ROI-AI could help governments cut debt — but unevenly: Joachim Klement

AI promises to deliver significant productivity gains to individuals and businesses, but could it also help developed ​market governments reduce their budget deficits and pay down debt? It might. The estimated productivity gains from AI are highly uncertain. Chicago Federal Reserve Board ​economist Ezra Karger and his colleagues recently surveyed economists, superforecasters, AI experts, and the general public on ‌the question. The ​median estimate for the boost to US gross domestic product (GDP) growth from AI was around 0.5 percentage point per year through 2031. Economists, perhaps unsurprisingly, are much more conservative than the other groups, expecting only a 0.1 percentage point boost annually over the next five years.

But one thing everyone can agree on is that if productivity increases, GDP growth will rise. This can happen through higher corporate profits as businesses reduce costs or increase output. Or it could happen ‌through higher real wages if workers become more productive, increasing their bargaining power and ability to demand better pay. To put things simply, both businesses and workers pay taxes, which means that, all else being equal, increased productivity should indirectly lead to higher tax revenues for the government. This, in turn, should reduce the budget deficit and the debt-to-GDP ratio, unless revenue increases are outpaced by increased government spending.

Let’s consider some scenarios. The chart below shows my estimates for how the US debt-to-GDP ratio could change relative to the Congressional Budget Office's (CBO) baseline projections. I assume either a 0.1 percentage point or a 0.5 percentage point boost to annual GDP growth, in line with the ‌median estimates of economists and all survey participants noted above. Furthermore, I simulate two additional scenarios. In the first one, all the benefits of the productivity boost accrue to businesses in the form of higher profits. In the other one, half the benefits are captured by workers in the form of higher ‌real wages.

Because US income tax rates tend to be higher than corporate tax rates, government revenues usually rise more when workers capture a greater share of the productivity gains. Consequently, in the second scenario, the deficit would shrink more and the debt-to-GDP ratio would fall further relative to the CBO baseline. The CBO currently projects that the US debt-to-GDP ratio will reach 120% in 2036. With a productivity boost of 0.5 percentage points per year – again, with all else being equal – this ratio could fall to under 111% if all the benefits accrue to businesses and to below 110% if half of the productivity gains are captured by workers.

The US primary deficit — which excludes debt-servicing costs — in 2031 could thus be reduced from 1.9% to as low as 1.3%, potentially shrinking the full budget deficit ⁠from 5.9% to ​5.2% of GDP, depending on how the gains are allocated. That’s all theory, however. In ⁠practice, the situation is a lot more complicated.

One key concern is AI’s potential effects on inequality and unemployment across the population, which could result in a massive surge in welfare payments. Indeed, the cost efficiencies driving the productivity boost could come from businesses producing more with fewer workers. But the Budget Lab at Yale has produced a much more elaborate simulation – taking into account many of these concerns – ⁠and it reaches results similar to those of my crude approach.

The bottom line is that AI productivity gains may be able to improve the fiscal outlook for the US, though likely not nearly enough to truly get the country out of its debt spiral. The US is running chronically high deficits of more than 5% and has one of the lowest ​effective corporate tax rates anywhere in the industrialised world. The result is that – even in fairly optimistic scenarios – the benefits of the growth boost could remain largely with the private sector and do little to improve government finances. THE UK ADVANTAGE

However, in other countries with higher tax rates and smaller ⁠deficits, the AI productivity boost could be a real game changer. Let’s look at the UK. Using the same scenarios above, I consider the potential impact of an AI productivity boost on Britain’s public finances, starting with the baseline projections from the UK Office for Budget Responsibility (OBR) through 2031.

Because growth rates in Britain are lower than in the US, a productivity boost of 0.1 to 0.5 percentage points translates into a ⁠larger ​relative increase in profits and wages. Tax rates in the UK are also higher than in the US, which means the government should capture a larger share of these gains. The OBR currently estimates that the UK debt-to-GDP ratio will be 95% in 2031. Assuming an AI productivity boost similar to the one expected for the US, I estimate that this ratio could drop to about 89%. This is a six-percentage-point decline in five years, roughly twice the reduction in the US over the same period.

The picture improves further for the taxpayer when examining the potential impact on projected budget deficits. By 2031, a 0.1 percentage point boost ⁠to UK GDP growth from AI could reduce the budget deficit by a full percentage point. If the GDP boost from AI is instead at the higher end of the range, the budget deficit could shrink by up to 1.3 percentage points compared with current projections. For a country forecast to ⁠run a deficit of less than 3% of GDP in the late 2020s, this means ⁠it could halve its deficit if the benefits from AI come to pass.

One caveat is that the GDP boost from AI could be lower in the UK than in the US given that Britain has a less dynamic economy. And, of course, the problem with any such forecast is – as always – the word “if”. (The views expressed here are those of Joachim Klement, an investment strategist for Panmure Liberum.)

Enjoying this column? Check out Reuters Open Interest (ROI), your essential new source for ‌global financial commentary. Follow ROI on LinkedIn, and X. And ‌listen to the Morning Bid daily podcast on Apple, Spotify, or the Reuters app. Subscribe to hear Reuters journalists discuss the biggest news in markets and finance seven ​days a week.

(Writing by Joachim Klement Editing by Marguerita Choy and Anna Szymanski)