From gas to AI, the ‘Dutch disease’ that comes from California

Whilst the US stock market prepares for Anthropic’s mega-IPO, AI continues to provoke a mixture of enthusiasm and fear. The latter has grown in comparison to the former. Progress does not stand still, of course, and AI is outpacing all existing information technologies. But attention is increasingly focusing on the potential negative implications and possible remedies.

One aspect that has received little attention is the impact on the rest of the economy. A boom in one sector is not confined to that sector alone; when it exceeds a certain scale, it triggers mechanisms that reverberate throughout the entire economy and call into question macroeconomic policies: fiscal, monetary and trade policies. Beyond a certain scale, the success of one sector can throw others into crisis, undermining a structural balance that depends, in part, on the diversification of the economy. Industrial policies, which are normally distinct from macroeconomic policies in terms of their aims and instruments, end up intersecting with them.

For those of a certain age, all this is reminiscent of ‘Dutch disease’, a dominant theme in the 1960s and 1970s. In 1959, natural gas fields were discovered in the plains around Groningen, in the north-east of the Netherlands. This led to a boom in state revenue, as well as benefits for individuals and businesses arising from an abundant and cheap energy supply. An unexpected windfall? Not quite. The Dutch guilder appreciated, undermining the competitiveness of the manufacturing sector; the Netherlands is an economy highly open to trade. The balance of payments remained healthy thanks to the energy sector, but domestically the economy was in a state of disarray: high costs, unemployment, the high cost of living, industrial decline and the deterioration of physical capital and know-how. When, a few years later, the United Kingdom discovered oil in the North Sea, the problem spread to the island across the Channel. The Economist magazine identified the problem and coined the term.

Before drawing parallels with certain current phenomena, it is worth recalling the remedies that economists proposed and that some governments adopted at the time. The first: to isolate the energy sector from the rest of the economy, setting aside the revenue to reduce its inflationary impact. This is an appropriate remedy if the shock is deemed to be temporary, and it is therefore advisable to build up reserves. This is the approach taken by some energy-producing countries, such as Norway and Saudi Arabia. The second strategy consists of measures to protect non-energy sectors, through subsidies and other means. This approach was adopted by the Netherlands itself, as well as by the pre-Thatcher United Kingdom.

In the regions where AI is primarily being developed – the United States, and California in particular – phenomena are being observed that are in some ways similar to those triggered by the Dutch gas crisis, albeit in different ways. The variable that transmits the effects to the rest of the system is not the exchange rate: the dollar has depreciated over the last two years. Rather, it is interest rates – particularly long-term rates – which have soared due to the mountain of debt generated by the high-tech sector. Except within this sector, high interest rates discourage investment and growth in other sectors. Data from Oxford Economics illustrate the divergence in US investment trends: robust in the IT sector, to fund data centres and other AI-related facilities, whilst stagnant or declining elsewhere. The impact of the boom on the wider economy is limited by the fact that labour demand in the IT sector is minimal. Returns on investment are astronomical but highly concentrated amongst a small number of people; consequently, they do not generate aggregate demand. What remains – and is significant – are the effects in terms of a slowdown in investment, growth and innovation in other sectors. The economy is becoming polarised and losing its diversity. The process is still in its early stages; where will it lead? The direction is relatively clear, but the scale of the phenomenon and its ultimate outcome are far less so.

Traditional remedies do not seem suited to what is happening today. The Dutch gas revenues that used to flow into the state could be temporarily set aside in a piggy bank. Today, they enrich a small group of individuals in the form of capital gains on shares that are not (yet) listed. The sovereign wealth fund route is off the table. It remains possible – and probably advisable – to protect manufacturing sectors, which are exposed to competition and hampered by financial conditions. Might we therefore conclude that Trump’s tariffs and the devaluation of the dollar – encouraged by this US administration in various ways – have some logic to them? Unfortunately, this ex-post rationalisation does not hold water: in terms of sectoral and geographical distribution, as well as their scale, the tariffs imposed on so-called ‘Liberation Day’ have little to do with a rational strategy of containment and rebalancing. By contrast, the rise in borrowing costs, which is strangling much of the economy, has a great deal to do with this administration’s unpredictable policies.