An internal candidate and an external candidate for the same role are almost never judged on the same evidence. The internal is judged on performance ratings, tenure and manager advocacy; the external on interviews and tests. Those evidence types differ in reliability and measure different things, so one shortlist is not a comparison.
That is the measurement problem sitting underneath every internal talent marketplace, every skills-based mobility programme and every "build versus buy" debate. It is worth separating from the business question, because the outcome data on build-versus-buy is reasonably clear while the measurement practice is mostly not.
What the outcome data actually shows
The most-cited study on this is Matthew Bidwell's analysis of the US investment banking arm of a financial services firm, covering 5,260 workers across 7,129 job spells between 2003 and 2009, published in Administrative Science Quarterly. Compared with people promoted into the same jobs, external hires:
- were paid around 18% more in base salary on entry
- showed significantly lower performance on all three performance measures, with the gap persisting through the first two years and converging over about three
- had a 61% higher hazard of involuntary exit and a 21% higher hazard of voluntary exit
The paper's own explanation is not that external hires are worse people. It is specific skills plus incomplete information: internal movers arrive already holding firm-specific knowledge, and the organisation already has years of observation on them. External candidates have neither advantage, and the pay premium is partly compensation for the risk both sides are taking.
Read the full paper rather than the summary (Bidwell, 2011, journal record) and note the limits: one firm, one industry, one period, before the current era of remote and cross-border hiring. It is strong evidence that internal moves can be underrated. It is not a universal law.
Why the internal case is weaker than it looks
If internal candidates arrive with real advantages, the obvious conclusion is "promote more". The problem is the evidence organisations actually use to decide which internal candidate.
Three findings should make anyone cautious about track record as the primary input:
Sales performance predicts promotion, and predicts worse management. Benson, Li and Shue analysed 214 firms, 156 million sales transactions and over 1,500 promotions into management. A doubling of a salesperson's sales credits raised their promotion probability by 14.3% relative to base rate. A doubling of a new manager's pre-promotion sales was associated with a 7.5% decline in the sales performance of their subordinates (NBER summary; Quarterly Journal of Economics, 2019).
The pattern replicates where performance is measured objectively. A 2026 longitudinal study of all 329 NBA head coaches from 1947 to 2020, including the 157 with prior NBA playing careers, found prior playing performance predicted later coaching success only weakly and stably — standardised coefficients of roughly 0.13, 0.12 and 0.11 across the first three coached seasons, with no strengthening over time (Schleu et al., 2026).
Tenure is close to noise. In the revised meta-analytic estimates in Sackett and colleagues' update to the personnel selection literature, years of job experience carries an operational validity of about .07, against .42 for structured interviews, .40 for job knowledge tests and .33 for work samples (author copy, *Journal of Applied Psychology*, 2022).
None of this says internal candidates are a bad bet. It says that "strong performer in the current role, five years in the business, manager speaks highly of them" is a weak instrument for predicting performance in a different role — and it is the instrument most internal decisions run on.
The two evidence bases, side by side
What you typically hold on an internal candidate
- Multi-year supervisory ratings, usually from one or two raters, against a framework not designed for the target role
- Observed behaviour in the current role, which may share few requirements with the target role
- Manager endorsement, which is advocacy rather than measurement and is unevenly distributed across teams
- Tenure and project history
What you typically hold on an external candidate
- A structured or semi-structured interview against the target role's requirements
- Possibly a test or work sample, scored against a defined standard
- Reference checks of unknown reliability
- No firm-specific knowledge, and no observed behaviour inside your context
The internal file is longer and deeper but almost entirely about a different job. The external file is thinner but aimed directly at the job being filled. Combining them on one shortlist without deciding what the common evidence is means the decision gets made on whichever file the panel finds more vivid — and a colleague they know is always more vivid than a stranger with a score.
How to make the two pools comparable
- Write the target role's requirements before you look at either pool. Four to six requirements, each stated as observable behaviour at a defined level. If a requirement cannot be observed, it cannot be assessed.
- Define a common core of evidence that both pools complete. A structured interview on the same guide, plus one role-referenced work sample or scenario set. Both pools, same instrument, same rubric.
- Treat track record as additional evidence, not substitute evidence. Internal candidates keep the advantage of a known record; they do not get to skip the common core because of it.
- Score against a defined standard, not against the pool. A criterion-referenced cut lets you conclude "nobody in either pool is ready", which a ranking cannot.
- Calibrate the panel, and keep it the same across pools. The panel that interviews internal candidates on Tuesday and externals on Thursday is not the same instrument twice unless it has worked through the rubric on anchor examples.
- Run the adverse impact check on promotion, not just hiring. Under the Uniform Guidelines, a selection procedure is any measure used as a basis for an employment decision, and employment decisions explicitly include promotion. Internal selection sits squarely inside scope; so does a manager-nomination step.
- Record the decision. Requirements, instruments, scores, standard, who decided, and the impact figures. The *Standards for Educational and Psychological Testing* expect the validity argument to be documented for the use being made of the scores — and promotion is a different use from hiring, even with the same test.
Running step two is the part most organisations skip on cost grounds, which is usually a false economy: a short role-referenced scenario set across both pools costs a fraction of one mis-promotion. Pay-as-you-go assessment credits exist partly for this case — AssessAll's are ₹30 / US$0.50 per assessment, with AI-graded scenarios scored on the same rubric whichever pool a candidate comes from, so the internal cohort is not excluded by budget.
When each approach is genuinely right
Hire externally when the target role's core requirements are largely absent inside the organisation, when the role demands a deliberate break from current practice, or when the internal pool is too small for a real comparison. Bidwell's performance gap is an average over a pool of external hires, not a prediction about a specific candidate.
Promote internally when firm-specific knowledge is a large share of the job, when the requirements overlap substantially with roles you can already observe, or when the development value of the move is itself part of the return. Just make the choice on role-referenced evidence rather than on the comfort of familiarity.
The takeaway
The build-versus-buy argument is usually conducted as a debate about which pool is better, when the real defect is that the two pools are measured on non-comparable evidence. Define the target role first, make both pools complete the same core assessment, and let the internal candidate's track record add to that evidence rather than replace it.