A ministry does not, on the whole, place more of its money in its own minister’s seat

We followed every federal minister and deputy minister who held a Dewan Rakyat seat between January 2021 and September 2026, through the Muhyiddin, Ismail Sabri and Anwar cabinets: 221 terms. For each term we set the ministry’s money placed in the holder’s own seat against what that seat’s share of every other ministry’s money would give it, and ranked the seat among all 222 on that ratio. 169 terms had enough awards to test. On average the holder’s seat ranked at the 50th percentile, where chance puts it; 20 landed in the top tenth, where chance puts 17. Had each holder’s seat received one extra fair share of its ministry’s money, the average would have been the 87th percentile, so a pattern that size is ruled out. Single seats still stand out, and Table 1 lists them.

A ranking, not a finding about any person: a seat can rank high because work there was due, and tender boards, not ministers, award contracts. Money placed in a seat is an estimate from the award’s title (Method).

How the test works

A term is one MP in one ministry’s minister or deputy minister post without a break: back-to-back posts in the same ministry, and the ministry’s renames, make one term, and senators (33 posts) have no seat to test. The 221 terms are 100 ministers’ and 121 deputy ministers’, held by 124 MPs in 113 seats; 75 of those seats have two or more terms, and 109 of the 222 seats had no minister or deputy in the window. So 221 counts terms, not seats, and is close to 222 by coincidence: every tested term’s seat is ranked among all 222, held or not.

Some seats draw more of every ministry’s money than others: cities, state capitals, seats with a hospital or a camp. So a minister’s seat is not compared with an average seat but with itself. Take Ahmad Zahid Hamidi, Minister of Rural and Regional Development since December 2022 and member for Bagan Datuk. In his term to September 2026 the ministry placed RM 22.7 million in Bagan Datuk. Bagan Datuk’s share of all the other ministries’ money in the same months would have given it RM 5.7 million. That ratio ranks Bagan Datuk at the 87th percentile of the 222 seats for that ministry in that term. Every tested term gets the same treatment, and under no favouritism their percentiles spread evenly from 0 to 100 (Figure 1).

Figure 1. The 169 tested terms by the percentile their holder’s seat ranked at for the holder’s own ministry, in tenths. The dashed line is what chance gives each tenth.

The seats that rank highest

Twenty terms put the holder’s seat in the top tenth. Some rest on one contract: Pontian’s RM 38.0 million from the Works ministry, against RM 3.1 million from its pull, while Ahmad Maslan was Deputy Minister of Works, is one award, the widening of the Pontian Besar bridge on Federal Route 5 (97th percentile). Others are small sums that a single cleaning or catering contract can move. The list also runs the other way: the Works ministry placed none of its seat-attributable money in Kapit during Alexander Nanta Linggi’s term as Minister of Works, against RM 26.9 million from Kapit’s pull (3rd percentile), and Education placed RM 10.9 million in Fadhlina Sidek’s Nibong Tebal against RM 22.3 million (36th).

Table 1. The twenty tested terms whose holder’s seat ranked highest for the holder’s own ministry, January 2021 to September 2026: the term’s length in months inside the window (a term begun before January 2021 counts from then, one still running to September 2026), the ministry’s money placed in the seat during the term, what the seat’s pull gives (its share of all other ministries’ money in the same months), the seat’s percentile among 222 seats, and whether the seat’s ratio was higher or lower than in the rest of the window. Tap a heading to sort.

Inside the term and outside it

A second reading compares each seat with itself over time: its ratio during the term against its ratio in the rest of the window, when someone else held the post. Of the 165 terms where both can be measured, 81 show a higher ratio during the term: a coin toss. Ministers (50th percentile on average) and deputy ministers (50th) look alike, and shifting every term six months later, to allow for awards that take time to be made, gives the 49th.

Table 2. The tested terms whose seat ranked lowest; “Show all” lists all 221 terms, including those not tested because the ministry placed fewer than 30 awards in seats during the term or the term was shorter than six months. Tap a heading to sort.

Limits

Only money that can be pinned to a seat counts: 60% of the placed money. An award naming only a district split between several seats is left out rather than shared.
A ministry that places few awards in seats cannot show a pattern; 52 terms are untested.
Senators who held posts (33 posts) have no seat and are left out, as are Prime Ministers’ and Deputy Prime Ministers’ own titles; their ministries are tested where they held one.
Ministers in the Prime Minister’s Department share one purchasing name, so each is tested on the whole department.
The test finds broad patterns. It cannot rule out a single favour, nor show why any seat ranks where it does.

Method

Office holders: the English Wikipedia articles on the Muhyiddin, Ismail Sabri and Anwar Ibrahim cabinets (read 16 October 2026): each post, holder, seat or senatorship, and the dates they took and left office, including acting posts. Back-to-back posts of one person in one ministry, and a ministry’s renames, make one term.
Window: January 2021 to September 2026, by the date the award was accepted. A term is tested if it ran six months or more inside the window and its ministry placed 30 or more awards in seats during it.
Seats: DOSM’s outlines of the 222 parliamentary seats. An award goes to a seat when its title names a facility or locality whose every point in the award’s district lies in one seat; to a district’s main seat when that seat holds 80% or more of the district’s voters (GE15 roll); otherwise it is not counted for any seat.
Pull: a seat’s share of all other ministries’ seat-placed money in the same months. Expected money = the ministry’s seat-placed money in the term × the seat’s pull. Ratio = (money + d) ÷ (expected + d), with d 1% of the average expected, so seats with nothing still rank. Percentile: the holder’s seat’s rank among all 222 seats on that ratio.
Under no effect the percentiles are uniform, so their mean is 50 with a standard error of 2.2 for 169 terms. Power: the same test with an extra fair share (expected money) added to each holder’s seat gives a mean of 87.