Dr Jing Li

2 April 2025

Since the beginning of American President Donald Trump’s second term in the White House, a new round of trade war has unfolded on a full scale. As of March 2025, not only has a 25% tariff been imposed on imported steel and aluminium, but targeted tariffs have also been implemented on major trading partners such as China, Canada, and Mexico, further escalating global trade tensions. In the face of America’s increasingly hard-line trade protectionist policy, the EU, Canada, China, etc. are attempting to alleviate the tariff pressure through litigation, negotiations, and countermeasures. However, the US has gone beyond these actions. On 26 March 2025, Trump announced that, effective on 2 April, all imported automobiles will be subject to a 25% tariff and plans are under way to impose the same tariff on non-American components of vehicles assembled on US soil, starting one month later. In addition, the “reciprocal tariffs” implemented simultaneously will step up the trade war.

In the wake of Trump’s first term as US President, tariffs were already at the core of his economic strategy. He firmly believes in them as a tool to boost domestic manufacturing and employment, framing tariffs as an important means of “Making America Great Again”. However, in actual fact such a policy has triggered grave concerns and adverse impacts around the world. Classical economic theories have long demonstrated that tariff hikes will hinder economic growth through various channels. First, tariffs will raise import costs, directly undermining consumer interests and weakening their purchasing power, in other words, compromising the overall consumption power. Second, rising import costs will also drive up production costs for domestic enterprises, implicating related industries through supply chains, and thereby leading to diminishing output. Subjecting American companies to liquidity pressure, this can even set off a chain reaction that affects the entire investment environment. Nonetheless, Trump remains convinced that the economic benefits derived from tariff-hike pressure will outweigh potential losses. Meanwhile, with the continued escalation of the trade war, companies and consumers in the US and around the world will face even greater harm.

Increased tariffs: a source of enemies on multiple fronts

The reciprocal tariffs are introduced by the US to escalate the trade confrontation from a bilateral to a multilateral one. In response to Trump’s non-discriminatory tariff policy, national governments and multinational companies are reassessing the trade landscape  and their respective investment strategies for the medium to long term. The new “reciprocal tariff” measures will not only take into account the tariff levels imposed by other countries on the US, but will also comprehensively consider the amount of subsidies provided by each country for its industries as well as any potential unfair trade practices from the US perspective. This goes to show that the so-called “reciprocal tariffs” are actually trade rules redefined based on American interests, particularly business interests. The imposition of these tariffs will no doubt be met with countermeasures from the rest of the world. Besides further intensifying uncertainties in the global economy, this will expose the US economy to even greater challenges.

History does not repeat itself but the similarities are astonishing. America’s well-known Smoot-Hawley Tariff Act in 1930 was a radical move at trade protectionism. To resolve domestic overcapacity, protect domestic industries, and aid farmers in hardship, the Act raised duties to a record high on more than 20,000 imported goods. Following the passage of the Act, retaliatory tariff measures by other nations dealt a serious blow to the US economy, resulting in a nearly two-thirds reduction in US imports and exports during the Great Depression. Despite the lack of hard evidence that retaliatory tariffs were the direct cause of the Great Depression, the pressure these tariffs exerted on the US market and the numerous disputes surrounding the Act were undoubtedly a catalyst for the severe economic recession in the US.

Today, Trump is still trying to create national wealth by raising tariffs to tackle a string of domestic economic problems, including employment. Nevertheless, judging by the tariff strategy during his first term, punitive tariffs this time are unlikely to achieve any significant effect. As it has turned out, this protectionist policy often backfires. The negative effects of the trade war, including inflation and investment uncertainty, will harm both American individuals and companies. The tariff war has been expanding since Trump’s second term began. If the reciprocal tariffs are implemented as scheduled, the new tariff regime will remain a threat issued by Trump to fight for national interests, and the expected countermeasures by other countries will also plunge the US economy into a predicament similar to that caused by the Smoot-Hawley Tariff Act.

China seeking international allies to break deadlock

2024 trade data shows that China, the EU, Mexico, Vietnam, and Taiwan were the top five export partners with the largest trade deficits with the US. Given that China has been levied an additional 20% tariff since the start of Trump’s second term, under the framework of reciprocal tariffs, China is unlikely to sustain any further impacts. Subject to the upcoming announcement by the US government on 2 April 2025, the details of tariff implementation warrant close attention.

Furthermore, unlike the trade war initiated in 2018, the current tariff hike is obviously not just targeted at China but also affects many other countries, including several American allies. Hence, to control trade costs, companies around the world will need to adopt different approaches to their investment strategies.

During the China-US trade war, numerous enterprises would choose to cooperate with Vietnam, Mexico, etc to bypass the US duty increases placed on China. In contrast, this business strategy may be much less effective for Trump’s upcoming tariff measures, leading to skyrocketing uncertainties in the investment environment and making “exiting” China not the best approach to evade rising tariffs. Enterprises now need to reassess their international investment strategies and previous decisions about investing in China to address the latest changes in the global trade landscape.

In response to the US protectionist policy, countries worldwide will proactively seek new  opportunities for international cooperation. This reflects efforts to overcome the US tariff barriers, ease pressures from the trade war, and safeguard their economic benefits and growth potential.

Against this backdrop, China should capitalize on current international cooperation opportunities and strive to develop mutually beneficial collaborations. In particular, in the domains of trade, investment, and technological innovation, China can further pursue such mechanisms as the Regional Comprehensive Economic Partnership (RCEP), which will be instrumental in cementing closer partnerships.

Through deeper international cooperation, not only can China strengthen its influence in the world economy, but it can also significantly contribute to an open, diverse, and mutually beneficial global economic order.

Professor Guojun He

26 March 2025

As of early 2025, the global cryptocurrency market was valued at over US$3 trillion (see Note 1). Cryptocurrency has made a leap from a specialized domain with a minority following in its early days to a global financial ecosystem. Unlike traditional financial markets, the cryptocurrency market operates 24 hours a day, seven days a week and transcends borders. At present, hundreds of publicly listed companies, hedge funds, family offices, and pension funds have invested in crypto assets. In July 2024, the BITCOIN Act was introduced in the US to establish a Strategic Bitcoin Reserve, underscoring the status of cryptocurrency as a strategically significant asset class.

Meanwhile, crypto mining is similarly an international industry worth billions of US dollars. To date, at least 13 crypto mining companies have been listed on the NASDAQ Index in the US, facilitating the creation of emerging industrial chains characterized by specialization and scale. From the production of specialized mining hardware and large-scale mining operations, to electricity trading and carbon credit management, crypto mining has evolved into a hybrid industry that integrates energy, technology, and finance, far beyond mere computation. With technological advancements and increasingly clear regulatory conditions, the cryptocurrency market is set to expand its influence in the world financial system and to play an even more important role in the international monetary system in future.

However, according to frequent media reports, the energy consumption of crypto mining is equivalent to that of a medium-sized country. This has not only made the industry the enemy of environmental protection but was also part of the reason the Beijing government imposed a carpet ban on it in 2021, citing “high energy consumption and carbon emissions; low contribution to the national economy; limited impact as a driving force for industrial development and technological advancement; and the detrimental effects of its unmonitored and chaotic growth on promoting high-quality economic and social development as well as energy saving and carbon reduction”. Meanwhile, latest research studies illustrate that under a properly-designed policy for electricity prices, crypto mining is likely to benefit the expansion of renewable energy and the reduction of overall carbon emissions. Prior to 2021, China had been the world’s largest crypto mining centre, accounting for over 65% of the global mining hashrate (see Note 2). In May of the same year, the National Development and Reform Commission and other government departments jointly announced plans to comprehensively crack down on and prohibit crypto mining (see Note 3). The considerations underlying the decision were the pressure of energy consumption on the objectives of “dual energy consumption control”, the need for financial risk management, and concerns over carbon emissions. Crypto mining activities were then transferred to the US, Kazakhstan, etc. In early 2025, the US became the biggest crypto mining market while legitimate crypto mining activities almost disappeared in China.

Utilizing excess renewable power

Crypto mining can help carbon emissions mitigation in that it encompasses three key features of renewable energy: high fixed cost of initial development; low marginal cost of electricity generation; and imperfect correlation between electricity supply and demand. Not only do these features hamper market potential but the grid systems would be affected if they are not flexible enough. In addition, given the “memoryless” nature of mining activities (i.e. no disruption to mining operations after shutting down and restarting within a very short time), they can serve as an ideal “adjustment device” for renewable energy power grids, compensating for the supply and demand imbalance of renewable energy. Such a mechanism can benefit the rapid development of renewable energy infrastructure in the Chinese Mainland.

While the National 14th Five-Year Plan affirms the goals of substantially accelerating the installation of clean energy infrastructure, including wind and solar power, the development of renewable energy faces the challenge of an “oversupply of wind, solar, and hydroelectric energy”. Despite China’s global dominance in installed capacity of renewable energy in 2024, the oversupply rate of solar energy in some regions had exceeded 10% while that of wind energy had reached up to 15% (see Note 4), causing severe wastage of clean energy and even resulting in negative power prices. Integrating crypto mining into the power grid adjustment system not only absorbs and utilizes surplus power but also overcomes the volatility of renewable energy, stabilizes grid operations, and generates economic benefits.

With a proper electricity pricing policy, introducing crypto mining can enhance investment in renewable energy production and curb overall carbon emissions under suitable conditions. In Western China, where energy resources are abundant, the adoption of crypto mining will help to augment the economic feasibility of renewable energy and encourage the expansion of renewable energy production.

For example, America’s Electric Reliability Council of Texas (ERCOT), a Texan electrical grid operator, has devised a large, flexible load curtailment programme, in which almost all operational large, flexible loads are crypto mining facilities. During the winter storm Elliott in December 2022, the crypto miners’ hashrate drastically declined to 38% of the gross cryptocurrency network hashrate on the same day, providing the grid with perfectly stable service (see Note 5). China can learn from this model and design a regionally differentiated policy based on the energy structure to regulate crypto mining. In the case of the hydroelectricity-rich Sichuan province, crypto mining activities should be permitted during wet season and be restricted or suspended during dry season. As for Inner Mongolia and Xinjiang, which are rich in wind power and solar power respectively, a dynamic electricity pricing mechanism should be implemented in line with the respective wind and daylight intensity.

In terms of policy, electricity pricing should be dynamically adjusted in accordance with grid power supply and demand. Miners should be charged lower prices when there is an abundant supply of renewable energy but higher prices in case of insufficient electricity supply to encourage power conservation. Such a policy framework involves the following four key elements. First, build a regulatory system that distinguishes crypto mining powered by renewable energy from that relying on fossil fuels, providing policy support for the former while imposing tight restrictions on the latter. Second, implement a dynamic electricity pricing mechanism; adjust pricing according to renewable energy supply and grid load; ensure priority for mining activities to consume excess renewable power; and require mining companies to install smart electricity meters and energy monitoring systems, reporting real-time energy usage data and ensuring regulatory transparency. Third, establish a carbon emissions responsibility mechanism, mandating mining companies to reach specific carbon intensity targets or purchase carbon quotas to offset their emissions. Fourth, consider incorporating crypto mining as part of the local energy transformation pilot schemes to evaluate its actual impact on promoting renewable energy adoption and minimizing carbon emissions.

If crypto mining is to be reinstated in the Mainland, priority could be given to the following three pilot regions.

Sichuan: Given the enormous hydroelectric power resources and low electricity prices in the province, crypto mining using hydroelectricity can be allowed during the wet season to utilize excess hydroelectricity.

Inner Mongolia: With the Autonomous Region’s wind-energy resources in abundance and wind energy in oversupply, a crypto mining industrial park designed for wind-energy utilization could be zoned to balance the volatility of wind energy by leveraging the flexibility of crypto mining.

Xinjiang:  Due to the Autonomous Region’s ample solar-energy resources and isolation from electricity load centres, crypto mining can serve not only as a means of utilizing surplus solar power but also as a new approach to linking crypto mining with industrial poverty alleviation.

Clear carbon emissions monitoring and evaluation systems should be in place in these pilot areas to ensure that crypto mining is instrumental in cutting carbon emissions instead of exacerbating environmental burdens.

A win-win delivering economic and social benefits

Apart from environmental benefits, resuming crypto mining can also bring economic and social value, e.g. opening up industrial growth avenues in less-developed western regions to create employment opportunities and generate local tax revenue. In addition, the electricity expenditures of crypto mining can provide a stable income source for renewable energy investments, thereby facilitating the growth of local clean-energy industries.

While the relationship between crypto mining and carbon emissions is more complex than it appears, a proper regulatory framework and pricing mechanism can support the adoption of renewable energy and help control of emissions. This necessitates that stakeholders move beyond black-and-white thinking and collaborate to strike a balance between technological innovation and environmental protection.

As a front-runner in the advancement of renewable energy and a major contributor to carbon abatement, China is well-positioned to explore the symbiotic development model for crypto mining and renewable energy, offering Chinese ideas and solutions for global energy transformation and climate change. Equating power consumption to carbon emission is a simplistic way of thinking. To limit carbon emissions in future, we need more innovative and complex solutions. With appropriate regulation, crypto mining can become a viable option worth reconsideration by China.

Note 1: https://www.coingecko.com/zh/global-charts

Note 2: https://www.gate.io/zh/learn/articles/60-bitcoin-mining-and-energy-consumption-statistics-for-2023-you-need-to-know/1265

Note 3: https://www.ndrc.gov.cn/xxgk/zcfb/tz/202109/t20210924_1297474.html?code=&state=123

Note 4: https://zjic.zj.gov.cn/ywdh/nyhj/202408/t20240812_22695794.shtml

Note 5: https://ceepr-mit-edu.eproxy.lib.hku.hk/climate-impacts-of-bitcoin-mining-in-the-u-s/

Dr Maurice Tse and Mr Clive Ho

19 March 2025

The American company OpenAI has dominated the Generative AI market since launching ChatGPT in 2022. AI-driven innovation has become the Holy Grail for players in the field. To curb China’s advancement in this area, the US has banned the export of high-end chips to the country and restricted Chinese users from accessing ChatGPT, making it necessary for Hong Kong users to rely on Virtual Private Networks to use the service. In 2024, a Chinese start-up turned the tables by introducing a generative AI model, DeepSeek, aiming for ChapGPT’s leading position.

By January 2025, DeepSeek took over ChatGPT’s spot as the most downloaded free app on Apple’s US App Store, causing Nvidia’s stock price to drop by 18% in a single day. Many cannot help but wonder: will DeepSeek simply replace ChatGPT?

New star making a splash

Neither DeepSeek nor ChatGPT is trained to play chess. Last month, however, under the arrangements of Levy Rozman, an American chess International Master, they did just that in a chess game. In the end, DeepSeek pulled a trick from Sun Zi’s Art of War and, with a clever stratagem, turned the tide to clinch victory against ChatGPT from the verge of defeat.

DeepSeek was developed in Hangzhou in 2023 by the Chinese hedge fund High-Flyer. Launched on 20 January 2025, the new open-source model DeepSeek-R1 provides free services as both a mobile application and desktop version, immediately attracting attention in global AI circles. The new model, trained on the database available up to July 2024, can integrate updated data from web sources as necessary.

DeepSeek-R1 is equipped with AI assistant functions, including writing song lyrics and making business development plans, or even preparing a recipe based on the contents of a fridge. With the ability to communicate in multiple languages, it excels in English and Chinese in particular. As R1 reveals its chain of thought when answering questions, making the reasoning process transparent, users gain a clear understanding of the AI system’s logic. This enables them to learn in reverse how to deconstruct complex problems and arrive at problem-solving solutions through practice. From International Mathematical Olympiad (IMO) questions to gossip news, or from quantum physics to Chinese literature, users can visualize the thinking process of an AI machine. The resulting cognitive impact is conducive to providing a good starting point for humans to enrich their own thinking through AI.

Works wonders at minimal cost

DeepSeek is not only lauded as “true OpenAI” because of its open-source model but also impresses with its basic model V3, which took merely two months to train and just US$5.58 million to develop. As reported by Sina Finance, the 10,000-plus graphics processing units (subsequently increasing to 50,000) purchased by DeepSeek cost far less than those acquired by renowned labs such as OpenAI and Google.

The fact that the US has been able to produce high-performance chips while denying China access to related technologies has long been regarded as an enormous advantage in the AI race. DeepSeek’s success offers food for thought for the industry, prompting reflection on whether improving high-end chips is the only key to perfecting AI models. By leveraging just 2,000 of Nvidia’s H800 chips, DeepSeek has been able to rival OpenAI’s capability with one-tenth of the latter’s training cost. This is a testament to DeepSeek’s excellent algorithmic design and effective resource utilization.

Two rival models going toe to toe

The primary difference between ChatGPT and DeepSeek is that the former is tailored for conversational applications, focusing on task-specific intelligence within the realm of Artificial Narrow Intelligence whereas DeepSeek aims to achieve Artificial General Intelligence (AGI). Language support is another major difference. DeepSeek is a Chinese company with an emphasis on Chinese language and culture, rendering it a strong contender in the Chinese market. In comparison, as the current market leader supporting multiple languages, ChatGPT holds a greater advantage with its wide user base among professionals worldwide.

OpenAI has built a powerful ecosystem around ChatGPT, encompassing application interfaces, plug-ins, and partnerships with tech giants such as Microsoft. This household name in AI benefits from a community of active developers who lend support to its continued improvement and innovation.

Meanwhile, DeepSeek is no slouch as a newcomer, making it ChatGPT’s formidable rival. Without substantial fine-tuning, ChatGPT may struggle to ensure the same level of accuracy in a specialized environment. DeepSeek is customized to efficiently handle specific datasets or domains, especially in sectors such as finance, healthcare, or legal documents. Its long-term goal of attaining AGI may facilitate the advancement of AI systems with even greater adaptability in future.

Facing direct headwinds as a force against the flow

Despite DeepSeek’s advanced technology, its data source has been mired in controversy. The chatbot has disclosed in its responses to users that its training process may have utilized ChatGPT’s output data, potentially violating OpenAI’s agreement.

Furthermore, DeepSeek is subject to stricter censorship in terms of certain content areas (e.g. personal finance), which limits some users’ experiences. In China, internet services are required to embody “core socialist values”. This means that Mainland chatbots must comply with government regulations regarding politically sensitive issues.

According to NewsGuard’s chatbot audit report in January 2025, DeepSeek could only provide accurate information about news topics 17% of the time, ranking it tied for 10th out of 11. Notably, all the other chatbots tested were Western models. The report points out that DeepSeek’s fail rate was 83%, compared with the average fail rate of 62% among other AI models. Microsoft and OpenAI have launched an investigation into whether any close associates of DeepSeek have used dishonest means to steal large amounts of data via the OpenAI application interface. Despite the fact that DeepSeek has readily become the top downloaded app in Apple’s App Store, its technology is inevitably called into question by competitors, and the pressure is piling on the US in the AI domain.

The introduction of the DeepSeek-R1 model in early 2025 coincided with America imposing tighter restrictions on the export of AI technology to China. Companies such as OpenAI have already warned that China’s AI models could catch up with or even surpass their American counterparts in future. China is explicitly aiming to become the world leader in AI technology by 2030 and plans to invest tens of billions of US dollars to support the sector. Currently, at least three Chinese labs, namely DeepSeek, Alibaba, and Moonshot AI, have launched AI models reportedly on a par with OpenAI o1.

DeepSeek’s success demonstrates that Chinese companies have overcome the hurdles involved, indicating that their development teams are fast approaching cutting-edge technological levels. The fact that the R1 model can be run on a personal computer greatly contributes to the popularization of AI applications. Its exceptional performance also means that powerful reasoning systems will soon become widespread. In view of the model’s ability to run on local hardware, it is likely to bypass America’s export controls.

With the spectacular debut of DeepSeek, AI has triggered not only revolutionary change. While creating unprecedented growth opportunities in law, accounting, healthcare, education, and transport, it has also given rise to unignorable risks.

Hidden risks not to be overlooked

Since AI systems are often trained on historical data, they are inevitably prone to the influence of embedded biases. Without proper action to address these biases, overt discrimination could occur. It is imperative to adopt both preventive and corrective measures, particularly in recruitment, lending, healthcare, and law enforcement. Amidst the rampant hacker attacks today, AI systems are susceptible to malicious data input due to adversarial sample attacks, data pollution, or model theft, which can result in erroneous or harmful decisions. This could end up with dire consequences in life- or property-critical domains, such as self-driving vehicles or financial systems.

Moreover, AI can also be used to generate convincing fraudulent content. For instance, increasingly prevalent deepfakes have fuelled the spread of false information. Deepfake videos of political figures can even sway elections or provoke social upheaval, seriously compromising institutional credibility and social cohesion. The weaponization of AI, whether applied to drones or cyber warfare, will further augment the potential for abuse and pose a formidable threat to global security.

Undoubtedly, the gains and risks of AI are equally far-reaching. For this very reason, it is essential to strike a balance between innovation and responsibility in the face of such challenges. Only through ethical governance, concerted efforts, and mutual vigilance by the international community can we fully harness AI’s power to benefit society and the economy while minimizing its dangers.

Dr Stephen Chiu

12 March 2025

US president Donald Trump recently announced significantly raising import tariffs on Canadian and Mexican products, with a surge as high as 25% on vehicles and automotive parts. Despite being aimed at relocating American manufacturing industries and minimizing trade deficits, this could actually trigger a series of profound negative impacts.

Below is an analysis of how this policy will undermine economic cooperation and productivity in North America, particularly within the automotive industry.

Cross-border interconnected industrial chains among US, Canada, and Mexico

The North American auto industry chain forms a highly-integrated, multinational supply system that fully leverages the resource and workforce advantages of America, Canada, and Mexico. Let’s take Ford’s popular  F-150 pickup truck as an example of the massive cross-border transportation of parts and components involved in the automotive manufacturing process. This demonstrates the industrial coordination under the North American free-trade system.

Engines from America and Canada: Some engines, such as EcoBoost V6, are manufactured in plants in Michigan, US while the rest, e.g. V8, are manufactured in factories in Windsor, Canada. These products are then transported to the US for assembly.

Transmission system from America and Mexico: the 10-speed automatic transmission system for the F-150 is produced in plants in Michigan, US and Chihuahua, Mexico. Leveraging Mexico’s cost-efficient labour force, a portion of the automotive components are transported to the US for assembly.

Automotive body parts from Mexico: Metal finishing in Mexico is cost-effective and technologically mature. The completed aluminium body panels are also transported to the US for final assembly.

Electronic components from Canada and Mexico: High-tech electronic control units are sourced from Canada while labour-intensive wire harnesses are made in Mexico.

Final assembly in America: Final assembly for F-150 is conducted in factories in Dearborn, Michigan and Kansas City, Missouri.

Throughout the entire manufacturing process, parts and components are transported across borders multiple times. For instance, under special circumstances, parts and components manufactured in Mexico may be shipped to Canada for further processing. Such cross-border movements highlight the substantial degree of coordination within North American automotive industry chains.

As a result of the tax cumulative effect of this cross-border manufacturing approach, the tax burden on manufacturers is actually much higher than the Trump administration’s 25% tariff may suggest on the surface. Suppose the components are manufactured in the US at a cost of US$10. They are then delivered to Mexico for processing at an additional cost of US$10 before being returned to the US for final assembly. The tariff is US$5 at a 25% tax rate. Since the processing vale in Mexico is US$10, the tariff amounts to 50% of the processing cost in Mexico. Clearly, while the tariff rate is set at 25%, the actual cost burden across different parts of the supply chain is much heavier, especially for the labour-intensive operators in Mexico.

The above impact will not only increase the overall manufacturing cost but will also pose a challenge to the North American free-trade cooperation system. Enterprises are compelled to reconfigure their supply chain layouts, compromising overall economic efficiency.

Economies of scale hindered by high tariffs

From an economic perspective, the tariff policy will push up manufacturing costs in the two following aspects. First, direct costs: tariffs acting as an additional expense directly escalate the costs of manufacturing parts and components. For example, wire harnesses or transmission systems made in Mexico will become more expensive as a result of the tariffs. These additional costs will eventually be reflected in the sales price.

Second, indirect costs: New Trade Theory, pioneered by Paul Krugman, emphasizes that one of the advantages of international trade is that it enables companies to fully leverage economies of scale, i.e. the unit cost of production diminishes with expanding output. However, the tariff policy will lead to the automotive industry’s scale of production shrinking. As rising prices of components suppress demand, overall production will dwindle as a result. The average production costs of enterprises will shoot up as output plummets, further weakening their competitiveness.

In the aforementioned Ford Motor case, if the company cannot afford the tariffs, it will not be able to import enough low-cost components from Mexican plants, causing America’s manufacturing capacity to drop. This means that Ford will no longer be able to manufacture best-selling models like the F-150 at low costs and will, therefore, struggle to maintain its price competitiveness in the market.

Talk of American manufacturing chains returning just wishful thinking

A core objective of Trump’s tariff policy is to bring manufacturing back to the US. Unfortunately, this objective is neither practical nor economically efficient.

One formidable obstacle to restoring a large number of jobs back to the US is the inadequate labour supply. Given the historically low unemployment rate and current shortage of skilled workers for technical positions in manufacturing, any attempt to move all cross-border production chains to America will lead to manpower shortages in other industries, further driving up wages and inflation rates.

Moreover, labour costs in the US are much higher compared with countries like Mexico. Even if enterprises establish factories on American soil, they would be unable to replicate the same low-cost operations achievable in Mexico. This will severely weaken the competitiveness of American automobiles in markets worldwide.

The international division of labour in manufacturing lies in professionalism and maximizing efficiency. While Mexico specializes in the production of labour-intensive parts and components, Canada and the US focus on high value-added research and final assembly. Such a division of labour approach has equipped the North American automotive industry with global competitiveness. Nevertheless, if all production activities were concentrated in the US, companies would not only lose the advantages of economies of scale, but productivity would also sharply decline, ultimately dealing a blow to economic coordination throughout the North American region.

The lesson from cutting off one’s nose to spite one’s face

Trump’s tariff policy serves as a reminder of the Smoot-Hawley Tariff Act from the last century. Aimed at protecting American agriculture and industries, the Act was passed in 1930 and imposed high tariffs on over 20,000 imported goods. Despite its implementation, the legislature not only failed to achieve its purpose but also triggered retaliatory tariffs from other countries, significantly slashing international trade and further exacerbating the Great Depression.

The ultimate failure of the Act taught us a valuable lesson: unilaterally raising tariffs will not merely harm the domestic economy but also compromise the stability of the global economic system. Trump’s tariff policy may well repeat the same mistakes, with a heavy impact on the overall economic coordination in the North American free-trade region, especially when Canada and Mexico take punitive measures.

While Trump’s policy to levy tariffs on imports from Canada and Mexico seeks to bring manufacturing jobs back to the country, it completely underestimates the cumulative burden on producers, not to mention the production cost hikes due to direct costs and the loss of economies of scale. The writing has long been on the wall regarding the Smoot-Hawley Tariff Act: without due consideration for the complexity of the global supply chains, the tariff policy is set to inflict irreparable damage on both the domestic economy and the international trade system.

Given the advanced development of globalization today, international trade collaboration remains crucial for enhancing productivity and economic growth. The US should re-examine its trade policy to avoid making the wrong moves again and to maintain the competitiveness and resilience of the North American automotive industry by deepening regional cooperation.

Professor Yulin Fang and Professor Xiaojie Zhang

5 March 2025

With the deep integration between big data and artificial intelligence (AI) today, data has become a core asset for corporate development. Continued technological breakthroughs in and the extensive application of AI have catapulted the value of data to unprecedented heights. The vast majority of companies are striving to derive valuable business insights from data analytics.

According to a study by Fortune Business Insights, the estimated market size of data analytics, valued at US$41.05 billion in 2022, is expected to reach US$279.31 billion by 2030, with an average annual growth rate of 27.3% between 2023 and 2030. This goes to show the rising importance of data analytics teams.

Persistent challenges in the age of data

By extracting key information from a sea of data, an outstanding data analytics team can provide a company with robust support for strategic decision-making, resulting in enhanced operational efficiency and a strengthened competitive edge in the market. Through in-depth analytics of sales data, enterprises can get a precise grasp of market demand and to optimize product planning. Through data mining of user behaviour, companies can achieve personalized marketing while boosting user satisfaction and loyalty.

However, data analytics-team management is beset with problems nowadays. Despite injecting substantial resources into big data, AI, and machine learning, many companies have not been able to gain sustained commercial value from their investments. A report by research institution Gartner reveals that more than half of Chief Marketing Officers are dissatisfied with their company’s marketing data analytics team.

The reason for the above phenomenon is that big data analytics involves a complex system. Its success relies not only on the collection, storage, and management of data assets and the use of appropriate analytics tools but also on an efficient information interaction model and a synergistic mechanism among team members. Given the multitude of interwoven factors, including data, analytics tools, team operations, and corporate environment, any single issue can result in the failure of the entire analytics project. It would be advisable for enterprises to have a viable management strategy in place to facilitate the effectiveness of their analytics team and maximize data value.

Pivotal agenda for building an analytics team

As a matter of fact, companies are not at a dead end amid impediments to big data analytics. The solution is to form an expert data analytics team to unlock the potential of big data and maximize business value. Seven vital points are outlined below.

1. Consolidating data foundation and safeguarding data quality (see Note 1). This involves ensuring diverse and reliable sources, as well as instantaneous data updating for the analytics team. Data analytics encompasses multiple aspects, such as transactions, user behaviour, and market movements, thereby laying a solid foundation for subsequent data mining. When devising a scientific data analytics process and a quality management system, it is essential to establish data quality monitoring indicators and regularly evaluate data quality. This system is conducive to maintaining data accuracy and consistency.
2. Profound operational alignments and precise identification of business requirements (see Note 1). The analytics team should be in close collaboration with operations departments and be fully aware of their needs in order to provide targeted data analytics and solutions. The team must be properly positioned within the company to enable it to partner closely with operations departments. Through data collection, collation, and analysis, the analytics team can support the day-to-day business operations and offer the company tailor-made data analytics services. This can be accomplished by acquiring an in-depth understanding of the operations departments’ workflows and needs.
3. Establishing a clear communication mechanism and refining the management structure (see Note 2). It is crucial to map out a scientifically sound team management structure, where a well-established communication framework is indispensable. A dedicated data management department should be set up to coordinate data collection, collation, and analytics, delineating the responsibilities of each department engaged in the data management process. Good communication among departments helps to prevent responsibility shirking and work process duplication, thus enhancing work efficiency. Only through effective communication can all departments perform at their best, enabling optimal resource allocation and ensuring the smooth implementation of data analytics projects.
4. Shaping transactive memories and strengthening coordination among team members (see Note 2). Since members of an analytics team each possess unique expertise and knowledge, it is necessary for them to thoroughly understand each other’s strengths to foster fruitful collaboration in the complex process of data analytics. In a comprehensive data analytics project, when the data collection members recognize the advantages of the data sources, the analytics members are well versed in various methods, and tasks are properly divided through a transactive memory system, the team will progressively upgrade its overall performance.
5. Stimulating creative integration and empowering corporate development (see Note 3). With abundant data sources, multiple analytics methods, and business knowledge, an analytics team will be equipped to implement creative integration. Take analytics on market conditions and competition, for example. Data on market dynamics and competitors alone will not suffice. This information should be creatively integrated with the company’s distinctive features and strategic goals to formulate a unique marketing strategy for its business.
6. Boosting knowledge management and facilitating collaborative creation (see Note 4). A high-impact analytics team is like a treasure trove of knowledge, with a wealth of professional expertise and data analytics skills accumulated from long-term practice. A sound knowledge management system is the key to this treasure trove, harnessing the team’s knowledge for efficient integration and innovative co-creation. Collating and archiving data analytics experience from a series of projects can develop an “intelligence toolkit”, paving the way for team members to draw inspiration and insights from it while continuously increasing professional capabilities. For new members, this undoubtedly serves as a “green channel” for integrating into the team, providing a shortcut to understanding the work mode and necessary knowledge, thereby empowering them to contribute to the team as quickly as possible.
7. Leveraging advanced technology and enhancing collaboration efficiency (see Note 3). An analytics team must be proficient in deploying various project management tools, such as Grantt charts, and the Kanban management system, to suitably chart project schedules, clarify duty allocations, monitor project progress, enhance the team’s collaboration efficiency and execution capabilities. This approach facilitates stable advancement and stimulates business development for the company. During the collaboration process, coordination techniques, e.g. real-time communication tools and online document-sharing platforms, should be used to break down communication barriers and achieve real-time sharing of information and synergy of efforts. With the geometric growth in generative AI technology, data analytics teams should strive to integrate it into their workflows, facilitating synergies between humans and generative AI to enhance efficiency and creativity.

Converging knowledge and action for a boundless future

Looking ahead, big data analytics technology is set to demonstrate more rapid growth and bring endless opportunities and possibilities for enterprises. In the face of this technological surge, promoting collaboration to refine the management coordination mechanisms of data analytics teams and build outstanding teams will enable all parties involved to fully tap into the value of big data. These concerted efforts will boost AI-driven efficiency, ushering in a new era of advancement.

Note 1: Zhang, X., Tian, F., Fang, Y., and Shen, H. “How to Promote Business Analytics Project Effectiveness: A Cross-disciplinary Bibliometric Analysis”. Industrial Management & Data Systems, under 1st round of revision.

Note 2: Fang, Y., Neufeld, D., and Zhang, X. 2022. “Knowledge coordination via digital artifacts in highly dispersed teams”. Information Systems Journal 32(3): pp. 520–43.

Note 3: Zhang, X., Fang, Y., Zhou, J. and Lim, KH. 2025. “How Collaboration Technology Use Affects IT Project Team Creativity: Integrating Team Knowledge and Creative Synthesis Perspectives”. MIS Quarterly, forthcoming.

Note 4: He, W., Hsieh, JJ., Schroeder, A., and Fang, Y. 2022. “Attaining Individual Creativity and Performance in Multi-Disciplinary and Geographically-Distributed IT Project Teams: The Role of Transactive Memory Systems”. MIS Quarterly 46(2): pp. 1035–72.

Dr Y.F. LUK

26 February 2025

Donald Trump’s comeback as the US President is characterized by the rapid introduction of various policies, which have been boldly and strictly implemented within just over a month. Seldom do newly inaugurated heads of state take such drastic and overwhelming actions. New domestic policies encompass the mass deportation of illegal immigrants and the retrenchment in federal agencies and jobs, including the virtually overnight closure of the United States Agency for International Development (USAID). His international policies are even more “revolutionary”. Beyond the sudden US-Russia détente, America’s about-face in its relations with long-standing European allies has utterly upended the international order built over the past eight decades.

One of the policy changes that has received less attention is the executive order signed by Trump on 3 February 2025 directing the Treasury and Commerce secretaries to propose a plan within 90 days for the creation of a US sovereign wealth fund (SWF), outlining funding mechanisms, investment strategies, governance model, etc. In his positive response, Treasury Secretary Scott Bessent believed that preparation work would be finished and the fund set up within 12 months.

In a nutshell, an SWF is a state- or government-owned and managed investment entity that allocates assets to domestic and foreign markets. SWFs are typically funded by revenues from natural resource sales, such as oil and minerals, e.g. those of Norway and Saudi Arabia. Others are financed by trade surpluses accumulated over the years, as in the cases of China and Singapore.

As for the purposes of SWFs, they primarily serve to finance national development objectives, stabilize government finances, and enhance the efficiency of national asset management. For instance, to raise its profile in international sports, Saudi Arabia has leveraged its Public Investment Fund to acquire football clubs, attract high-profile players like Cristiano Ronaldo with lucrative salaries, and sponsor the Women’s Tennis Tour (WTA) Finals a few months ago. These are some of the more entertainment-oriented events. Yet more often than not, SWFs are channelled into conventional areas, including infrastructure development projects, industrial policies, and financial investments.

Countering China while bypassing domestic restraints

The SWF executive order signed by Trump is said to be intended to strengthen financial sustainability, alleviate the tax burden on American families and small enterprises, and bolster US economic and strategic dominance on the international stage. This is just a broad-brush description and the details probably remain to be disclosed in the plan to be announced within 90 days. However, how these objectives related to fiscal capability, financial sustainability, and tax liability can be achieved is an issue for further examination. Simply put, there are two possible scenarios. First, upon the establishment of the SWF, the government injects new capital into the fund without changing the existing financial arrangements. This will, of course, mean more money in the government coffers, yet questions remain about whether the new capital should be used to set up an SWF or put to better use, such as reducing government debts and future interest payments. Second, with no new capital at its disposal, the federal government needs to raise funds through taxes or from the lending markets to finance the SWF. Should the returns on government investments be comparable to market returns, the government would be better off keeping funds invested in the market rather than risking a crowding-out effect by unnecessarily initiating an SWF.

In both scenarios, the SWF will produce several economic effects. Firstly, the government is unwilling nor does it find it appropriate to transfer the capital or resources managed through an SWF to the market. Secondly, the SWF can save on investment costs or boost returns due to its colossal size. Thirdly, market investments do not take into consideration social value, making private returns often lower than social returns, which is regarded a market failure. Nevertheless, an SWF has the flexibility to invest in areas where the market is absent. That is one reason why some commentators are in favour of the US establishing an SWF. They believe that America’s lack of an overseas infrastructure investment mechanism to counter China’s Belt and Road Initiative has undermined its global influence. The US government may have its International Development Finance Corporation to handle related matters but, subject to existing financial regulations, it would be hard-pressed to make investment decisions aimed at enhancing US global influence. The new SWF could bypass such limitations.

Apart from getting greater freedom to counter China in the international market, Trump’s directive about an SWF is in a way his attempt to start anew. This move would allow him to break free from certain financial constraints and minimize intervention from Congress while deflecting criticisms about the accelerating fiscal deficit. Furthermore, it can also serve to avert the uncertainty caused by the conflict between the Republican and Democratic parties in the event of an impending federal government’s debt ceiling. Regarding the legislation governing the SWF’s establishment, Congress has little reason to relinquish its regulatory authority. That being the case, Trump’s strong leadership and Republican dominance in both houses of Congress are poised to shift the power balance in support of his governance in the coming years.

While no mention is made in the executive order of the size of the proposed US SWF, a ballpark estimate of US$2 trillion is often cited in the commentariat. In comparison with the world’s three largest SWFs, namely Norway Government Pension Fund Global (US$1.74 trillion), China Investment Corporation (US$1.33 trillion), and SAFE Investment Company Limited (US$1.09 trillion), this is far from an insignificant sum. Where will this massive amount come from? When pressed for an answer, Trump said “from tariffs and other intelligent things”. What “intelligent things” may be revealed in due course. Yet, using tariff revenues as the primary funding mechanism for an SWF would be a time-consuming process. In 2024, the federal government’s total income stood at US$4.92 trillion, with tariffs accounting for merely US$80 billion. Clearly, tariffs are a long way from being a viable primary funding source for the proposed SWF. Compared with his other policies during the presidential campaign, his braggadocio about tariff hikes has so far turned out to be much ado about nothing. Even if the tariff revenue expanded fivefold in 2024, the income would still be less than US$400 billion. And this would have triggered a certain level of inflation and recession. On the other hand, raising other taxes would mean reneging on his campaign promise of tax reduction and the Republican economic philosophy. Bond issuance is likewise not an ideal financing option. For one thing, the US has already racked up a debt of US$36 trillion, which is a major concern in the financial markets worldwide. For another thing, the surging interest costs of bond issuance have compressed the future fiscal space of the American government.

It is the Treasury Secretary Scott Bessent, after all, who provides a clearer picture by stating, “We are going to monetize the asset side of the US balance sheet”. According to the information supplemented by the White House on the occasion of his remark, the value of assets directly held by the federal government is US$5.7 trillion. Nevertheless, as some commentators have pointed out, cash and gold only contribute US$1.2 trillion to these assets.

Official gold price fixed at US$42.22 per ounce

The amount of gold held by the US government is 261.6 million troy ounces but on the Treasury Department’s balance sheet, the gold is still priced at the same level of US$42.22 per troy ounce in the 1970s. Under the Bretton Woods system, the gold price was originally set at US$35 per ounce. In August 1971, after America unilaterally delinked the dollar from gold, there were still international efforts to save the Bretton Woods system by devaluing the greenback against gold but they were all in vain. The US official gold price has since stood at US$42.22 per ounce. Now that the price is approaching US$3,000, should the US government’s gold reserves be priced in accordance with market value, the asset value on its balance sheet would expand by US$773.7 billion. This may be what Scott Bessent meant by his remark: “to monetize the asset side of the US balance sheet” as a funding source for the SWF plan.

By revaluing its gold reserves, the US can generate substantial capital at no cost, which can be directed into the new SWF or deployed for other purposes. Stephen Miran, Trump’s nominee to chair the Council of Economic Advisors, once suggested that to devalue the US dollar, gold could be sold to buy foreign currencies. There is no telling how seriously this piece of advice should be taken. Given that many central banks have been buying gold heavily in recent years, the US government’s operations with its gold reserves are likely to prompt considerations for reincorporating gold into the monetary system.

Professor Heiwai Tang, Ms Shuyi Long, and Beining Liu
19 February 2025

The “Flying Geese Paradigm” was first proposed by Japanese economist Kaname Akamatsu in the 1930s to describe the transition of labour-intensive manufacturing in Southeast countries towards industries that are more capital- and technology-intensive. The concept reveals that Japan, as the “lead goose”, led the way in completing industrialization, driving the development of other countries through foreign direct investment. This created an interdependent and gradient-advancing system of division of labour in the East Asian economy, characterized by a structure similar to a formation of flying geese. The theory has been validated through economic practice in the decades after the Second World War. Japanese corporations, through their global expansion strategy, not only shaped the manufacturing landscape of East Asia but also set an example for the eventual globalization of enterprises in China, South Korea, and beyond.

Nowadays, more and more countries have come to realize the importance of industrial policy and government intervention. Fierce competition has kept nations vigilant against one another in the fields of business and technology. Apparently, the pivotal theory is no longer compatible with the current economic realities. As the world economy is once again at the crossroads of change: automation, artificial intelligence, geopolitical conflicts, and supply chain reorganization are redrawing the map of global manufacturing. Is the Flying Geese Paradigm, once the theory governing relocation of industries, still applicable? How useful is it as a reference for understanding China’s economic transformation today and its economic interactions with other countries?

In 1985, to address the interest rate hike and continued appreciation of the greenback, the US convened the Group of Five (the US, Japan, the UK, France, and West Germany) at the Plaza Hotel in New York to sign the Plaza Accord. This far-reaching agreement sought to achieve an orderly devaluation of the US dollar against the world’s major currencies. Japan was subjected to the greatest impact, with the Japanese yen appreciating by over 100% within two years. The price competitiveness of Japanese products declined as a result, dealing a serious blow to Japanese exports, particularly in the automobile and electronics industries.

In response to the impact of exchange rates and surging domestic costs, Japanese corporations embarked on an unprecedented wave of overseas production relocation: leading companies in automobile, electronics, and home appliances industries moved their production bases to other places in the region, primarily South Korea, Taiwan, Hong Kong, and Singapore, which are subsequently known as the “Four Asian Little Dragons”. This strategy not only aimed to mitigate exchange rate fluctuations but was also based on the logic of the Flying Geese Paradigm. After completing its industrial upgrading on home soil, these enterprises shifted low value-added manufacturing segments to lower-cost regions. Meanwhile, the latecomer economies were able to gradually advance their own industrialization process by hosting the Japanese enterprises. By the early 1990s, with the economic rise of the Four Little Dragons, the Japanese manufacturing industry further expanded towards Southeast Asia. Thanks to their proximity with supportive policies, Thailand, Malaysia, and Indonesia became new destinations for this expansion.

Take Toyota’s marketing strategy for North America, for example. The company integrated into the US industrial chain by establishing a production base and investing in local research and development (R&D). First launched in the US in the 1980s, the Toyota Camry saloon was not only popular in the North American market but was also the company’s first flagship model for the international market. Toyota surpassed Ford and GM in the early 2000s and 2008 respectively to become the world’s best-selling vehicle manufacturer. This goes to show the high returns from focus on brand-building strategy.

Japanese corporate globalization mired in the “Lost Decade”

While Japanese enterprises gained new development opportunities through globalization, the Japanese economy paid a heavy price in the process. As a result of the massive relocation of production lines overseas, Japan’s manufacturing industry was gradually hollowing out. Given limited employment prospects, stagnant investment in manufacturing, and a prolonged depression in gross domestic product growth, the Japanese economy entered the “Lost Decade”. Japanese companies responded by reflecting on the sustainability of their cost-driven globalization strategy. With the advent of the 21st century, the globalization strategy of Japanese enterprises began to change alongside the shifting global economic landscape. Particular attention was focused on brand building and technology acquisition, as evidenced by large-scale overseas mergers and acquisitions as well as R&D input.

Plagued by continued external suppression from the US and sluggish domestic market growth in recent years, Mainland Chinese companies find themselves at the crossroads of change, similar to the situation encountered by Japanese corporations four decades ago. However, with the world economy having undergone earth-shattering changes in the intervening years, is the Flying Geese theory still applicable today? Can Mainland companies likewise become the “lead goose” and create a breakthrough window for Asian countries?

It is noteworthy that the Flying Geese model in the 1980s arose from a truly unique historical period. The “gradient advancement” described by this theory was, first and foremost, dependent on international cooperation during the heyday of globalization over the past few decades. Not only did developed and emerging markets need to collaborate with each other, but they also had to set aside competition and animosity. Amid the escalating momentum of deglobalization in the current climate, protectionism and unstable geopolitics are likely to set a thorny path for industrial and technological transfer.

Automated production a hurdle to the “Flying Geese”

In the wake of the 2010s, faced with the pressures of protectionism, Japan, once the “lead goose”, initiated structural adjustments to its globalization strategy, sparking trends of industrial-chain reshoring and relocalization among major Japanese corporations. In 2022, Sony announced its partnership with Taiwan Semiconductor Manufacturing Company Limited (TSMC) to build a chip manufacturing plant in Kyushu. Instead of allocating resources to the next development stage as originally planned, developed countries started to promote industrial-chain reshoring or near-shoring in order to manage the risk resilience of their industrial-chains. Meanwhile, developing countries in dire need of foreign direct investment and technology transfer are left with dwindling economic development opportunities.

The large-scale application of technological advancements and automated production is another reason why the Flying Geese Paradigm is challenged. Automated production, which can drastically reduce the cost of labour-intensive products, eliminates a key element of this theory: the incentive for companies to shift production chains to lower-cost regions when robots can replace cheap labour. A study by the World Bank finds that once the number of industrial robots in a country exceeds a certain threshold, there is a negative correlation with foreign direct investment in that country. In other words, the more well-developed automated production is in a country, the lower its motivation to invest in industrial transfer to other countries.

Although the China-US rivalry has posed multiple hurdles to international trade, it has also injected fresh vitality into the long-static global economic landscape. America’s attempt to control the rise of China through various policies has prompted businesses to adopt the “China + 1” strategy. Designed to diversify geopolitical risks by balancing a foothold in the Chinese market and maintaining a production base there while seeking suppliers in nearby countries, this strategy has enabled more and more small to medium economies to participate in global supply chains. The main beneficiaries include the Association of Southeast Asian Nations (ASEAN) members such as Vietnam, Thailand, and Indonesia, as well as Central and South American countries such as Mexico and Brazil. For instance, in the ASEAN region, China’s direct investment exceeded US$25 billion in 2023, representing a 34.7% year-on-year increase. The diversification of supply chains, which have long been dominated by a few major trading countries, is conducive to the establishment of diverse international rules and systems.

Growth impetus for emerging markets from overseas expansion of Chinese companies
The global outreach of Chinese companies has also introduced a new business model for developing countries. The latest study by Harvard Business School Professor Josh Lerner et al. (see Note) reveals that with the growth of Chinese enterprises in the past 20 years, their business model has gradually exerted a positive influence on emerging markets. Chinese venture capitalists not only play an increasingly significant role in global investment but have also considerably expanded their investments in emerging markets. In addition, since these investments are typically channelled into sectors dominated by Chinese companies, including educational technology, digital technology, and financial technology, these sectors in emerging markets worldwide experience notably faster development compared to other sectors.

More importantly, in sectors dominated by Chinese enterprises, local investors in emerging markets have also begun to adopt these companies’ investment and business models in making investments. This shift has strengthened their commitment to entrepreneurship and innovation. The study by Lerner et al. concludes that China’s outward investments have a clearly positive impact on the economic and social development of developing countries.

In comparison with the period when Japanese corporations relocated production abroad to peripheral countries, the changes in funding and business models brought to emerging markets by Chinese enterprises and investors represent a modern interpretation of the Flying Geese theory. Despite the changing times, international cooperation is still a core element of economic growth. The global economy is also bound to progress further towards diversification and technology. Countries, both large and small, along with businesses across various sectors, will have greater opportunities to join supply chains worldwide.

Diversification, nevertheless, cannot be taken for granted but requires concerted efforts by all nations to achieve. The Donald Trump administration’s high-tariff policy against US trading partners and even American allies in recent months is likely to trigger a global trade war. Once the global economy is embroiled in the trade war, the cost of multilateral trade will become even higher. To safeguard their own interests, different countries may be inclined to cut outward investments and cooperation. In the long run, this will only undermine their confidence in multinational cooperation, making development even harder for low- to middle-income countries.

Note: Josh Lerner, Junxi Liu, Jacob Moscona, and David Y. Yang. 2024. “Appropriate Entrepreneurship? The Rise of China and the Developing World”, Havard University Working Paper.

踏入2025年，中國人工智能（AI）領域取得標誌性突破。深度求索（DeepSeek）推出的R1模型，以557萬美元的超低訓練成本（僅為GPT-4的5%）媲美美國OpenAI o1模型的性能表現。

從「追趕」到「並跑」的轉折

R1模型這一成就的關鍵在於其獨創性的科技路徑：利用純強化學習技術擺脫對監督微調數據的依賴，藉助「群體相對策略優化」（Group Relative Policy Optimization）演算法實現推理能力的自主進化。與此同時，其動態精度調節技術有助華為旗下昇騰AI平台的算力成本降低70%，而性能損失則只低於5%。尤其觸目的是，R1的開源模式打破了算力壟斷，支持開發者將大模型能力蒸餾為1.5B參數的小型版本，因而在東南亞、中東等新興市場迅速滲透。

美國OpenAI對此迅速作出回應，上月緊急發布了o3-mini推理模型。該模型主打可調節推理強度，在高強度模式下對科學、科技、工程、數學（STEM）領域展現出無可匹敵的統治力——在AIME 2024數學測試中取得87.3%的準確率，勝過R1的79.8%。其物理類比能力也被開發者譽為「教科書級別」，例如在四維超立方體內反彈小球的程式設計中，o3-mini對幾何結構與運動軌跡的分析精度高達97%。然而，其輸入輸出成本仍是R1的8倍，凸顯出閉源模式的效率瓶頸。

生態構建與地緣滲透

中國市場的獨特優勢體現在場景驅動的生態整合上。透過開放應用程式介面與行業定製策略，國內企業已深度嵌入醫療、教育、金融等垂直領域。以教育界為例，作業幫的大型語言模型通過逐步分析，提升中學生數學解題的效率達37%。這種本土化適配能力預計將推動內地AI大模型市場在2026年突破700億元人民幣，使用者每周平均使用頻率達4.5次，遠超歐美市場的2.1次。

美國企業則試圖依靠多模態AI功能穩固霸主地位。GPT-4o通過文本、語音和視覺的融合能力，在微軟Office智能助手中自動生成會議紀要和簡報排版等功能，企業用戶續費率高達89%。另一方面，o3-mini在非英語市場的表現卻顯露不足，在中文理解一環的準確率只及百度文心一言的82%，拓展東南亞市場的計劃也就舉步維艱。更為嚴峻的是，OpenAI預計2023至2028年的累計虧損將高達440億美元；公司行政總裁阿爾特曼（Sam Altman）罕見地承認「閉源策略可能犯下歷史性錯誤」，並啟動o3-mini免費開放，以吸引開發者。

美國科技霸權重構

近年來，美國在AI大模型領域憑藉閉源商業模式和專有技術，一直維持國際領先地位；然而，中國 DeepSeek-R1冒起，馬上引發一場嶄新的科技變革。R1模型通過開放創新的方式，達至核心演算法和模型共用，有利於前沿技術迅速反覆運算，同時也大幅降低科技壁壘。這一舉措對美國長期以來依賴封閉體系構建的科技霸權造成直接衝擊。

DeepSeek-R1 的開源，使環球科研機構、企業和獨立開發者能夠在一個開放透明的平台上，不但共同改進和應用AI技術，而且逐漸重塑科技標準和行業規則。各方藉助這一開源平台進行研發合作，加速達至創新成果和構建自主智能知識產權，從而開創國際科技生態中多極互動的新局面。這種模式正在促使科技管治由單一壟斷轉向開放協同，並為國際社會建立科技公平競爭的規則樹立典範。

面對異軍突起，美國政府與科技巨頭不得不調整其基於閉源模式的科技霸權體制。為了維持競爭優勢，美國正在探索如何在保障商業機密的同時，更加積極地參與制定國際規則與拓展跨國合作，致力平衡開放創新與保障所有權之間的關係。DeepSeek-R1的成功實踐表明，開源能夠釋放創新活力，亦能加速全球科技生態的健康演進，進而對美國的科技主導地位產生實質性挑戰。

香港充當科技冷戰的安全地帶

顯而易見，AI的迅猛發展和大規模模型在各行各業的廣泛應用，催化中美之間的劇烈競爭，呈現科技冷戰的危機。在這一背景下，特區政府應聚焦其國際金融中心、獨立法治與開放市場的制度優勢，著力打造一個中立監管平台連接中美雙方，而成為緩衝調節的試點城市。事實上，為確保AI既能安全發展，又能保持公平競爭，香港金融管理局現正和科技企業、研究機構攜手推出生成式AI沙盒，構建一個涵蓋模型安全、數據隱私、倫理審查與風險預警的專屬平台，各類企業和監管機構可在受控、獨立的環境下對AI演算法、數據處理流程和應用場景進行全面測試，及時識別與化解潛在風險。此一項目可為中美兩國的科技對話提供實證支援。

此外，特區政府積極搭建國際對話平台，主動與歐美、東南亞及中國內地的監管機構建立常規化的合作機制，全力在中美科技冷戰中發揮避風港的作用。通過定期舉辦以AI倫理、數據管治與風險管理為主題的政策研討會、專題論壇及雙邊交流，各方圍繞AI應用規範、模型透明度及跨境數據流動等熱點議題，進一步展開溝通與協作。除了有助於平衡各國利益、推動科技與監管同步向前，這種開放協商模式也為國際統一AI監管標準提供寶貴的香港實踐藍圖，進一步鞏固特區在全球科技管治中的樞紐地位。

開源與閉源博弈下的展望

在全球AI領域風雲變幻、競爭白熱化的時代，中美大模型的爭鋒既展現了AI技術的無限創新潛力，亦暴露了開源與閉源兩種模式各自面臨的風險與機遇。中國依託開源協作、演算法優化和務實創新不斷突破，而美國則通過閉源策略、全能整合及高性能應用牢牢佔據部分關鍵市場。這種互為競爭且互補的格局，正驅動全球科技標準和監管制度迅速演變，挑戰傳統管治模式。
憑藉作為國際金融中心並具備健全法制與開放市場的強項，香港特區政府宜利用跨境監管沙盒、多邊對話和區域合作，構建全鏈條技術監管制度，為應對中美大模型競爭提供全新管治思路。唯有政府、業界與國際監管機構協同創新，在開源活力與閉源保障之間取得平衡，香港才更能用好其豐富經驗和前瞻探索，在環球多極競爭中搶佔先機，為未來數碼生態管治創立務實的範例。

章逸飛博士

港大經管學院經濟學高級講師

（本文同時於二零二五年二月十二日載於《信報》「龍虎山下」專欄）

特區政府近6個財政年度中，有5個出現赤字，累計超過5千億元【圖】。面對此一難題，財政司司長將於本月底發表新一份《財政預算案》，有何妙策可以應付？

2024–2025年度首8個月的赤字為1,432億元（已包括發行債券收入）；羅兵咸永道估算全年赤字為948億元，而財政儲備將下降至6,398億元，相當於約10個月政府開支，屬有記錄以來最低水平，最高峰曾達28個月水平。若公共財政繼續入不敷支，並無措施出台扭轉趨勢，估算未來4年赤字亦將居高不下。政府必須盡快全面檢討支出，同時在管理財政及其對公眾的影響之間小心取得平衡。

圖  香港特區政府的財政狀況

（億港元）

資料來源：香港特區政府庫務署

以公債挹注基建

政府近年推出銀色和綠色債券，並積極增幅以應對財赤。去年2月陳茂波發表預算案時，提出2024年度將發債1200億元，其中零售部分為700億元（500億元為銀色債券、200億元為綠色債券及基礎建設債券）。當時訊息非常清晰，500億元銀債與基建債計劃無關。

發售銀債旨在為本港長者提供回報穩定的投資產品，雖沒有二手市場，但投資者可以在債券到期前讓政府提前贖回。值得留意的是，2024年9月根據基礎建設債券框架發行的第九批銀色債券，首度將債券資金用於基建工程。政府認為可更好地管理大型基建所需現金流，令惠及經濟民生的項目早日落成。預計2028–29年度，債務對本地生產總值比率介乎9%至13%。

過去發行的銀色債券均由金融管理局（金管局）負責投資，如去年投資回報有5.5%，稍高於銀債息率，政府不能償還債券的機會極微，但本財政年度開始，卻直接將銀債的資金用於政府公務工程，投資回報變得複雜，風險比以前高。縱使目前政府違約的機會不大，然而隨着發行債券的金額倍數擴大，風險也會大增。無獨有偶，目前銀色債券的發行額度正與財赤增加而同步擴大。例如2020–21年度財政赤字2,325億元，銀債就從此前每年發行30億元躍升至240億元，以致表面上錄得盈餘294億元。

回溯政府在2016年首次發行銀債，目標之一是協助長者抵抗通脹，但近年本港通脹偏低，如2023年的通脹率只是1.7%。除了發行首年，銀債息率一直高於通脹率，尤其是2021年的通脹率僅為0.6%，銀債息率卻達3.5%；認購量又由最初5手大增至去年21手，難免有以公帑補貼長者投資回報之嫌。筆者認為，公帑理應集中支援最需要照顧的弱勢長者。銀債推出已近10年，究竟如何能達到當年推廣長者投資市場的本意，值得深思。

千億債券救近火

綠色債券是政府可持續債券計劃的重要組成部份，為應對氣候變化、轉型至低碳經濟體等理念的綠色項目融資，反映了香港與綠色債券市場的國際標準接軌。本年度預算案列出，綠色債券計劃與基礎建設債券計劃的合共借款上限為5千億元，全部撥入基本工程儲備基金。根據預算案的中期估算，政府2025–26年度可轉虧為盈，盈餘63.3億元，但若扣除發債收支，則至2027–28年度才會錄得盈餘約141億元。

銀債在2020年起發行規模由30億元增加至150億元，翌年更倍增至300億元；同一時間在特區政府帳目中來自債券發行的淨收益，由2019–20年度首次有63億元年收入，隨後兩年倍增至193億元及291億元，可見認購銀債金額全數撥入財政儲備。上年度銀債發行額達550億元新高，並同時發行綠色債券，以致債券淨收入高達716億元。顯而易見，沒有這筆龐大收入，就會錄得1,700億元財赤，而非只是1千億元。政府不應將各類債券所募集的資金視為收入，更不應在同年將債款花掉，公眾才得以了解財赤的嚴峻程度。

新一批銀債集資不再放在外滙基金投資，意味着政府潛在收入減少，財赤亦會比預期多，加上賣地收入減少，財政儲備恐怕更快耗盡，政府有必要及早設法開源節流。金管局稱，未來會繼續按基礎建設債券框架發行銀債，或就合適基建項目發行綠債，視乎屆時的公共項目而定。在財政儲備將要跌穿6千億元的情況下，發行基建債券融資不能單靠儲備作支持，發債成本應與基建項目回報掛勾，但基建投資屬長遠性質，北部都會區、交椅洲人工島發展談不上何時能提供收益，恐怕融資成本將遠高於發行銀債的息率。政府應詳加解說，以釋公眾疑慮。

經濟需另譜新章

結構性赤字其實有好有壞，例如美國列根時代減稅所引發的財政不平衡，在不同程度上有助於經濟結構提升和改革。相反，日本的結構性財赤更關乎經濟政策出現問題，特區政府應引以為鑑，設法推出減低支出、提升收入的政策，否則單靠發債度日，難免有債務危機。政府亦需考慮，持續性發債會否影響金融市場的穩健性及評級。香港一如新加坡和挪威，是世界上極少數擁有積累財政儲備的經濟體，然而透過開源節流來削減財赤，實在談何容易！經濟不景時開徵新稅，必然備受社會反對，甚至因被視為有違簡單稅制的優良傳統，結果得不償失。至於削減公共開支，則會遇到既得利益者的阻力。教育、醫療、社會福利開支，極其量只有望凍結增長。

另一燙手山芋是公務員薪酬。筆者不建議公務員全面減薪，因每每引致私人機構效法，亦會進一步受到打擊消費者和投資者信心，而影響疫後經濟復甦。當然，現時19萬名公務員的龐大架構中，可刪減部分職位以便節流。政府不妨帶頭推動人工智能，提升工作效率，此舉更能釋出正面訊息。發債以外，政府亦可考慮下調股票印花稅的稅率，吸引更多資金流入股市，並制定吸引高端消費遊客的政策，以收開源之效。

財政赤字令特區政府忽略或擱置長遠的策略，無疑是非常危險的傾向。政府的本能反應是削減開支、增加收入。歸根究柢，解決方案還需依賴經濟增長。本港經濟的核心問題，在於地產市道崩潰之後，缺乏新的火車頭和增長動力。這既屬地產泡沫的後遺症，也是外在環境（特別是中國因素）出現重大轉變使然。根據金管局的研究報告，香港的潛在本地生產總值增長率從1980年代初的8%，跌至1998至2000年的3.5% 至4%，反觀南韓、中國台灣及新加坡的增幅卻達4%至6%。

香港經濟因疫情低迷，疫後經濟復甦緩慢，並失去發展方向。國內城市冒起，更迅速動搖香港的自信。特區的生產性投資、效率及人力資源質素，整體難以令人樂觀，加上供求錯配，這座城市的潛在增長率正逐漸下降。這必然影響政府財政收入的穩健性（revenue buoyancy）。由於開支方面出現名義剛性（nominal rigidity），財政赤字便會惡化，單以削減開支、增加收入的傳統方法來平衡財赤，就難以奏效。除非能顯著地提高生產效率，否則財政赤字只會揮之不去，更可能引致惡性循環。如何令經濟發展有所創新，重拾增長動力，才是關鍵所在。

參考資料

Jiming Ha and Cynthia Leung. “Estimating Hong Kong’s Output Gap and Its Impact on Inflation.” HKMA Research Memorandum, November 2001.

謝國生博士   港大經管學院金融學首席講師、新界鄉議局當然執行委員

何敏淙  香港大學附屬學院講師

院線大戰客廳沙發：香港戲院如何扳回一局

從新冠肺炎疫情至今，香港電影院經歷了一波關門大潮。截至2024年4月30日，累計16間戲院結業，其中包括UA連鎖影院在2021年3月全線結業。【註1】這波關門大潮似乎還遠未結束，上月12日MCL Cinema Plus+荷里活廣場戲院宣布結業，而以播放小眾電影知名的高先電影院，也傳出將於租約期滿後結束營業。【註2】

電影院紛紛倒閉並非香港的獨有現象，在作為電影大國的美國，同樣難逃厄運。根據英國一家研究機構統計，目前美國的影院銀幕總數為36,400，比2019年減少12%。世界最大的影院公司AMC娛樂，從2019年開始就陸續關閉了169家影院，雖然疫情之後又增加了60家，但是規模遠比疫情前要小。至於全球第二大連鎖影院巨頭Cineworld，也關閉了接近15%的影院。

影院頭號勁敵

衝着戲院而來的最大對手，無疑是Netflix、Disney+等串流媒體。比起動輒花上百元購買1張門票，同樣是大概100元的串流服務月費，卻可換來豐富的電影、電視節目選擇，對於消費者來說自然更加划算。加上用戶可以舒舒服服地躺在自家沙發上，隨時隨意暫停播放或重播，不難理解為什麼愈來愈多人選擇串流服務來看電影。

無庸置疑，新冠肺炎來襲令影院市道雪上加霜。疫情期間本港戲院共停業267日，市民被迫呆在家中，通過串流媒體觀看電影電視。3年多下來，消費習慣逐漸改變。據《華爾街日報》報導，疫情前有60%的美國消費者認為，看電影應去戲院，疫後降至只有35%；同期選擇在家看電影的美國人，則由40%增至65%。【註3】

作為影院的合作方，電影製作公司亦轉投串流媒體的懷抱。荷里活各大電影製作公司為了既在院線大銀幕上賺足戲票，又想在串流媒體上分一杯羹，通常會在新電影發行的首90天，授予電影院獨家放映權。【註4】消費者如果迫不及待先睹為快，就只能花錢進戲院。等到這個影院獨家檔期過去了，串流媒體才可加以播放。

在串流媒體盛行的今天，獨家檔期對於影院的重要性也就不言而喻。然而疫情令所有影院停運，電影製作公司只好將影片直接在串流媒體上發行。儘管獨家檔期在疫後恢復，但大幅減半至只有45天。盈利窗口大大縮短之際，經營成本又告上升，經營影院生意可謂舉步維艱。

力挽狂瀾之道

影院能否贏過客廳沙發？雖然串流媒體優勢明顯，影院卻可提供無法取代的觀影體驗。如何放大這一強項，是影院能否打贏家中沙發的核心關鍵。比方說，影院的超大銀幕和震撼音效能讓觀影體驗全面升級。在影院總體經營困難的情況下，大銀幕IMAX加杜比音效卻逆勢增長。2023年IMAX的年收益上漲25%，達到2億1400萬美元，其全球（除中國內地）票房收入的市場份額也從2019年的1.7%升至2.1%。【註5】為了強化這一優勢，美國德克薩斯州具百年歷史的連鎖電影院B&B，甚至安裝了一塊寬7層樓、高4層樓的超大銀幕以及一塊270度的環形銀幕，讓觀眾完全沉浸在電影特效之中。

即便是座椅，也有不少影院想勝過家裡的沙發。為了提高觀影體驗，一些影院不惜重金設置可調整溫度和傾斜度的座椅，甚至有服務員端上餐飲，觀眾們一享眼福之餘，還能一飽口福。阿聯酋的連鎖影院Vox更提供和牛豪華大餐，讓看電影變成奢華享受。另有創新者，如在亞洲擁有連鎖影院的韓國公司CGV，不僅在電影放映中為觀眾帶來風雨以及氣味的4D體驗，還在放映前提供音樂、遊戲、餐飲一條龍服務，把觀影塑造成獨特文化體驗。

影院與客廳沙發能否握手言和

電影分類眾多，一個很簡單的標準就是片長。《經濟學人》分析了從荷里活的黃金年代1930年至今10萬多部在全球放映的電影，發現平均片長的增幅為32%，從1930年代平均1小時21分鐘增至2022年的1小時47分鐘。這種趨勢在熱賣的電影中尤其顯著；2022年IMDB網站的評分排名中，十大電影平均片長為2個半小時，足足比1930年代的十大電影長了50%。【註6】

對於觀眾來說，這樣的長片很難集中精力盯着家中的小熒幕一口氣看完，而斷斷續續地看又讓觀影體驗大打折扣。反之，影院的大銀幕配合好音效，若再提供附加服務，就足以吸引消費者步出家門，走進影院，好好享受電影時光。電影製作公司的這種加時策略，在2019年漫威電影《復仇者聯盟：終局之戰》中取得豐厚回報。這部長達3小時的超級英雄電影成為2019年票房冠軍。近期北美電影票房排行榜就有不少這類長片，其中包括長達3小時的2024年奧斯卡最佳影片《奧本海默》。

這意味著電影院和串流媒體可以針對不同影片細分市場：影院專門放映長片或大製作，串流媒體則主攻其他類型的作品；各有側重，和氣收場。

影院應為跳出舒適圈另謀對策

上電影院和在家看戲還有一個很大的不同，就是在影院內和朋友或陌生人坐在一起，這就使得影院有可能成為社交娛樂的場所。影院不妨跳出電影的局限，另闢蹊徑。比如AMC影院正在考慮直播美國欖球聯賽；美國連鎖電影院B&B不僅貼心地為小朋友提供遊樂場地，還為大人提供保齡球、攀岩、匹克球、酒吧以及舉辦私人宴會的場所。其他影院為了鼓勵單身男女入場，甚至在放映前的等待時間組織配對約會活動，讓害怕一個人看電影的年輕人多了一個走進影院的理由。

電影院PK客廳沙發的大戰尚未結束。戲院如果不想舉手投降，只能變陣迎戰，由「觀影場所」打造成「升級體驗的休閒社交場所」，方能有望在這場沒有硝煙的戰役中收復失地。

註1：https://www.hk01.com/%E7%A4%BE%E6%9C%83%E6%96%B0%E8%81%9E/1015167/%E6%88%B2%E9%99%A2%E7%B5%90%E6%A5%AD-%E7%96%AB%E6%83%85%E8%87%B3%E4%BB%8A16%E9%96%93%E5%80%92%E9%96%89-ua%E5%85%A8%E7%B7%9A%E8%90%BD%E5%B9%95-%E5%8C%97%E4%B8%8A%E6%BD%AE%E6%AF%94%E9%98%B2%E7%96%AB%E6%8E%AA%E6%96%BD%E6%9B%B4%E5%85%87

註2：https://www.hk01.com/%E9%9B%BB%E5%BD%B1/1084551/%E9%AB%98%E5%85%88%E9%9B%BB%E5%BD%B1%E9%99%A2%E7%AA%81%E5%82%B3%E6%BB%BF%E7%B4%84%E5%BE%8C%E7%B5%90%E6%A5%AD-%E6%9C%89%E8%82%A1%E6%9D%B1%E6%9B%BE%E8%A1%A8%E7%A4%BA-%E7%94%9F%E6%84%8F%E5%94%94%E4%BF%82%E4%B8%8D%E5%A5%BD-%E8%80%8C%E4%BF%82%E5%B7%AE

註3：“How going to the movies is changing, in charts,” The Wall Street Journal, August 6, 2024.

註4：“Movie-theater industry pain intensifies even as pandemic eases,” The Wall Street Journal, April 4, 2023.

註5：“Cinemas may be dying. But IMAX and the high end are thriving,” The Economist, February 28, 2024.

註6：“Why films have become so ridiculously long,” The Economist, October 14, 2023.

范亭亭
港大經管學院市場學首席講師

（本文同時於二零二五年一月二十二日載於《信報》「龍虎山下」專欄）