what makes them worth retaining through a downturn.

The first peripheral group holds jobs that are full-time and permanent but built on skills available in the labour market. Terms are less generous, turnover is expected and tolerated, and the group supplies numerical flexibility through recruitment and attrition, not through contract form.

The second peripheral group holds explicitly contingent arrangements: part-time, temporary, fixed-term, job-share, subsidised trainees. Numerical flexibility here is immediate.

The external group is not employed at all — agency staff, subcontractors, the self-employed, outsourced functions.

The logic is that firms buy security and commitment where skills are specific and buy flexibility where they are not, and that the boundary between core and periphery is drawn by skill specificity rather than by importance to the business. That distinction is worth stressing: a cleaner may be essential and still sit outside the core, because the skill is available elsewhere.

Why firms pursue it

Demand variability. Seasonal, cyclical or unpredictable demand makes a fixed workforce either idle or inadequate for much of the year.

Cost. Peripheral labour typically carries lower hourly cost and, more significantly, lower fixed cost, since the obligation ends with the contract.

Speed. Adjusting headcount through contract expiry is faster and cheaper than redundancy.

Focus. Distancing allows management attention to concentrate on activities where the firm has an advantage.

What the evidence shows

Three findings recur and should be stated rather than asserted.

Numerical flexibility is far more widely adopted than functional. Functional flexibility requires training investment, job redesign and frequently the renegotiation of working practices; numerical flexibility requires a different contract. Surveys since the 1980s have consistently found the second widespread and the first patchy.

The structure is frequently a description after the fact. Firms rarely design a core-periphery structure deliberately. They make a series of separate decisions — outsource this, use agency staff for that, keep this team permanent — and the resulting pattern resembles the model without anyone having applied it.

The effects on performance are mixed. Numerical flexibility reduces cost and reduces commitment, training and accumulated knowledge at the same time. Studies find positive, negative and null effects depending on the outcome measured and the context, which is what should be expected of a practice whose costs and benefits fall on different variables.

What flexibility costs the organisation

The case for peripheral working is usually made on cost, and the costs on the other side are real and systematically under-counted.

Recruitment and induction repeat. Every departure and replacement carries an administrative cost, a period of lower productivity, and the time of a colleague who has to supervise. Where turnover is high by design, these are continuous and not occasional.

Knowledge does not accumulate. A worker who leaves takes what they learned about the particular machine, the particular customer, the particular exception. In operations where the exceptions are what create value, this is the dominant cost and the hardest to see, because the loss never appears as an event.

Quality and safety variance rises. Less experienced workers make more errors, and the effect is measurable in sectors that count incidents. Where the work carries a safety or regulatory dimension, the cost of an error can exceed years of the saving that produced it.

The core carries the difference. Where a periphery absorbs variability, the core absorbs the supervision, the training and the correction. This shows up as workload on exactly the employees the firm has decided are hardest to replace, which is how a flexibility strategy produces turnover in the group it was designed to protect.

None of these is an argument against peripheral working. The argument is that a decision taken on hourly cost alone is taken on one of four relevant numbers.

Flexibility and the psychological contract

The framework's least discussed consequence is what it does to expectations on both sides.

The core-periphery structure makes the employment relationship explicitly conditional, and conditionality is visible to the people it applies to. Employees in the periphery generally understand their position accurately, and the research on contingent workers finds that those who chose the arrangement report satisfaction comparable to permanent staff while those who did not report markedly lower commitment.

That distinction — voluntary against involuntary contingency — accounts for much of the variance in the outcome studies, and it is absent from the original model, which treats peripheral status as a property of the job, not as something with a person's preference attached.

For practice it produces a usable rule. Where contingent arrangements match what workers want — students, people with caring responsibilities, those holding several roles by choice — the costs above are much smaller. Where they do not, the organisation is paying the costs and receiving a workforce that would leave for a permanent position elsewhere, which is the worst available combination.

The criticisms

It describes intention, not practice. The strongest objection, made early and repeatedly. Atkinson presented a model of what a firm might do; the evidence that firms were pursuing an integrated flexibility strategy was thin then and has not strengthened.

The core-periphery division is cruder than reality. Real workforces contain many gradations, and individuals move between categories. A long-serving part-time employee with deep firm-specific knowledge fits neither box.

It normalises a distributional choice. Presenting flexibility as an organisational requirement conceals that the model allocates security to one group and insecurity to another, and that the allocation is a decision. Critics have argued the framework supplied a vocabulary that made that decision easier to describe as necessity.

The costs to the periphery are outside the frame. Income volatility, absence of training, weaker statutory protection and the health effects of insecure work are real and appear nowhere in a model organised around the employer's requirements.

Flexibility runs one way. The model discusses the employer's need to vary labour and says almost nothing about employees' need to vary their own working arrangements. Flexible working in the modern sense — where the employee initiates — is a different subject that shares a word.

Where the discussion has moved

Three developments have changed the terrain since 1984, and an answer that notes them is describing the present and not the model.

Regulation has tightened at the margins. Rules on agency workers, fixed-term employees, part-time workers and, more recently, on the classification of the self-employed have raised the cost of some peripheral arrangements and narrowed the gap the model relied on.

Employment status has become contested. Whether a worker engaged through a platform is employed, self-employed or an intermediate category has been litigated repeatedly, and the answers have not been uniform. Distancing is legally riskier than it was.

Skill shortage has weakened the numerical route. The model assumes peripheral skills are freely available in the labour market, which is what makes them substitutable. Where they are not — and in several sectors they have not been for some years — the firm discovers that its periphery is as hard to replace as its core and has none of the commitment, which inverts the logic the structure was built on.

Employee-initiated flexibility has become a retention issue. Where the model treated flexibility as something the employer needs, the current bargaining is substantially about where and when employees work, which reverses the direction and is the more active question in most organisations now.