This post was originally published on this site.
Listen to this article
If there was any doubt that weâre in an artificial intelligence (AI) race, Donald Trumpâs 21 January 2025 announcement of a US$500 billion private investment in building data centres and advancing AI chips removes it. Even the UK government is joining in, unveiling an AI plan on 13 January 2025 with 50 recommendations to boost innovation and efficiency.
All this was before Chinaâs DeepSeek rocked the foundations of the AI world, delivering the same results as Western companies at dramatically lower cost.
With AI coming in fast, a common question is how we can minimise the risks from the technology. But an equally important question which doesnât get asked as much is: how can we maximise its benefits? After all, for any given level of risk, we will want to make sure we are getting the biggest bang for our buck.
This might sound like a tech question. But itâs largely a question for macroeconomics. Olivier Blanchard said recently that DeepSeek was âprobably the largest positive total factor productivity shock in the history of the worldâ.
So we used a computable general equilibrium model to simulate exactly that: an AI boom where productivity takes off across countries as firms and households adopt new AI technologies to explore how the benefits to GDP from this boom differed depending on the governmentâs policy settings.
We identify four key lessons for policymakers in Asia and around the world if they want to get the most out of the AI boom.
First, the AI boom will see a major increase in investment, along with a significant increase in demand for savings to finance that investment. This could lead to a problem â if the demand for savings outstrips the supply of savings, interest rates rise, dampening the investment boom and offsetting some of the benefits to the community.
Unsurprisingly, we found that the countries that have access to more savings will do better than the countries that donât. For policymakers, this means that governments should continue to reduce debt and deficits and find more efficient ways to deliver services, remove restrictions on foreign investment and get comfortable running trade deficits since exports will fall as the exchange rate appreciates with capital inflows.
Second, countries will need to compete for the worldâs scarce savings. Countries with a higher return on capital unsurprisingly do better at attracting savings to finance the AI boom and hence reap more of its benefits.
For policymakers, this means two things â lifting productivity and tax reform. Simplifying and reducing taxes on capital and investment will increase the return on capital, encouraging more savings to flow into the economy and delivering bigger benefits from the AI boom. It will also lead to better energy markets, improved resource management and the removal of duplication in development approvals.
Third, financial markets will need to direct the capital to where itâs needed. Key to this is flexibility. Countries with more financial rigidities experience the same negative consequences as countries with insufficient savings â the investment boom is muted.
For policymakers, this means making sure that financial regulations are not unnecessarily preventing firms and households from getting the capital they need. Bank asset risk weightings, along with regulations on pension funds and fintech, should all be reviewed with this in mind. It also means being strategic and targeted about where countries can most succeed in the AI tech stack â in Australia, for example, these opportunities lie in applications, infrastructure and datasets.
Fourth, itâs not all about capital. Booming industries will need workers and many new AI businesses will be created by workers who leave their current jobs. Policymakers need to make sure that workers can switch jobs easily, that they are equipped with the right skills and that the entrepreneurs who start new businesses arenât stopped from doing so.
This means getting rid of the things that make it hard for workers to switch jobs. Non-compete clauses should be scrapped, taxes that punish relocation should be phased out and barriers to entry to certain occupations (think: licencing requirements) should be reviewed. Regulatory and tax inconsistencies between sub-national governments should also be addressed.
Upskilling the workforce will be key. The AI boom creates new jobs and higher wages, particularly in STEM fields, which will encourage more people to train in those areas. But the lag-time in this upskilling means that it is crucial to get ahead of the curve and use direct subsidies to start upskilling quickly in STEM capabilities. And developed economies need to increase skilled migration, not reduce it. If businesses canât expand due to a lack of skilled workers, the economy wonât get the full benefits of the AI boom.
The AI race is an ironman event, not a track meet, and success depends on getting all the settings right. Policymakers must approach it comprehensively. Getting our macro settings right should be the first step.
Amit Singh is Managing Partner and Adam Triggs is Partner at Mandala, an Australia-based economic consultancy.