Risks & Safeguards
21 identified risks with safeguards, residual ratings, and monitoring triggers. Every risk has an accountable owner and defined escalation path.
Residual Risk Profile
delivery Risks
delivery
Quarterly Prime Ministerial chairmanship is not sustained beyond the first year, and the Task Force decays into an officials' committee with no convening power - the exact failure mode that defeated the MIB Implementation Unit and MITRA
Safeguard
Chairmanship, quorum and meeting cadence written into the Cabinet decision and terms of reference; a standing calendar fixed twelve months ahead; automatic publication of the dashboard within 14 days whether or not the meeting is held, so non-convening is itself visible; a named Deputy Chair to preserve continuity without diluting authority
Monitoring Trigger
Any quarter in which the meeting is not held or is not chaired as constituted
delivery
Delivery capacity in the lead ministries is already committed, and MIB 2.0 competes with existing priorities rather than being additive
Safeguard
Every programme runs on an existing delivery agency rather than creating a new one; only the delivery secretariat and the navigator cadre are new establishment; programme volumes are deliberately capped below the source drafts' ambitions; phase gating allows scope reduction without failure
Monitoring Trigger
Any lead ministry reporting inability to resource its programme at the quarterly review
legal Risks
legal
Ethnically-defined eligibility for public benefits is challenged under Article 8 of the Federal Constitution, with the procurement and quota elements most exposed
Safeguard
The ethnic procurement set-aside and participation quota are REMOVED (CNF-024); eligibility for every material benefit is defined on NEEDS-BASED criteria (household income, documentation status, tenure insecurity) with outreach - not entitlement - targeted to the community; Attorney-General's Chambers clearance is sought before any instrument is issued (VAL-01)
Monitoring Trigger
Any drafting instruction that reintroduces ethnicity as an eligibility condition
constitutional Risks
constitutional
Federal programmes encroach on state jurisdiction over land, local government and non-Islamic religious institutions
Safeguard
Participation by state consent only, facility by facility and project by project; federal funding confined to safety, accessibility and community-service functions; land tenure findings documented for referral to state authorities rather than determined federally; a standing item at the National Land Council or an equivalent federal-state forum
Monitoring Trigger
Any state declining to participate or any land determination attempted federally
fiscal Risks
fiscal
The incremental new funding requirement of approximately RM820 million over six years is not appropriated, or is appropriated only in part, leaving Phase 2 and 3 commitments unfunded
Safeguard
Phase gating: Phase 2 scope is authorised only after the Phase 1 independent evaluation; no Phase 3 commitment is entered before Year 5 appropriation; each phase is separately costed so a phase can be scaled without stranding earlier investment; conservative scenario (RM1,188m) presented as the fallback operating envelope
Monitoring Trigger
Any appropriation below the phase envelope in any supply year
fiscal
Existing-funding assumptions prove wrong - in particular the assumed 55% existing-funding share for SJKT works - and the true incremental requirement is materially higher than RM820 million
Safeguard
Every existing-funding share is declared in COSTING_ASSUMPTIONS.csv and flagged as provisional; MOE confirmation of the SJKT share of the national dilapidated-schools programme is a gating validation item (VAL-11); the model is rebuilt from assumptions by script so any correction reprices the whole portfolio in one step
Monitoring Trigger
MOE or MOF confirmation differing from the assumed share by more than 10 percentage points
fiscal
RM246.456m of the portfolio is classed development expenditure (PRG-03, PRG-04, PRG-12). In Malaysian practice development expenditure is appropriated through the Malaysia Plan rolling-plan process administered by the Ministry of Economy, not by a Cabinet decision plus annual supply. A six-year plan commencing 2026 straddles two Malaysia Plans
Safeguard
The Ministry of Economy must be engaged as a co-lead on the development-expenditure component and the requirement entered into the applicable rolling plan; the fiscal framework decision (D5) is explicitly framed as approval in principle only; the Public Finance and Fiscal Responsibility Act 2023 compliance position must be established before any formal submission
Monitoring Trigger
Any development-expenditure line not entered in the applicable Malaysia Plan rolling plan
fiscal
The stated defence to the “why only one community” objection is that the delivery model is extensible to any community facing similar structural barriers. That defence creates an unpriced contingent liability if Cabinet extends it
Safeguard
The extensibility claim is stated as a property of the DELIVERY MODEL (front door, family profile, navigator, pathway, dashboard) and expressly not as a commitment to replicate the programme costs for any other community; any extension is a separate Cabinet decision with its own costing; no extension is costed or implied in this envelope
Monitoring Trigger
Any proposal to extend the programme to another community without a separate costing
data protection Risks
data_protection
Household and documentation records are linked or misused, chilling take-up among precisely the population the plan targets
Safeguard
The documentation caseload system and the household progression record are SEPARATE systems with no automatic linkage (CNF-032); enrolment in the progression record is voluntary and consent-based; a Data Protection Impact Assessment is a Phase 1 gating deliverable and must be published before any household record operates; access logging and periodic audit
Monitoring Trigger
Any proposal to link the two systems or to use documentation status in benefit targeting
political Risks
political
The plan is characterised as ethnic preferential treatment, damaging both its prospects and inter-community relations
Safeguard
Needs-based eligibility throughout; explicit disclosure that Indian absolute poverty (5.4%, 2022) is BELOW the Bumiputera rate (7.9%), so the case is made on structural exclusion and adverse trend, not on relative deprivation; the delivery model is explicitly designed to be extensible to any community facing similar structural barriers; all electoral and partisan content removed (CNF-019, CNF-030)
Monitoring Trigger
Public characterisation of the plan as a quota or preference programme
political
A six-year horizon necessarily spans at least one general election. The plan’s central mechanism is personal Prime Ministerial chairmanship, which does not survive a change of Prime Minister or administration by design
Safeguard
Chairmanship and cadence written into the Cabinet decision rather than into a personal undertaking; multi-partisan membership so the body is not owned by one administration; every deliverable that can be completed within Phase 1 is front-loaded so that a discontinuity does not strand incomplete work; the independent evaluation and Auditor-General audit continue irrespective of chairmanship
Monitoring Trigger
Any change of Prime Minister, Cabinet or machinery of government
measurement Risks
measurement
Eleven of sixteen KPIs have no baseline, so the plan could report activity for two years without being able to demonstrate outcome change
Safeguard
Baseline establishment is the principal Phase 1 deliverable and is itself a KPI (KPI-16); no numeric outcome target is set where no baseline exists - the Year 2 target is the baseline itself; independent mid-term evaluation at end-Year 3 gates Phase 3
Monitoring Trigger
Any KPI without a published baseline at end of Year 1
programme design Risks
programme_design
The matched-savings instrument (PRG-11) cannot be established because PNB declines to participate or MOF does not approve the design
Safeguard
PNB participation is sought, not assumed; the instrument is product-agnostic and can be delivered through an alternative qualifying savings product; the programme is classed Provisional and its RM200m central cost is explicitly at risk; a decision point is set at end of Year 1
Monitoring Trigger
No agreed product design by end of Year 1
leakage Risks
leakage
Grants and subsidies are diverted, repeating the governance failures that have damaged confidence in community programmes
Safeguard
Payments made to institutions and verified accounts rather than in cash where possible; published recipient lists for all grants above a threshold; annual Auditor-General audit tabled in Parliament; open competitive procurement; immediate referral of irregularities to the MACC; the delivery secretariat holds no programme funds
Monitoring Trigger
Any audit qualification or any grant paid outside published criteria
equity Risks
equity
Support concentrates in accessible urban localities and misses estate-origin, rural and remote populations - the households with the greatest need
Safeguard
Mobile delivery is built into PRG-01 and enrolment outreach into PRG-02; all KPIs are disaggregated by state and, where relevant, by locality type; the quarterly dashboard reports geographic distribution so concentration is visible within one quarter
Monitoring Trigger
Any programme where the top three states exceed their share of the eligible population by more than 20 percentage points
exclusion Risks
exclusion
Needs-based eligibility using documented household income excludes informal and undocumented households, who are the most vulnerable
Safeguard
Alternative means-assessment pathways for households without formal income documentation, including community verification and social-welfare officer assessment; documentation status is explicitly NOT a barrier to PRG-01 or PRG-02 participation
Monitoring Trigger
Any programme where the proportion of applicants rejected for want of income documentation exceeds 10%
perverse incentive Risks
perverse_incentive
Poverty graduation targets create pressure to exit households from support before they are stable, producing measured success and real harm
Safeguard
Graduation is defined by movement in two or more domains sustained over 24 months, not by exit from assistance; re-entry is permitted without penalty; the independent evaluation specifically tests for premature exit
Monitoring Trigger
Any rise in re-entry rates or in exits without sustained improvement
double counting Risks
double_counting
Reach figures are summed across programmes, overstating unique beneficiaries in public reporting
Safeguard
Reach and unique beneficiaries are modelled separately with a stated overlap factor (ASM-030); the machine-verification script tests for duplicate programme-beneficiary combinations; reach figures are never summed across programmes in the proposal or dashboard
Monitoring Trigger
Any published figure that sums reach across programmes
evidence Risks
evidence
The plan's own baselines rest on 2014-vintage data and provisional derivations, and re-estimation materially changes programme scale
Safeguard
All 2014-vintage figures are labelled by year and are not used for present-day statements; DOSM re-estimation of the Indian low-income household count and child cohort are gating validation items (VAL-02, VAL-09); the costing model rebuilds from assumptions by script, so a corrected population reprices the portfolio deterministically
Monitoring Trigger
Any DOSM re-estimate differing from the planning assumption by more than 20%
procurement Risks
procurement
Capital works under PRG-03 and PRG-04 are let without competitive tender or to related parties
Safeguard
All works follow standard Treasury procurement instructions with no exemption sought; award lists published; the prioritised remediation schedule is published in advance so departures from audited priority order are visible
Monitoring Trigger
Any award outside the published priority order or any direct negotiation
reputational Risks
reputational
Publishing disaggregated ethnic data on civil service composition, school condition or poverty produces adverse comparison and political friction
Safeguard
Publication is committed to in advance as a condition of the plan, so selective disclosure is not available; data are published in full national context rather than for one community alone; the reporting framework is described as a model extensible to any community
Monitoring Trigger
Any request to withhold a scheduled publication