Mega-Project, Mega-Irregularity

The full numbers

The Royal Audit Authority’s Annual Audit Report 2024–25 (Volume I) records Nu 9,987 million in audit observations. Of that:

• Punatsangchhu-II Hydropower Authority (PHPA-II): Nu 7,790 million (78% of the total)
• All other agencies, ministries, autonomous bodies, corporations, and projects combined: Nu 2,197 million (22%)

PHPA-II is a single 1,020 MW hydropower project on the Punatsangchhu river. It was approved in 2008, commissioned in December 2024 (the original target was 2017), and has now been under construction for 17 years — most of the working career of every engineer who joined the project at graduation (see paradox #3). The specific FY 2024–25 PHPA-II observations include Finance & Revenue Management failures, incorrect steel-price escalation indices producing Nu 99.5M and Nu 26.3M of overpayments to construction firms, and unresolved contractor variations against documentation that the audit team could not fully reconcile.

This is the Chapter 10 angle on the same data that paradox #40 (“Eight billion ngultrum hiding in one project”) frames from a different direction. Paradox #40 says: the irregularity number has grown 25× since 2017, and the integrity-score survey hasn’t moved — that is a story about what gets measured. This paradox says: the mega-project absorbs the bulk of the institutional attention available for irregularities of all kinds, and the second-order audit-and-compliance apparatus to monitor a multi-decade build at that scale has not yet been built — that is a story about what gets supervised.

Imagine this

A senior Royal Audit Authority auditor closes the FY 2024–25 file in March 2026 and opens the next year’s work plan. PHPA-II alone has 17 years of construction history, four prime contractors across the lifecycle, an Indian Government-financed cost base that runs into the tens of billions of ngultrum, multiple cost-revision protocols negotiated with Delhi, and a steel-price-escalation formula that — by the auditor’s own work — has produced over Nu 125M in overpayments in a single fiscal year. The audit team assigned to PHPA-II is six people. The audit team assigned to every other agency in the country is roughly forty.

The arithmetic is unforgiving. If six auditors produce Nu 7,790M of flagged observations on one project in one year, the ratio of audit-attention-per-rupee-of-spend on PHPA-II is structurally lower than the ratio on a small dzongkhag office that gets visited once every three years. The auditor is not finding more irregularity in PHPA-II because PHPA-II is uniquely problematic. She is finding more because there is uniquely more to find — and because every comparable project anywhere in the world produces comparable findings when the audit team finally gets through the books.

The follow-up cycle is the second bottleneck. The flagged observations sit in the report. The supervisory follow-up that would compress them — the parliamentary Public Accounts Committee hearing, the OAG’s review of any criminal-accountability cases, the corrective-action plan from PHPA-II management itself — runs on a tempo measured in years. The cases age. The construction phase ends. The operational phase begins. The irregularity sits in the file, no longer the subject of a live decision.

Where this came from

Three structural facts compounded:

1. The project was designed at Crown scale; the oversight architecture was not. PHPA-II is a Crown-level strategic asset — a multi-decade, multi-billion-dollar, bilateral G2G undertaking with India that ranks alongside Tala, Mangdechhu, and Punatsangchhu-I in the country’s strategic-energy portfolio. The strategic-planning office that articulated the project (Gross National Happiness Commission, Ministry of Energy & Natural Resources, DGPC predecessor structures) operates at the scale appropriate to a national-level capital project. The audit-and-supervisory office that would close the loop on a project of that magnitude — RAA, ACC, OAG, the parliamentary committees, the project-specific oversight committees — has not been resourced at the same scale. The mismatch is structural, not malicious.
2. Mega-projects globally produce mega-irregularities — but in mature jurisdictions, the supervisory apparatus catches up. Hinkley Point C (UK) had cost-overrun audit cases through the entire 2016–2024 construction period; the UK National Audit Office maintained a permanent Hinkley-specific desk. Sasan UMPP (India) generated multiple CAG reports across its construction window; the CAG followed the project from approval through commissioning. Kemano (Canada) had a dedicated provincial-auditor unit for the full build. Bhutan has not yet built the equivalent dedicated PHPA-II audit-and-resolution desk that would close findings within the project’s construction window rather than after it ends.
3. The criminal-accountability cascade has slowed. The PHPA-II case load includes contractor-overpayment matters that would, in a mature jurisdiction, move from RAA finding → OAG referral → ACC investigation → prosecution within 24–36 months. Several cases referenced in the 2024–25 audit file have origins going back to 2018–2020 audit cycles. The construction phase ended in December 2024. The criminal-accountability phase has not closed.

Why this matters now

PHPA-II is one project. The pipeline includes Kholongchhu (600 MW, currently negotiating revival), Wangchhu (570 MW), Bunakha (180 MW), Chamkharchhu-I (770 MW), Dorjilung (1,125 MW), and Sankosh (2,560 MW) — projects together representing roughly Nu 800 billion of capital expenditure over the next 15–25 years. If the audit-and-compliance apparatus that handled PHPA-II at the PHPA-II ratio is scaled forward to the full pipeline, the FY 2024–25 Nu 7,790M irregularity finding in PHPA-II is not a one-off — it is a forecast of what the next decade looks like at higher scale, unless the second-order infrastructure is built deliberately.

The window to build it is now, before Sankosh and Chamkharchhu-I move from DPR into construction.

What it should be

• A dedicated mega-project audit desk inside RAA with continuous in-construction monitoring (six-month flagging cycle, not annual), modelled on the UK NAO Hinkley desk
• A project-specific Public Accounts sub-committee in Parliament with standing authority to call mega-project management at any quarterly milestone
• A statutory deadline for OAG referrals of RAA findings — e.g., 24 months from flag to disposition (referral or closure), tracked publicly
• Cost-escalation formulas and contractor-variation protocols pre-audited at the DPR stage, not after the variations have already been paid
• A published mega-project audit dashboard showing flagged-vs-resolved trajectories for each major project, refreshed quarterly

How others do it

• United Kingdom (Hinkley Point C) — UK National Audit Office maintains a dedicated Hinkley desk; published five major value-for-money reports across 2016–2024; cost-overrun findings discussed in House of Commons within months of issuance
• India (Sasan UMPP and other mega-projects) — CAG follows projects from approval through commissioning; project-specific audit reports laid before Parliament; criminal-accountability referrals to CBI move on defined timelines
• Canada (Kemano completion) — British Columbia Auditor General maintained a dedicated project unit through the construction window; findings folded into provincial fiscal reporting in real time
• Norway (Mongstad CO2 capture, oil-platform overruns) — Office of the Auditor General reports tabled in the Storting within months of audit completion; mega-project findings drive parliamentary investigative committees with named accountability
• Bhutan: PHPA-II construction completed December 2024; the FY 2024–25 audit findings sit in the RAA report; supervisory follow-up runs on a multi-year tempo; criminal-accountability cases referenced in the audit file have origins going back to 2018

The question we should be sitting with

If a single mega-project produces more flagged irregularity than the rest of the public sector combined, the question is not whether the project is uniquely problematic. The question is whether the audit-and-supervisory apparatus that would catch up to a project of that scale has been built. And if it has not been built for PHPA-II, what does that imply about the eight comparable projects in the pipeline?