MGAP7001 Public Governance and Policy Implementation: Comparative Perspective
The Complete Study & Assessment Guide · MGAP7001 Overview
- The University of Hong Kong
- First Semester, 2026-27
- Postgraduate
- Ten teaching topics
- Four graded components
What this course is actually about
MGAP7001 examines how public policies are designed and implemented, and how far they reach the people they were meant to reach.
- Assessed by written policy work Participation, two policy memos, a team report with a presentation, and an individual research paper.
- Hardest step Naming an analytical frame and then applying it, rather than summarising what the readings said.
- Watch the clock Two deadlines are set at 8am and one at noon, and the late penalty compounds steeply after the first day.
- Cheapest mark available The paragraph explaining your own contribution wherever AI assistance was used.
How MGAP7001 is assessed
| Component | Weight | Format |
|---|---|---|
| Class participation | 10% | Attendance at lectures and tutorials, with regular contribution to discussion of the assigned readings |
| Policy memos | 20% | Two memos on two different topics, each written as three slides and each due at 8am on the morning of that topic's lecture |
| Team Policy Project: written report and presentation | 40% | Teams of about six. A report of at least 1,500 words plus a 20-minute presentation, drawing on at least two course readings and at least eight additional sources |
| Individual Policy Research Paper | 30% | About 1,000 words stating a governance challenge and the analytical frame used to examine it |
The published weights total 100%. The course materials do not state a hurdle or minimum-mark requirement for any component, so none is claimed here in either direction; confirm the current position on the course Moodle page. Three published rules carry their own arithmetic: a late submission loses 5% of the mark on the first calendar day and 10%, 20% and 40% on the second, third and fourth, and work submitted five calendar days after the deadline is not accepted; written work receiving AI assistance requires a paragraph explaining the student's own contribution at each key step, on penalty of a 25% reduction in coursework marks, and the same applies to the team project.
Assessment structure
Segment widths reproduce the published percentage weights and total 100%.
What MGAP7001 covers
Ten source-led chapters move from why governance matters, through the machinery that turns a decision into a delivered service, to the conditions under which a reform survives long enough to matter.
Why Governance Matters in a Changing World
Read state capacity and societal capacity as two measures, not one dialTwo capacities, measured separately
Governance is a relationship, not an amount of government, and the opening topic exists to make that difference operational.
- State capacity and societal capacity as separate measures
- Weak, despotic and inclusive arrangements
- Conditional effects: why a shock has no fixed sign
- Policy feedback and its two directions
- Choosing the unit of comparison honestly
Implementation in Centralized and Federal Systems
Couple the three streams at an open window, and separate the level of devolution from the fact of itFrom a decision to a delivered service
Public governance courses often start by explaining how a subject ever reaches a decision at all, and this chapter marks the turn toward a different question: what happens once a decision exists.
- Why three separate streams must meet before a decision is made
- Policy entrepreneurs and the narrow opening a window actually offers
- Decentralization as a choice of level, not only of degree
- Why a center bypasses its own regions when it devolves power
- Reading state capacity within one country, and what two cases can prove
Accountability Systems: Design, Trade-offs and Metrics
Tell measurement, formal authority and disclosure apart from answerability with a consequenceWhat an accountability system actually has to contain
Public administration scholarship treats accountability as a design problem with four working parts: a stated goal, a named actor who must answer for progress toward it, an indicator that stands in for the goal because the goal itself usually cannot be observed directly, and a consequence that follows once the indicator is reported.
- Goals, actors, indicators, and consequences as the four parts a working system needs
- Administrative and professional forums operating inside an organization's own hierarchy
- Political and market forums operating outside the organization through votes or exit
- Self-enforcing, enforceable, and unenforceable directives as a test of real authority
- Disclosure documents versus disclosure announcements in shaping trust and dispute channels
Policy Instruments for Resources and Outcomes
Name the target behaviour and the obstacle before choosing the toolAn instrument choice is a diagnosis before it is a decision
Introductory treatments of policy tools tend to read like a catalogue: a designer can reach for a rule, a payment, a persuasion campaign, a training program, or a redesigned agency, and the choice looks like a matter of preference or political convenience.
- Naming the target behavior before naming any instrument
- Matching authority, money, information, training and reorganization to the obstacle that is actually present
- Procurement as an instrument carrying industrial, social and environmental goals
- Executive and unilateral instruments: speed weighed against reversibility
- Government spending's electoral return as a separate currency from its policy return
Public Engagement and Stakeholder Voice
Ask what decision right moved, who could reach it, and whose absence goes unrecordedAn invitation is not influence
Each reading in this topic is a clinical note on a different way a well-intentioned consultation fails, and together they supply a diagnostic sequence worth running over any participation mechanism: what power does it hand over, who can physically reach it, and whose absence will go unnoticed.
- Rungs of engagement, from informing to delegating
- Participation and trust: what the experiment separated
- The asymmetry between opening and closing a channel
- Structural inequality in who reaches a forum
- Procedural levers that outperform persuasion
Collaboration Across Sectors, Local and Global
Check interest alignment, then measurability, before any discretion is sharedTwo gates, checked in order
Collaboration is the vocabulary governments reach for when a problem crosses their own boundaries, and the fieldwork here is unusually honest about how poorly that vocabulary predicts behaviour.
- Interest alignment as the first gate
- Measurability as the condition for granting discretion
- Transaction costs and why coordination is withheld
- Conception, construction and performance errors
- Institutional distance as an advantage as well as a liability
State Capability for Complex Problems
Measure capability per service and locate the break at one of three levelsCapability has a grain, and the usual grain is wrong
Index a country, a ministry or a council on structure, budget and staffing and you produce a single number that assumes resources spread evenly across everything the organisation does.
- Why an aggregate capacity score misleads
- The paired-service design and what it isolates
- Three levels at which a complex service breaks
- Grading a problem definition rather than admiring it
- Political sustainability as a design-stage criterion
Policy Innovation and Diffusion
Separate learning, competition, imitation and coercion behind one adoption curveFour mechanisms, one observable
Diffusion is the topic where a striking pattern most easily substitutes for an explanation.
- Learning, competition, imitation and coercion
- Why the same curve fits four different stories
- Policies, decision systems and ideas as different cargo
- Carriers: who physically moves a practice
- Reinvention, and the units that never adopt
Governance in the AI Era
Digital intermediaries, inclusion of actors against inclusion of outcomes, and a broken mediatorDigital delivery does not create direct contact
The assumption that an online service reaches the citizen directly is usually implicit rather than argued, and it is what justifies counting a service as accessible once it exists.
- Digital intermediaries and the coverage-to-use gap
- Close and hierarchical intermediaries compared
- Inclusion of actors against inclusion of outcomes
- Moralising and regulating as a sequence
- Testing a mediator, and what a broken link proves
Public Policy as an Engine for Human Progress
Backlash as negative feedback, absent preferences, and phase-ins that buy timeBacklash, defined precisely
The course closes on the question it has been circling since the first topic.
- Backlash as an extreme case of negative feedback
- Four cells: actor level against threat type
- Sequencing redistribution rather than maximising it
- Absent preferences and why silence is informative
- Designing a formula that survives its own losers
It works comparatively, setting centralized against decentralized and federal systems, and it draws its analytical frames from social science and policy studies: problem-oriented governance, collective impact, executive-centred action, policy diffusion and the policy window.
The organising question across all ten topics is the distance between a decision taken at one level of government and a service delivered at another.
How the ten topics fit together
The sequence has a spine. The opening topic establishes that governance is a relationship rather than a quantity, and that state capacity and societal capacity have to be measured separately.
Topics two to four are the machinery: who decides, who is answerable, and which instrument gets used. Topics five to seven are the constraint: whose voice enters the process, who else has to be at the table, and whether the state can actually do the thing.
Topics eight to ten are the dynamics: how a practice spreads, what a new technology does to the whole arrangement, and why some reforms are reversed before they mature.
What the assessment rewards
Every graded task asks you to do something with a literature rather than report it.
The policy memos ask for the significance of two different readings on one topic, the agreement or disagreement between them and why, two critical questions and the policy relevance. The team policy project asks what works, what does not work, under what conditions and with what consequences, addressed to an office that could act on it.
The individual research paper asks you to state a challenge and then name the analytical frame you are bringing to it. A student who summarises five articles accurately and takes no position has answered a different question from the one set.
How to use this guide
Work one chapter at a time.
Each names the frames its topic supplies, works one case in full so you can see how long a qualification should be and where it goes, and then tests the frame against a case it does not fit. That second move is what separates a strong answer from a competent one, because a frame that explains everything explains nothing.
Chapters close with practice prompts shaped like the assessed tasks and a short list of terms you should be able to define without notes.
Confirming the operational details
Assessment labels, weights, deadlines and penalties follow the published course materials for this offering. Teaching explanations and practice cases are independently authored.
Some things students often want are not published in those materials, including a referencing style, a grade descriptor table and any minimum-performance condition attached to a component. This guide does not fill those gaps with a guess in either direction; confirm them on the course Moodle page or with the teaching team early enough that the answer can still change what you write.
Worked example · free
Turning a governance question into an answerable one
- 3Refuse the question as posed and name what would count as improved delivery for this service.
- 3Separate the fact of devolution from the level chosen, and say which one the evidence speaks to.
- 3State the observable that would distinguish a capability gain from a reallocation of visibility.
Key terms
- Policy Implementation
- Policy Implementation is the chain of action between an authoritative decision and a change in the behaviour of the people the policy targets, and it is where most stated goals are lost.
- State Capacity
- State Capacity is a government's ability to reach into its territory and actually carry out a decision, resting on fiscal, legal and administrative machinery rather than on formal authority.
- Policy Window
- A Policy Window is a short opening in which a recognised problem, an available solution and favourable politics can be coupled, and it closes whether or not anyone acted.
- Accountability Forum
- An Accountability Forum is the setting that determines who owes an account to whom, what information counts as an account, and what follows from the judgement.
- Policy Instrument
- A Policy Instrument is the tool a government selects to change behaviour, such as authority, money, information or organisational change, and each fails for different reasons.
- Target Compliance
- Target Compliance is whether the people a policy acts on behave as its designers intended, and it is the final and least theorised link in the implementation chain.
- Collaborative Governance
- Collaborative Governance is a delivery arrangement in which a public body shares discretion with partners outside government, gaining judgement and losing direct control in the same transfer.
- Policy Diffusion
- Policy Diffusion is the spread of a practice across jurisdictions, describable as a pattern but explicable only once the mechanism driving adoption is identified.
- Policy Feedback
- Policy Feedback is the way an existing programme reshapes the politics that follow it, building constituencies that defend it or mobilising opposition against it.
- Political Sustainability
- Political Sustainability is the likelihood that a policy survives long enough to produce its intended effects, and it can trade against both efficiency and equity.
MGAP7001 FAQ
How is this course assessed, and what is each component worth?
Four graded components make up the whole mark: class participation at 10%, two policy memos together worth 20%, a team policy project comprising a written report and a presentation at 40%, and an individual policy research paper at 30%. The published weights total 100%.
There is no examination component in the published structure, so revision should be organised around producing written policy analysis to deadline rather than around recall under timed conditions.
What exactly does a policy memo have to contain?
Each memo is written as three slides and covers two different readings on one assigned topic. The required elements are the significance of the key arguments and findings in both, the agreement or disagreement between them together with the reason for it, two critical questions, and the policy relevance connected to your own field experience, a report or a news article.
The three-slide limit is a real constraint, since it forces a choice about which disagreement matters.
How strict is the late penalty?
It compounds sharply and then stops entirely. One calendar day late costs 5% of the mark for that piece; two, three and four days cost 10%, 20% and 40% respectively; anything submitted five calendar days after the deadline is not accepted and receives zero.
Note also that several deadlines are set at a time of day rather than simply a date, so a submission on the correct day but later in the afternoon is still late under the schedule.
What has to be written if AI assistance was used?
A paragraph explaining how the student further contributed to each key step of the research and production process. The requirement applies to any written work that received such assistance, and the same obligation falls on a team for the team policy project, where the role of the assistance and each member's own contribution must both be explained.
Failure to provide it results in a 25% reduction in coursework marks, which makes it among the most expensive omissions available.
What is an analytical frame, and why does the research paper insist on one?
A frame is a named account of how something works that tells you which evidence matters and which comparisons are admissible: the three streams and the policy window, problem-oriented capability, an accountability forum, a diffusion mechanism. The paper asks for one because naming it converts a description into an argument that can be disputed.
Without a frame, a paper accumulates facts about an issue and gives the reader no way to judge whether the conclusion follows.
How should a team divide a project of this shape?
The report and the presentation answer the same four questions, so divide by analytical section rather than by member preference. One pairing takes what works and under what conditions, another takes what does not work and with what consequences, and someone owns the recommendation and the evidence behind it.
Since the report and the slides are both due at 8am on the first presentation day whatever day a team presents, the internal deadline has to sit well before that.
Questions students ask in each MGAP7001 chapter
- How do I compare two countries without falling into a structural explanation?
- Match the pair on the things you are not testing, then vary the one condition you are. If two polities share their institutional inheritance, their geography or their exposure to war and still diverge, structure cannot be carrying the explanation. State explicitly what the match holds constant, because a comparison that holds nothing constant is an illustration rather than evidence, and a reader will treat it that way. Chapter 01 · Why Governance Matters in a Changing World
- Why would a government create districts instead of the regions that were actually asking for more autonomy?
- Because districts are smaller and more numerous, no single one can accumulate enough population, budget, or political standing to challenge the center the way a full region could. Handing authority to that fragmented tier lets a government answer the general demand for local control while denying any unit the scale needed to organize a serious opposition, which is why local and regional transfers function as substitutes for each other rather than as two steps on the same ladder toward federalism. Chapter 02 · Implementation in Centralized and Federal Systems
- Why doesn't a published performance report count as accountability on its own?
- A report satisfies only the indicator element of the four-part design and says nothing about whether an account-holder can attach a consequence to what it shows. Accountability is better understood as a relationship with stakes: someone must judge the account-giver's information and be able to reward or sanction based on it. Without that step, a report can run for years without ever changing an incentive, which is why the sharper question about any dashboard is not how detailed it is but who acts on it and how. Chapter 03 · Accountability Systems: Design, Trade-offs and Metrics
- Why does naming the target behavior have to come before naming an instrument?
- Because an instrument is only as good as the obstacle it is built to remove, and that obstacle cannot be identified until the required behavior has been stated precisely enough to be observed. A designer who starts from a favored tool, such as a public information campaign or a subsidy, risks matching that tool to a behavior it was never suited to change, then blaming the target population when compliance does not follow. Stating the behavior first also forces a check on whether a program is self-executing or genuinely depends on someone outside government choosing to act. Chapter 04 · Policy Instruments for Resources and Outcomes
- Does consulting people improve trust in government?
- It raises trust somewhat, and delivering the project raises it more, but the two do not compensate for one another. The experimental evidence found no meaningful interaction, which means a consultation cannot repair the trust lost when a programme underperforms. Treat participation as worth doing on procedural grounds and as a sound investment where delivery looks likely, not as insurance against failure. Chapter 05 · Public Engagement and Stakeholder Voice
- When should a government contract rather than collaborate?
- When interests align on the output but not on the judgement calls, and when the result is specifiable in advance. Contracting keeps discretion inside government and pays an outside body to execute a defined task, which is the right trade whenever you could not observe how discretion was used. Collaboration is worth its risk only when the task genuinely needs judgement you could not have written down beforehand. Chapter 06 · Collaboration Across Sectors, Local and Global
- Why not just use a national governance indicator?
- Because it answers a different question from the one a service failure poses. An aggregate indicator averages across everything an administration does and therefore predicts the visible services well and the invisible ones badly, which is precisely the wrong way round for diagnosing a breakdown. Measure the named service, then locate the break at leadership comprehension, at office structure and procedure, or at how frontline staff are able to work. Chapter 07 · State Capability for Complex Problems
- How do I tell competition from learning in a real case?
- Look at the lag and at what adoption tracks. Learning needs enough time for an outcome to become observable and predicts that adoption follows published results; competition predicts adoption clustering geographically and arriving faster than any evaluation could. Economically consequential policies among neighbouring jurisdictions are the classic competitive signature, and a briefing that reports only the count cannot distinguish the two at all. Chapter 08 · Policy Innovation and Diffusion
- What should a digital inclusion strategy measure instead of availability?
- Independent completion rates, broken down by the groups with the lowest digital skills. Availability measures the supply side and will read as success in exactly the population that cannot finish a transaction unaided, which converts the unpaid labour of family intermediaries into a reported state achievement. If a strategy exists to serve low-skill groups, its headline measure has to be something only those groups can move. Chapter 09 · Governance in the AI Era
- Why not legislate the most progressive version immediately?
- Because it creates the opposing coalition before it creates the defending one. A reform whose costs are felt at once and whose benefits arrive later hands opponents a mobilised constituency during exactly the window when the policy has nothing to show, which is the standard route to repeal before maturity. Building the broad constituency first and raising progressivity afterwards reaches the same destination by a path that survives an election. Chapter 10 · Public Policy as an Engine for Human Progress
How to prepare for the assessments
Read for frames rather than for findings. For each topic, be able to state the frame in one sentence, the distinction it protects, and the case where it does not apply. Then rehearse compression, because three slides and a thousand words are the binding constraints here: take any argument and cut it to a hundred words without losing the mechanism or the counter-case.
Close every piece of written work by naming what would have to be true for your recommendation to hold.
Your AI Management tutor for MGAP7001
Stuck on a hard MGAP7001 question? Sia is AskSia’s AI Management tutor — ask any MGAP7001 Public Governance and Policy Implementation: Comparative Perspective question and get a clear, step-by-step explanation grounded in how the course is actually taught and assessed. Read this whole study guide free, then take your hardest questions to Sia.
