Contact us

Civic Questions System | EP22 Full Worked Example

You published the books.
I ran the numbers. They don't add up.

Four ledgers Pingtung County Government published itself: its administrative plan, its procurement notices, a summer work-study listing and its final accounts. We read every entry and worked every figure with the same method behind AAI, and laid out the raw numbers, the formulas and the sources. If you have to speak in council questioning or budget review, open it and use it.

See the four ledgers checked Start with a 30-minute call
  1. 01

    NT$4,833,277

    Four contracts in four years, one winner, and only one bidder every year

  2. 02

    1 of 117 indicators

    Across three departments' performance indicators, only one measures outcomes

  3. 03

    A14-07

    Demands “professional skills and experience”, posts a summer work-study job

  4. 04

    88 of every NT$100

    Of every NT$100 coming in, NT$88 is handed down by the central government

01｜Show of force

Four ledgers, checked line by line

Every card follows the same order: raw numbers, formula, what doesn’t add up, sources. We use only official source text and final accounts audited by the National Audit Office, so the other side can open every document and run the same numbers.

Ledger 01｜Procurement notices｜Government e-Procurement System

Pingtung Qinghang Youth Station, four contracts: NT$4,833,277 over four years, one winner, and only one bidder every year

Procuring entity: Pingtung County Government. Handled by the Youth Affairs Section, Department of Labor and Youth Development. All four were below the publication threshold, openly solicited price quotations or written proposals, followed the spirit of the most advantageous tender, and set no base price.

Raw numbers (award notices, verbatim)

Year (ROC)Case No.Award amountBudgetBidders
112E1120710-11,200,0001,200,0001
113E1130801-1680,000680,0001
114E1140613-11,476,9771,476,9771
115E1150316-31,476,3001,476,3001

Amounts in NT$. Years follow Taiwan’s ROC calendar (112 = 2023, 115 = 2026). The same company won all four.

Formula

1,200,000 + 680,000 + 1,476,977 + 1,476,300 = NT$4,833,277

Award amount − budget = NT$0 (all four)

Bidders: 1 + 1 + 1 + 1, 1 bidder for every contract, four years running

What doesn’t add up

0 lines of scope of work

  • None of the four notices lists the work to be performed. All they say is “For questions about the scope of services and related matters, contact the Youth Affairs Section of the Department of Labor and Youth Development directly.”
  • A proposal competition, four years running, and every year exactly one proposal. Competing against whom?
  • What the NT$4,833,277 paid for, and whether it covered video, cannot be found in the notices.

Sources

Ledger 02｜Administrative plan｜Pingtung County Government FY2026 Administrative Plan

Labor and Youth Development, Cultural Affairs, Communication: 117 indicators, 1 that measures outcomes

Each indicator is classified by the unit of measure in its “measurement standard” column. Output indicators measure whether something was done. Outcome indicators measure whether it worked.

Raw numbers (performance indicators in the plan)

DepartmentIndicatorsOutputOutcomeMixed / unclear
Department of Labor and Youth Development565312
Cultural Affairs Department484107
Department of Communication and International Affairs131300
Total11710719

Communication department, verbatim: “Produce 12 videos” (PDF p. 79, printed p. 75); “Write 1,200 or more press releases for this government” (PDF p. 78, printed p. 74).

Formula

Output: 107 ÷ 117 = 91.5%

Outcome: 1 ÷ 117 = 0.9%

Count all 9 unclear ones as output: 116 ÷ 117 = 99.1%. Count them all as outcome: output is still 91.5%.

What doesn’t add up

0 indicators for views, employment, retention or satisfaction

  • “Produce 12 videos”: no word on who they are for, no word on how many people watch. Shoot twelve and the target is met. A video with zero views still counts as one.
  • The only outcome indicator is, verbatim, “the quota employment requirement under Article 38 of the People with Disabilities Rights Protection Act”. Its measure is “number of obligated agencies (institutions) meeting the quota”, which counts how many agencies hired as the law requires.
  • Nobody watches the videos, so how does the job count as done? That question isn’t on the exam.

Sources

Ledger 03｜Work-study listing｜2025 Public Sector Summer Work-Study Program for College Students, recruitment guide

Hao Ping Zhi Bo (Pingtung career-video series): the post demands “professional skills and experience”, and it is a summer work-study job

Hao Ping Zhi Bo, a video series on the county government’s youth channel, was shot for its first three years by summer work-study students at the Department of Labor and Youth Development. The students did nothing wrong. Our numbers are about the side that posted the job.

Raw numbers (recruitment guide, p. 23, verbatim)

  • Job No. A14-07, “Hao Ping Zhi Bo: video production for summer work-study highlight programs”. 4 openings. Hiring unit: Department of Labor and Youth Development.
  • Duties: “Discover outstanding public-sector summer work-study programs, conduct interviews and produce videos.”
  • Requirements: “Must have professional skills and experience in video concept development, filming, editing, poster design, computer graphics, etc.”
  • Back in 2023 and 2024, the video descriptions themselves credited “summer social media interns, Youth Affairs Section, Department of Labor and Youth Development”.

Formula

Hao Ping Zhi Bo: 28 videos, total views = 3,044

2026, EP.01–04: 4 videos, 113 views in total

View counts are exact figures pulled video by video on 2026-10-15.

What doesn’t add up

Professional skills demanded, summer work-study offered

  • They took work-study pay and delivered professional videos. Summer ends, the students leave, the videos stay on the channel.
  • Once the videos are shot, whose job is it to get them seen?
  • Is this giving young people a chance, or using them like disposable chopsticks?

Sources

Ledger 04｜Final accounts｜Figures audited by the National Audit Office, via the key county and city statistics of the Directorate-General of Budget, Accounting and Statistics (DGBAS)

Of every NT$100 coming in, NT$88 comes from the central government. Locally raised taxes haven’t grown in four years. Spending is up almost 30%.

The self-financing ratio uses the National Audit Office’s official definition: (revenue − subsidies and grants received − centrally allocated tax revenue) ÷ revenue. Per-capita figures use registered household population at year end (population source: Ministry of the Interior).

Raw numbers (Pingtung County, FY2021 and FY2025 final accounts)

ItemFY2021FY2025
Self-financing ratio19.76%11.76%
Subsidies + allocated tax, share of revenue80.2%88.2%
Local taxes collected per capitaNT$6,895NT$6,644
Expenditure per capitaNT$57,129NT$73,711

Counties with similar dependence, self-financing ratio FY2021→FY2025: Taitung County −5.41, Yunlin County −3.21, Chiayi County −2.19, Pingtung County −8.00 percentage points.

Formula

19.76 − 11.76 = 8.00 percentage points

6,644 − 6,895 = −NT$251

73,711 ÷ 57,129 − 1 ＝ ＋29.0%

What doesn’t add up

Own taxes per head −NT$251, spending per head +29.0%

  • The national average is falling too: the simple average of 22 counties and cities went from 31.56% to 24.43%, down 7.13 percentage points. On the ratio alone, the county still has a case to make.
  • Local taxes collected per person haven’t grown in four years; they fell by more than NT$200. Spending per person rose from just over NT$57,000 to just over NT$73,000.
  • Relying on the central government is no crime. Dependence deepens every year, and spending keeps climbing. Who reconciles that account?

Sources

02｜Who follows up

The National Audit Office has named it.
The answers belong in the council chamber.

The National Audit Office writes about Pingtung's books every year.

Its audit report on Pingtung County's FY2023 general final accounts covered the county's international baseball stadium: construction costs were high, the engineering budget needed closer review, professional teams should be brought into the design, and the search for an operator should start early. "A letter has been sent requesting review and improvement."

The FY2024 report said it again: preliminary design work was delayed, and that year's budget execution rate was 7.05%. "A letter has been sent requesting concrete improvement measures and prompt action."

The FY2025 audit report on the special final accounts for Phase 5 of the Forward-Looking Infrastructure Development Program covered Pingtung's flood control: of 92 regional drainage channels, 90 have no completed management plan; of 28 stormwater sewer planning reports, 19 have gone more than ten years without review.

Name it, send a letter, write it again the next year. Once the audit is done, the report written and the letter sent, the procedure is complete. Whether the county fixed anything only shows up in the next report.

Where do these reports go? The report says so itself. The flowchart in the preface of the FY2023 report reads: "Included in the final accounts audit report and submitted to the Pingtung County Council and the Control Yuan for deliberation." Every line the National Audit Office writes goes to the council. The body that can make the county government stand at the podium and answer during budget review is the council.

Once a report reaches the council, someone has to open it, run the numbers and ask the question on the floor. The four ledgers above were worked out exactly that way: the question, the formula, the source, the county's likely explanation and the next follow-up, on your desk before the session starts. This method is built for people who have to speak in the council chamber.

Naming it: the National Audit Office does that every year.
Following up: that takes someone speaking in the council chamber.

This page is what we sell

The four cards above are what one method produced on Pingtung County Government’s public data: read the source first, compute entry by entry, state the other side’s most reasonable explanation up front, and keep only what that explanation cannot account for. The same method can work for your district.

  • Issue Search｜Finding the questions

    Every day we go through procurement, budget, final-accounts and subsidy data to find the questions worth asking, each with its original source and page number. That is how the four ledgers on this page were found.

    See the service →

  • AI Council Questions｜AI-assisted questioning scripts

    We turn questions into lines of questioning and council scripts: the facts, the sources, the government’s likely explanation and the next follow-up. We can also turn them into video scripts and caption cards like this episode’s.

    See the service →

  • AI Constituent Services｜AI petition triage

    Petitions and inquiries reaching your office are sorted automatically, matched with the relevant regulations and public data, and tracked one by one until they get a reply.

    See the service →

Your district’s books can be checked like this too.Start with a 30-minute call

03

How we work together

Start with a 30-minute conversation about your district, the issues you want to pursue and your office's daily workload. Afterwards we propose a scope and a delivery rhythm. The three services work on their own or together.

Call or email, whichever suits you: phone 0900-489-893, email info@4force.com.tw. Go to the contact section of the Civic Questions Service page →