In short: A nationalisation ratio is a stock, and a stock only moves by changing the flows into and out of it, so a recruitment quota that addresses hiring alone leaves the leaver flow untouched. Quota driven hiring buys attrition, because placing people into roles faster than supervision and development can absorb them produces early leavers, and the ratio drifts back towards trend a few years later. The binding constraint is the development pipeline rather than the offer rate, since the number of people who can reach independent sign-off in a period is set by the supervisors available to sign them off. A headcount ratio and a capability ratio are different measures, and reporting only the first allows a target to be met with people who are counted but not yet carrying the work.
A commitment gets made at board level. Raise the national share of the workforce from where it sits today to a higher figure by a stated year. By the time it reaches the people who have to deliver it, it has become a hiring number: each function gets a monthly recruitment target, recruitment gets measured on offers accepted, and progress gets reported as a percentage against a curve.
Two years in, the percentage is close to plan. Four years in it has slipped back, and the slip gets explained as attrition, treated as a separate retention problem to be solved with a bonus scheme.
Attrition here is the symptom of how the target was translated. A workforce ratio is a stock, and the only way to move a stock is to change the flows into and out of it. A recruitment quota addresses one flow, ignores the other, and puts the pressure on the part of the system with the least spare capacity.
What gets lost in the translation
A nationalisation target as written is a statement about the composition of a workforce at a future date. It carries no instruction about which roles, at which grades, holding what competence, or how the people in them are supposed to get there. The recruitment quota that comes out the other end is a rate, and it is silent on everything the target was about.
The losses in that conversion are systematic. Roles differ enormously in how long they take to fill with someone who can do the work unsupervised, and a percentage flattens the difference completely. A control room operator, a reservoir engineer, a marine cargo surveyor and a procurement analyst have times to competence that differ by an order of magnitude, and a quota expressed as a single ratio treats a hire into each of them as identical progress.
Organisations that hit the number early usually noticed this and filled the fast roles first, which is a rational response to the measure and means the ratio and the capability start moving apart in year one.
Why quota hiring buys attrition
Take a role with a real competence requirement and put someone into it below that requirement. Three things can happen, and the organisation picks which one by what it does next.
The gap closes through structured development. A defined syllabus, supervised practice against a named mentor, and a sign-off that the person can work unsupervised. This works, and it consumes a scarce resource: the time of people who are already competent.
The gap closes by narrowing the role. The title stays, the harder parts of the job migrate quietly to a contractor or an expatriate colleague, and the person holds a position they are not really doing. This is comfortable for a while and it produces a distinctive kind of leaver: someone two years into a job, aware they are not doing it, who takes the first offer that involves real work.
The gap does not close. The person is exposed, gets held to outcomes they were never equipped for, and leaves. In an operating environment that route sometimes takes an incident with it.
Two of the three end with the person gone, and the middle one is the more damaging because it is invisible on every report the organisation runs: a filled position and a met target get recorded while the work is being done by somebody else. When the leaver goes the ratio drops, the quota reappears, and the next cohort gets hired into exactly the same conditions, none of which were ever the thing being measured.
The pipeline is the actual constraint
Put recruitment aside and look at the system as a pipeline with a stock at the end.
The steady state number of qualified national holders in a role family equals the rate at which people become qualified multiplied by the average time they stay. Both terms are measurable and both are changeable, and the second is usually the cheaper of the two.
Work an example. A role family has 400 positions and the target is a 60 percent national share, which is 240 national holders. If the average tenure of a national holder in that family is four years, holding that ratio needs 60 qualified entrants every year, indefinitely.
Now look at the qualifying route. If time to competence is three years and the development route can carry 25 supervised places a year, the pipeline produces 25 qualified people a year and supports a steady state of 100 positions, which is 25 percent. Hiring 60 a year against a route that can qualify 25 puts 35 people a year into positions the system has no way of bringing to competence, and those 35 are the attrition that shows up two years later.
The same arithmetic shows where the movement is available. Raising average tenure from four years to seven cuts the required intake from 60 to about 34 without touching the training route at all. Retention work is usually cheaper than training capacity and it acts on the same stock.
What limits the training route is worth naming precisely. In most organisations it is the number of people qualified to supervise and sign off, drawn from the same senior technical population the target is trying to grow. Money and classroom seats are rarely the binding item. That circularity is why these programmes are slow: the capacity to produce competence is made of competence, and drawing it down to hit an intake number reduces next year's capacity to produce any.
Rig time, simulator hours, plant access during turnaround windows, and a supervisor's willingness to hand the controls to a trainee on a live unit are the other real constraints, and none appear in a recruitment plan.
Which roles are genuinely constrained
Before deciding what to build, map what is actually hard. Two variables sort most of it.
Time to competence. Measure it as elapsed time from entry to independent sign-off, from your own records rather than from a job description. Roles cluster cleanly: some are weeks, some one to two years, and some five years or more, unshortenable because they need exposure to events that happen a few times a year.
Domestic supply. How many people enter the national labour market each year holding the entry qualification for that discipline, and how many are realistically available to you rather than to every other employer chasing the same graduates.
The four combinations want different plans. Short time to competence with adequate supply is where the ratio gets carried, and those roles should be filled fast and completely because they were never the problem. Short time to competence with thin supply is an attraction and education problem with a two to four year lag before it resolves. Long time to competence with adequate supply is a training throughput problem, where the pipeline arithmetic above does most of its work. Long time to competence with thin supply will not move inside any target horizon, and it is where the parallel routes below belong.
There is a fifth category that decides more targets than the other four combined, and it is definitional. Contracted work sits inside the operation and frequently outside the headcount denominator. Whether a maintenance technician on a five year service contract counts toward your ratio depends on contract structure and on how the target was drafted, and organisations regularly discover late that a large share of the technical work they depend on was never in scope.
The headcount ratio and the capability ratio
A headcount ratio is nationals divided by total positions. It is the number in the commitment and it takes ten minutes to compute.
A capability ratio asks what share of the positions requiring independent technical judgement are held by nationals who exercise it. Approximate it by weighting positions by grade, by delegated technical authority, or by whether the position holds a sign-off that stops work when withheld.
The two diverge for a structural reason. National hiring concentrates where time to competence is shortest, which is the bottom of the technical grade structure, so the headcount ratio climbs while the layer carrying technical judgement stays roughly where it was.
The diagnostic takes an afternoon. Plot national share by grade band, from entry through the most senior technical grades, and look at the shape. A flat line means the headline ratio reflects capability. A curve falling away steeply above the middle grades means the number is carried by the bottom of the organisation, and the senior positions are held by people whose retirement or rotation dates you can already read off a spreadsheet.
Reporting both numbers changes the conversation, because the gap between them is a forecast. Where the capability ratio at senior grades sits well below the headcount ratio, the headcount ratio will fall when the current senior cohort turns over, and the size of that future fall is visible today.
Buying capability as a parallel route
Training builds capability at the rate the training system permits. Acquiring it runs on a different clock.
Joint ventures with an established operator or manufacturer, technology transfer obligations written into major contracts, secondment arrangements, and outright acquisition of a capable company all move capability faster than a development programme can. They also fail routinely, and the failures share one pattern: the obligation was written as an input.
A transfer commitment expressed as training hours delivered gets satisfied with training hours delivered. A commitment expressed as named positions transferred to named national holders by a stated date, with competence signed off by someone accountable for the work rather than for the contract, is enforceable and much harder to satisfy on paper. Clauses worth writing specify which positions, at what standard, verified by whom, by when, and what happens commercially if that date passes without the sign-off. Where the partner is also the party assessing competence, the assessment is worth moving somewhere else.
The learning curve literature that begins with T.P. Wright's 1936 paper on airframe production costs describes improvement as a function of cumulative units produced. Individual competence has the same shape, accruing with repetitions completed under conditions that matter. A transfer arrangement producing attendance without repetitions produces very little, whatever the hours recorded against it.
Which of the candidates generated this way to fund is a constrained portfolio problem covered separately in this series.
Measure retention at eighteen months rather than hires
Hires are an input and the easiest thing in the whole system to move, which is why targets built on them get met and then unwind.
Retention at eighteen months, by cohort and by role family, tells you whether the hiring produced anything. Eighteen months is long enough that people placed into roles they could not do have mostly gone, and short enough to give a signal while the cohort behind them is still in training.
Four measures alongside it cover most of what a quota misses. Time from entry to first independent sign-off, tracked as a distribution rather than an average, because the tail is where the problem lives. The share of positions carrying technical authority held by national staff, which is the capability ratio above. The number of critical positions with a named successor at a stated readiness date, which turns succession from an assertion into a schedule. And the count of positions reclassified downward in the last twelve months, which catches the paper route to the target and will not be a welcome number to publish.
The limit
Time to competence in specialised technical disciplines is measured in years, and no planning approach compresses it. A reservoir engineer, a senior process safety engineer or an experienced marine operations superintendent gets there through exposure to events that occur a handful of times a year, and nothing supplies those events faster.
The direct consequence is that a target with a short horizon over a population containing many such roles will be met on paper. The mechanisms are all cheaper than the real thing: reclassifying positions downward, redefining the denominator, moving work to contractors outside the count, and awarding titles ahead of competence. Anyone under pressure to report a number has access to all four.
A second limit sits on the measurement. Attrition to another domestic employer is a success for the national programme and a loss for you, so the same event scores differently depending on whose books are open. An organisation that gets good at producing competent national technical staff will lose some of them to competitors who never invested, and reading that as programme failure leads to the wrong response.
And the data usually is not there. HR systems record hires, leavers and grades; very few record time to independent sign-off, which happens in an operating system or on paper in a supervisor's file. Building that record is a prerequisite for everything above and takes a full cycle to accumulate.
Pull the last three intake cohorts and compute retention at eighteen months by role family; the families where it is worst are where the quota has been running ahead of the pipeline.