Rainer Kattel 专栏丨英国撤销科学、创新与技术部后的再观察

雷纳·卡特尔（Rainer Kattel）

伦敦大学学院（UCL）创新与公共目标研究所（IIPP）副所长，教授

今年7月，英国新任首相安迪·伯纳姆（Andy Burnham）上任第一天，便撤销了科学、创新与技术部（Department for Science, Innovation and Technology，DSIT），其团队、预算和下属机构则分别转入其他部门。

该部门成立还不到三年半。2023年，里希·苏纳克（Rishi Sunak）设立这一部门，意在表明科学技术对英国未来至关重要。

如今，该部门的职能被分为三部分：科学与创新职能划入扩大的商业、创新、科学与贸易部；数字政府职能，包括政府数字服务局（Government Digital Service，GDS），划入数字、文化、媒体与体育部；人工智能战略及公共部门人工智能应用，则交由政府中枢负责，划入内阁办公室和新成立的首相与内阁办公室。

这次事件的幕后决策者并不神秘。但是真正值得追问的是，这场调整究竟让英国失去了什么。

英国科学界和创新界对此作何反应？简而言之，各方态度不一：有人失望，有人不以为意，也有人私下表示赞同。更深入地看，这场调整不仅揭示出英国政府调整自身组织架构的方式，也反映出新政府真正看重的政策方向。

各方反应不一

最鲜明的批评来自英国皇家学会主席保罗·纳斯爵士（Sir Paul Nurse）。今年9月，他在英国下议院科学与技术委员会作证时表示，撤销DSIT“实际上是一个相当大的错误”。英国政府研究所（Institute for Government）此前也曾公开批评这一决定。而初创企业和科技行业团体则称，这是在错误的时间做出的错误调整，并担心科学研究和年轻企业在被纳入规模更大的商业部门后会受轻视。英国科学与工程运动组织（Campaign for Science and Engineering）认为，加强科学与商业之间的联系可能带来机遇，但同时警告，基础研究的优先地位可能因此下降。

另一些人对此更为淡定。前科学大臣格雷格·克拉克（Greg Clark）和乔治·弗里曼（George Freeman）均对这一调整表示欢迎。克拉克认为，政府部门之间存在过多职能壁垒。我在伦敦大学学院（UCL）的同事詹姆斯·威尔斯登（James Wilsdon）则提出了一个更为审慎的观点：相比政府部门的组织架构，科研经费、人才和需求对研究的影响更大。1995年至2023年间的大部分时间里，科学事务都归属商业部门；而为DSIT确定的2026—2027至2029—2030年度585亿英镑研发预算，也不会因这次调整而改变。

双方的观点都有道理，但在我看来，他们都忽略了这场变动最值得关注的问题。

在英国，调整政府架构很容易——或许太容易了

在英国，调整政府架构相当容易。首相可以根据自身意愿，设立、合并、重新命名甚至撤销政府部门和机构，也可以修改其职权范围，而且往往一夜之间就能完成。这类调整通常不需要经过宪法法院审查，也很少需要制定新法律。因此与大多数欧洲大陆国家相比，以及在某些方面截然不同的中国相比，英国的政府架构可塑性极强。

这种灵活性常被视为优势，有时也确实如此。它使政府能够迅速应对新问题，但也有代价。很多时候，机构重组本身就成了政策信号：设立一个部门，意味着某项议题受到重视；撤销或合并一个部门，则意味着政策方向发生变化。然而，政府往往没有说清楚更具体的问题：这次调整究竟要建设哪些新能力？将改变哪些工作惯例？又将如何判断调整是否奏效？

重组的代价是真实存在的，只是大多不容易直接看见。英国政府研究所估计，新设一个政府部门或进行一次中等规模的部门合并，直接成本约为1500万英镑。实际成本还要高得多，包括数月的工作效率损失、重新招聘人员、部署新的信息技术系统、员工面临的不确定性，以及大学、企业和科研资助机构等外部合作伙伴突然不知道该联系谁。一位参与议会调查的证人曾形象地说，我们应该暂时停止反复挪动甲板上的躺椅了。

这里有一个我此前讨论过的悖论。公共部门创新需要一种我和同事称为“敏捷的稳定性”（agile stability）的能力：既能开展试验、及时调整方向，又能依托相对稳定的组织，在多年时间里积累专业知识、工作惯例和信任，这一过程往往需要十年甚至更久。如果机构调整过于容易，重组就可能取代那些缓慢而不显眼、却真正能够提升组织能力的工作。制度创新由此可能沦为制度的反复折腾。从这个意义上说，变动过多和变动过少一样，都可能束缚政府的行动能力。

以政府数字服务局为例

政府数字服务局（GDS）的经历最能说明这一问题。GDS于2011年在内阁办公室内部成立。在随后的十年里，它成为推动英国政府数字化创新的关键力量，其工作模式也被其他国家借鉴。GDS采用的工作方式包括组建小型跨学科团队、开展用户研究、自主开发而非外包，以及采用开放标准。这些做法改变了一代英国公务人员对公共服务交付方式的认识。

2024年，GDS从内阁办公室划入DSIT，目的是建立更强大的政府数字化中枢。议会内部设立的一个委员会认为这一决定并不妥当，但在当时的议会任期内再度调整架构并不现实。两年后，政府还是再次调整了GDS的归属，只不过这次并未将其放回政府中枢，而是划入文化部门。短短两年左右，GDS已经第三次更换所属部门。与此同时，政府的人工智能工作组仍留在内阁办公室。多位专家曾向议会委员会发出警告：将技术交付团队与政策制定部门分开存在风险，反复调整组织归属也会打击员工士气。

当然，GDS实际承担什么工作，比它在白厅组织架构图上的位置更重要。但政府的组织能力存在于团队、工作惯例和协作关系之中，而这些恰恰是反复重组最容易破坏的东西。

从产业政策转向公共服务创新

不过，撤销DSIT并不只是又一次折腾，它也反映出政府政策重点的真正转变。坦率地讲，在伯纳姆领导下，英国政府正将工作重心从产业政策转向公共服务创新。

9月，伯纳姆以党魁身份在第一次党代会上的讲话明确表达了这一点。演讲的核心是一项国家护理服务计划（National Care Service），旨在提供使用者无需付费的公共护理服务，并在不增加借款的情况下筹资。伯纳姆认为，如果社会护理体系持续失灵，最终也会拖垮国民医疗服务体系（NHS）。此外，他还提出实施战后以来规模最大的市政住房建设计划、取消禁止水务公司公有化的规定、成立新的公共电网公司以挑战私人运营商，并改革技术教育和青年心理健康服务。

产业战略主要体现为推动地区再工业化的承诺，具体方案则有待后续公布。除了提及英国要成为“创新国家”，演讲几乎没有涉及科学和数字政策；人工智能主要作为英国担任G20轮值主席国期间的一项国际议题出现。

这是一次重要的政策转向。过去十年，英国乃至欧洲许多国家的创新政策讨论，主要围绕产业战略、增长目标，以及人工智能、量子技术等前沿领域的国际竞争展开。新政府则押注于另一种判断：英国面临的关键创新挑战，可能不在前沿技术领域，而在于如何组织和提供护理、住房、医疗和福利等公共服务。

英国并非个例。在世界各地的民主国家，政治议程的重心正逐渐从前沿技术转向日常生活的可负担性。在纽约，市长佐兰·马姆达尼（Zohran Mamdani）凭借一套以降低生活成本为核心的政策主张赢得选举，包括免费公交、冻结房租、普及托育服务以及由市政府运营的杂货店。这些主张被许多经济学家认为难以实施。曾组织经济学家支持其竞选的经济学家伊莎贝拉·韦伯（Isabella Weber），则在新书《反法西斯经济学》（Anti-Fascist Economics）中提出了更广泛的观点：当政府面对生活必需品价格持续上涨却无所作为时，就会加剧社会不安全感，而这种不安全感可能成为极右翼势力滋长的土壤。因此，让住房、能源、食品和护理服务保持可负担性，也是民主制度的重要基础。

伯纳姆将政策重点转向护理、市政住房、水务和电网，与这一思路属于同一类政策取向。这些政策主张的共同点，也是批评者容易忽略的，其重点并非花多少钱，而是政府能否真正把公共服务做好：能否日复一日地保障公交运营、建设住房，并提供有序有效的照护服务。些

重组背后的悖论

在这里，前面讨论的两条线索交汇在一起，形成了一个悖论。如果公共服务创新已成为新政府的优先事项，那么GDS所代表的能力就比以往更加重要。这些能力包括围绕用户需求重新设计服务、自主建设并掌握数字基础设施、快速迭代，以及将政策制定与一线服务实施衔接起来。

国家护理服务不可能仅靠立法和资金投入就建立起来。英格兰的社会护理服务由数千家服务提供者具体实施，由地方政府负责监管，同时又与NHS紧密交织。要将这些环节整合成一个协调统一的服务体系，需要共享数据和数字基础设施，更需要建立新的组织运行机制，把政策制定者、护理人员、家庭以及接受护理的人联系起来。

换句话说，政府最重要的政策目标，恰恰依赖于此次机构重组可能削弱的能力。再次调整GDS的归属，并将其与位于政府中枢的人工智能团队分开，可能削弱政府自身的创新能力，而这些能力恰恰是改善公共服务所必需的。

对英国以外国家的启示

每个国家都需要根据新的政策重点，定期调整自身组织体系。关键从来不在于政府是否应该调整组织架构，而在于如何调整才能避免破坏组织内部已经积累的能力。

我认为，有三点经验值得借鉴。

第一，区分政策使命与组织能力。政府的政策重点可以、也应当随形势变化而调整，但承载专业能力的团队和工作机制需要得到保护，并保持连续性。

第二，机构重组应明确说明预期建设哪些能力，以及通过什么标准衡量进展。否则，重组就只是在传递政策意图，难以转化为实际成效。

第三，当政府将工作重心转向公共服务时，数字化能力和服务交付能力就应被视为核心基础设施，而不是可有可无的技术辅助。

英国调整政府机构的便利，是一项难得的资源。运用得当，它有助于政府快速学习、适应变化；但如果使用过于频繁，政府就可能看起来一直忙于改革，却一再推迟真正需要时间的能力建设工作。未来几年将会检验——DSIT的撤销究竟属于哪一种情况。

The Moving Ministry

Britain’s vanishing science department, and the difference between institutional innovation and institutional churn

In Edmund Crispin’s 1946 detective novel The Moving Toyshop, the poet Richard Cadogan wanders into an Oxford toyshop late one night and finds a dead woman inside. When he returns in daylight, the toyshop has gone. In its place is a grocer’s, and nobody seems to remember anything else. British scientists may feel a little like Cadogan this autumn. A department that looked like a fixture of Whitehall has vanished, and its teams, budgets and agencies have reappeared behind other people’s shopfronts. Unlike in Crispin’s novel, there is no mystery about who did it. The more interesting question is what was lost in the move.

On his first day in office this July, Britain’s new Prime Minister, Andy Burnham, abolished the Department for Science, Innovation and Technology (DSIT). The department was barely three and a half years old. Rishi Sunak had created it in 2023 to signal that science and technology were central to Britain’s future. Its functions have now been split three ways. Science and innovation go to an enlarged Department for Business, Innovation, Science and Trade. Digital government, including the Government Digital Service (GDS), goes to the Department for Digital, Culture, Media and Sport. AI strategy and public sector AI adoption go to the centre of government, into the Cabinet Office and a new Office for the Prime Minister and Cabinet.

How have British scientists and the innovation community have responded? The short answer is: with a mixture of dismay, shrugs and some quiet approval. The longer answer says something interesting about how the British state changes itself, and about where the new government’s priorities really lie.

A divided response

The most prominent criticism came from Sir Paul Nurse, President of the Royal Society, who told the House of Commons Science and Technology Committee in September that abolishing DSIT was “actually quite a big mistake”. The Institute for Government, which also gave evidence, had already called the change a mistake in print. Start-up and tech industry groups called it the wrong change at the wrong time, and worried that science and young companies would be squeezed inside a much larger business department. The Campaign for Science and Engineering saw some opportunities in closer links to business, but warned about the de-prioritisation of basic research.

Others were more relaxed. Former science ministers Greg Clark and George Freeman both welcomed the move; Clark argued that government simply has too many silos. My UCL colleague James Wilsdon made perhaps the most sober point: the shape of Whitehall matters less for research than money, people and demand. Science sat inside a business department for most of the period from 1995 until 2023, and the £58.5 billion R&D budget set for DSIT for 2026–27 to 2029–30 has been locked in.

Both camps have a point. But I think both also miss what is most distinctive about this episode.

Change is easy in Britain — perhaps too easy

In the United Kingdom, changing the machinery of government is remarkably easy. A prime minister can create, merge, rename or abolish ministries and agencies, or rewrite their mandates, largely at will and often overnight. There is no constitutional court to ask, and rarely a need for new legislation. Compared with most continental European states — and, in a very different way, with China — the British state is extraordinarily plastic.

This plasticity is often presented as a strength, and sometimes it is. It allows governments to respond quickly to new problems. But it has a cost. Very often, the reorganisation itself becomes the message. Creating a department signals that an issue matters; abolishing or merging one signals a change of direction. What is frequently missing is the detail: what new capabilities the change is supposed to build, which routines it will alter, and how anyone will know whether it has worked.

The costs are real, even if they are mostly hidden. The Institute for Government estimates the direct cost of creating a new department or a mid-sized merger at around £15 million, and the true cost is certainly higher: months of lost productivity, re-recruitment, new IT systems, staff uncertainty, and partners outside government — universities, companies, research funders — who suddenly do not know whom to call. One witness to the parliamentary inquiry put it memorably: we should stop moving the deckchairs around for a little while.

Here lies a paradox I have written about before. Public sector innovation requires what my colleagues and I call agile stability: the ability to experiment and change direction, but built on stable organisations that accumulate expertise, routines and trust over years, often over a decade or more. When change is too easy, it becomes a substitute for the slower and less visible work of building capabilities. Institutional innovation turns into institutional churn. In that sense, too much change can be just as stifling as too little.

The case of GDS

Nowhere is this clearer than with the Government Digital Service. GDS was created in 2011 inside the Cabinet Office and became, over the following decade, the key driver of digital innovation in the British state and an internationally imitated model. Its approach — small multidisciplinary teams, user research, building rather than outsourcing, open standards — changed how a generation of public servants think about delivering services.

In 2024 GDS was moved from the Cabinet Office into DSIT, to create a stronger digital centre of government. A parliamentary committee concluded that this was the wrong call, but that reversing it at that stage of the parliament would be unrealistic. Two years later, it has been moved anyway, though not back to the centre: GDS now finds itself in the culture department, its third home in about two years, while the government’s AI taskforce sits in the Cabinet Office. Several experts warned the parliamentary committee about separating technical delivery teams from policy-making, and about the toll that repeated moves take on staff morale.

It is fair to say that what GDS does matters more than its address on the Whitehall map. But capabilities live in teams, routines and relationships, and those are precisely what repeated reorganisations disrupt.

From industrial policy to public service innovation

Yet the dissolution of DSIT is not only about churn. It also reflects a genuine shift in priorities. To put it bluntly: under Burnham, the focus of the British government is moving from industrial policy towards public service innovation.

His first party conference speech as leader, in September, made this unmistakable. Its centrepiece was a National Care Service, free at the point of use and funded without borrowing, on the argument that broken social care will break the National Health Service too. Alongside it came the biggest council house building programme since the post-war era, the reversal of the ban on public ownership of water companies, a new public grid company to challenge private operators, and reforms to technical education and youth mental health services. Industrial strategy appeared mainly as a promise of regional re-industrialisation, with a fuller plan to follow. Science and digital policy went unmentioned beyond a nod to Britain as an “innovation nation”; AI featured mainly as an international agenda for Britain’s G20 presidency.

This is a significant reorientation. For the last decade, British innovation policy debates — like those in much of Europe — were dominated by industrial strategy, missions for growth, and competition in frontier technologies such as AI and quantum. The new government is betting that the most important innovation challenges for the British state lie elsewhere: in how care, housing, health and welfare are actually organised and delivered.

Britain is not alone in this. Across the democratic world, the political centre of gravity is shifting from frontier technologies to the affordability of everyday life. In New York, Mayor Zohran Mamdani won office on an affordability agenda — free buses, a rent freeze, universal childcare, city-run grocery stores — that many economists dismissed as unworkable. The economist Isabella Weber, who organised economists in support of his campaign, makes the broader case in her new book, Anti-Fascist Economics: when governments stand by as the cost of essentials rises, they feed the insecurity on which the far right thrives, and affordable housing, energy, food and care are therefore a foundation of democracy itself. Burnham’s turn to care, council homes, water and the grid belongs to the same family of ideas. What these agendas share, and what their critics tend to miss, is that they are less about spending than about the state’s capacity to provide: to run buses, build homes and organise care well, day after day.

The paradox at the heart of the reorganisation

Here the two threads meet, and the result is a paradox. If public service innovation is now the government’s priority, then the capabilities that GDS embodies — the capacity to redesign services around users, to build and own digital infrastructure, to iterate quickly, to connect policy with frontline delivery — matter more than ever. A National Care Service will not be built by legislation and money alone. Social care in England is delivered by thousands of providers, overseen by local authorities, and intertwined with the NHS. Making it work as a single service will require shared data, digital infrastructure, and, above all, new organisational routines that connect policy-makers, carers, families and the people being cared for.

In other words, the government’s central ambition depends on exactly the type of capabilities that its first reorganisation has unsettled. Moving GDS once more, and separating it from the AI teams at the centre, risks weakening the state’s own capacity to innovate in the very services the government cares most about.

Lessons beyond Britain

What might readers in China take from this? China, too, reorganises its innovation institutions — the 2023 reforms that created the Central Science and Technology Commission and the National Data Administration are a recent example. Every state must periodically realign its organisations with new priorities. The question is never whether to change the machinery of government, but how to change it without destroying the capabilities it contains.

Three lessons seem to me to travel well. First, distinguish between mandates and capabilities: political mandates can and should change with priorities, but the teams and routines that carry capabilities need protection and continuity. Second, a reorganisation should come with an explicit account of which capabilities it is meant to build and how progress will be judged; otherwise it remains a signal of intent and little more. Third, when a government shifts its focus towards public services, its digital and delivery capabilities become core infrastructure, not a technical afterthought.

Britain’s ease of institutional change is a remarkable resource. Used well, it allows a state to learn and adapt quickly. Used too often, it becomes a way of looking busy while the slow work of building capabilities is postponed once again. The next few years will show which of these the end of DSIT turns out to be.

本文翻译：上海市科学学研究所邓彤。文章观点不代表主办机构立场。

◆ ◆ ◆

编辑邮箱：sciencepie@126.com