Topic 32 o 64· Governmental performance an delivery metrics

Metrics for trust an legitimacy

Legitimacy an support is yin o the three legs o Mark Moore's "strategic triangle" in Creating Public Value (1995) — alangside public value itsel an operational capacity — an it is the leg maist aften left unmeisurit, acause unlike a budget or an ootpit coont, legitimacy haes nae obvious single nummer attached tae it. Trust an legitimacy metrics are the faimily o proxy meisures governments uise tae fill that gap: institutional trust surveys, oversicht-body confidence ratings, complaint an appeals data, an political/legislative support indicators.

Why it maitters

Moore's argiment is that a public manager wha delivers real value but loses political an public legitimacy will in the hinner lose the authorisin environment needit tae keep deliverin it — fundin gets cut, mandates get narrowit, an the service is stairvit nae maitter hou guid its ootcomes are. Legitimacy is thairfore nae a public-relations afterthocht bolted ontae a delivery scorecard; it is a load-bearin input tae whether the mission can continue ava, that is why it sits as a co-equal perspective in a public value scorecard raither nor a fitnote. The OECD's "Trust in Government" survey programme is the leadin cross-national attempt tae quantify this: it tracks the share o citizens athort OECD member states wha say they hae confidence in their national government, an its lang-run data shaws trust is highly sensitive tae shocks — baith the 2008 financial crisis an the COVID-19 pandemic produced sharp national-level swings, aften follaed bi ainly pairtial recovery, wi the OECD's analysis consistently findin that perceived competence (daes government deliver whit it says it will) an perceived fairness/integrity (is government seen tae act athoot corruption or favouritism) are the twa strangest drivers o the trust figure, distinct frae satisfaction wi ony single transaction. Governments increasingly try tae operationalise legitimacy at a mair granular level an aw — the UK's independent regulators an inspectorates (the National Audit Office, the Parliamentary and Health Service Ombudsman, sector regulators like Ofsted an the Care Quality Commission) function as institutionalised legitimacy checks, convertin "daes the public still trust this service" intae auditable ratings.

The maths

Trust an legitimacy is a framewark-shapit topic whase uisable quantitative proxies are:

Institutional trust index (OECD-style)
  = % o survey respondents answerin "aye" tae a confidence-in-government question,
    trackit ower time, disaggregatit bi demographic group

Legitimacy proxy set (nae single nummer substitutes for the construct):
  - Upheld complaints per 1,000 service uisers (ombudsman or internal complaints data)
  - Judicial review / appeals success rate agin the body's decisions
  - Independent regulator/inspectorate rating (e.g. "outstandin" tae "inadequate" bands)
  - Legislative/oversicht committee confidence votes or critical report frequency
  - Freedom o information request volume an disclosure/refusal rate, as a proxy
    for perceivit transparency

Legitimacy is corroboratit, nae calculatit: a defensible legitimacy assessment
triangulates several o the abuin raither nor relyin on ony single proxy.

Wrocht example

National tax authority: legitimacy triangulation for an annual public value report.

OECD-style trust proxy (depairtment-specific confidence survey):
  58% o respondents say they trust the authority tae "treat me fairly" (doun frae
  64% twa year afore)

Complaints data:
  Upheld complaints: 4.2 per 1,000 taxpayer interactions (up frae 3.1 per 1,000)

Ombudsman referrals:
  Referrals tae the independent Adjudicator's Office: 1,850 in the year, o that
  61% upheld in full or pairt agin the authority (up frae 48% the prior year)

Readin athort aw three: trust is fawin, upheld complaints are risin, an
independent ombudsman findins are increasingly sidin agin the authority —
three independent signals convergin on the same direction, that is whit mak
this a credible legitimacy finding raither nor noise in ony yin series.

A single yin o thir figures muivin wad be waik evidence; three independent meisures muivin thegither ower the same period is the pattern that mak a legitimacy claim defensible.

Legitimacy metrics are seldom produced bi a single team's dashboard, that is itsel the design lesson: big reportin pipelines that can ingest an reconcile data frae independent external sources (ombudsman casewark systems, regulator ratings feeds, survey vendors) raither nor architectin legitimacy reportin as an internal-only metric, acause internally-sourced legitimacy claims ("we rate oorsels as trustworthy") cairry little evidential weicht — the same independence problem notit for the legitimacy perspective in a public value scorecard. Complaints an appeals data pipelines deserve the same data-quality rigour as ony ootcome pipeline feedin peyment bi results contracts, sin an unner-reportit or puirly-categorisit complaints dataset quietly unnerstates a legitimacy problem afore it becomes visible in a trust survey a year efter. See citizen satisfaction metrics for the transaction-level coonterpairt tae this institution-level meisure, an public value for Moore's full strategic-triangle framewark this leg belangs tae.

Pitfaws

  • Treatin satisfaction as a proxy for legitimacy: a citizen can be satisfied wi a single transaction's interface while distrustin the institution owerall (or vice versa) — see citizen satisfaction metrics for why the twa maun be reportit separately.
  • Relyin on a single self-reportit metric: an internally-rin trust survey wi nae independent corroboration (ombudsman data, regulator ratings) is easy tae dismiss as self-markin; triangulate.
  • Ignorin demographic disaggregation: aggregate national trust figures can mask sharply divergin legitimacy amang specific groups (bi age, ethnicity, income, or region) — the OECD's ain Trust in Government releases disaggregate for exactly this reason.
  • Readin a single shock-driven dip as a permanent trend: trust figures muive sharply roond crises (financial crashes, pandemics, high-profile scandals) an pairtially recover; a single post-shock data pynt shuidnae be extrapolatit intae a lang-run decline athoot mair data.

Sources