Ludovica Ambrosino, Jenny Chan and Silvana Tenreyro
Recent technological advances raise an important question for policymakers: will higher productivity lead to disinflationary or inflationary pressure? A coming wave of AI-driven productivity growth is often described as a disinflationary tailwind that would allow central banks to hold interest rates lower without reigniting inflationary pressures. Yet faster productivity growth can just as plausibly call for higher, not lower, interest rates. By raising expected future income and the returns to investment, it stimulates consumption and investment today, pushing up the natural rate of interest. Neither view is entirely wrong and our model reconciles the two by showing that the answer depends on the timing, permanence, and sectoral origin of the productivity shock.
The intuition that producing more output from the same inputs should lower prices is a partial equilibrium argument, as it describes how productivity affects supply while holding demand fixed. In general equilibrium, higher expected income and returns also raise consumption and investment. Whether inflation rises or falls therefore depends on the balance between expanding supply and demand, and, crucially, on how monetary policy responds.
Methodology
To illustrate these dynamics, we analyse three scenarios for how a 10% rise in productivity can unfold: a temporary increase, a one-time and permanent increase, and a gradual and permanent increase (Ambrosino et al (2026)). All three scenarios capture the partial equilibrium intuition that higher productivity allows firms to produce more output per unit of input. Each scenario gradually builds up the demand-side effect to show how the overall impact on inflation depends on the interaction between supply, demand and expectations. The sectoral incidence of the shock will also matter: we begin by discussing the implications of higher productivity in the services sector, before considering the same three scenarios in the tradables sector.
We capture these dynamics in a small open-economy New Keynesian model with two sectors: a non-tradable sector, or ‘services’ sector, whose outputs are priced and sold only at home (such as haircuts or restaurant meals), and a tradable sector, whose goods are traded internationally and priced with reference to world prices. Households consume both types of goods, save, and supply labour. Firms in each sector hire workers and capital and adjust prices so that inflation depends on both current costs and expectations of the future. A central bank sets interest rates in response to inflation
When higher productivity temporarily lowers inflation
We start with a textbook example: a temporary increase in the level of productivity (Chart 1, Column 1). Consider a new technology that allows firms to produce more efficiently. Production costs fall, and firms can supply more goods and services. Early evidence suggests that this pattern may be happening in hiring for AI-exposed occupations. This increase in supply places downward pressure on prices. In this scenario, the productivity shock causes a temporary fall in inflation because it leads to a one-off adjustment in the price level rather than a permanent reduction in the inflation rate. Once the shock dissipates, inflation returns to its steady-state level. The natural real rate of interest also falls temporarily, before recovering alongside inflation as the shock fades.
When higher productivity increases demand
The demand-side effect is stronger if the increase in productivity is permanent rather than temporary. If technology permanently raises the economy’s productive capacity, households and firms expect higher income and profits in the future. These expectations can affect behaviour today. Households may increase consumption because they expect future income to be higher, while firms may increase investment because the expected return to capital rises as productivity increases. As a result, aggregate demand begins to increase alongside the expansion in supply. Business investment and household spending both move ahead of realised productivity gains, as many argue is happening now with investment in AI infrastructure. This pattern is also visible in how AI-related earnings news moves equity and bond markets.
These two forces may offset each other. Higher productivity expands supply, while higher expected income raises demand. Whether inflation rises or falls therefore depends on the relative strength of these two effects. If the expansion in productive capacity dominates, inflation falls. But if demand responds strongly, the effect of higher productivity can be less disinflationary or even neutral for inflation (Chart 1, Column 2). The natural real rate of interest is little changed in this scenario, reflecting how closely the expansion in supply and the strengthening in demand offset one another.
When higher productivity is inflationary
Instead of an immediate increase in productivity, consider a scenario where productivity increases gradually over time (Chart 1, Column 3). This pattern is plausible if general purpose technologies diffuse slowly through the economy. For example, firms may need to reorganise production, complementary innovations need to be developed, or bottlenecks in skills or infrastructure may slow adoption. As a result, productivity gains materialise only gradually over time. Historical examples of general-purpose technologies, such as electricity and information technology, show a similar pattern: productivity gains materialised over many years as applications developed and firms reorganised around the new technology. Early UK industry data suggests that this pattern may be starting to repeat.
Again, if households and businesses expect productivity to rise in the future, they anticipate higher future incomes and profits, which changes behaviour today. They may start spending and investing before those gains actually materialise. Firms may invest more to take advantage of higher expected returns, while households can increase consumption because they expect higher future income. This increase in investment and consumption raises overall demand.
This creates a scenario where demand rises first while supply takes time to catch up. If demand grows faster than supply, inflationary pressures can emerge. This increase in demand shows up as a higher natural real rate, which requires monetary policy to tighten in order to dampen inflationary pressures (Chart 1, Column 3). This dynamic is not just theoretical. Similar debates occurred during the technology boom of the late 1990s: strong productivity growth initially coincided with low inflation, prompting then-Federal Reserve Chair Alan Greenspan to argue that productivity gains were holding down inflation (Greenspan (1999)). But in the same speech, Greenspan also warned that rising equity prices and wealth effects were fuelling domestic demand and tightening labour markets faster than productivity gains could offset, and that wages would eventually outpace productivity, leading to inflationary pressures. This is broadly what occurred: demand kept outpacing supply and the Federal Reserve raised interest rates as inflation rose steadily until the 2001 recession (Furman (2026)).
Chart 1: 10% increase in service productivity
Notes: This chart shows the responses of various macroeconomic variables following a 10% productivity shock in the services sector, under three timing assumptions. The exchange rate is defined as the domestic-currency price of foreign currency, so a decline corresponds to an appreciation of the domestic currency (a rise corresponds to a depreciation).
Where productivity gains happen matters
So far, these scenarios have described a productivity gain in the services (non-tradable) sector. However, the inflationary consequences also depend on where productivity gains materialise.
In our model, prices for internationally-traded goods are pinned down largely by world prices, so a productivity gain in the tradable sector does not show up mainly as lower prices for tradable goods. Instead, the adjustment happens through higher wages and income, which raise demand for services, a sector where supply cannot expand as quickly. The resulting rise in services prices can dominate, so aggregate inflation rises even though productivity has improved. This is the classic Balassa-Samuelson mechanism (Balassa (1964) and Samuelson (1964)), applied to the timing of a tradable-sector productivity gain.
The sectoral incidence of a shock can reverse the pattern described above. A front-loaded productivity gain is disinflationary when it happens in services, but inflationary when it happens in tradables, because it leads to demand-driven services inflation rather than a fall in the tradable sector’s own costs (Chart 2, Column 2). A gradual productivity gain yields the opposite pattern: while this had been inflationary when the productivity increased in services, it is now disinflationary when the productivity shock occurs in tradables (Chart 2, Column 3). In this case, the exchange rate appreciates in anticipation of the future productivity gain more quickly than domestic resources can be reallocated, placing downward pressure on imported and tradable goods prices immediately (Broadbent et al (2024)).
Chart 2: 10% increase in tradables productivity
Notes: This chart shows the responses of various macroeconomic variables following a 10% tradable-sector productivity shock, under the same three timing assumptions. The exchange rate is defined as the domestic-currency price of foreign currency, so a decline corresponds to an appreciation of the domestic currency (a rise corresponds to a depreciation).
Policy implications
For central banks, higher productivity is neither inherently inflationary not disinflationary. The inflationary consequences are a priori ambiguous because productivity affects both supply and demand. The overall impact depends on how quickly productive capacity expands relative to demand, whether the shock reflects a temporary level effect or a persistent increase in growth, how expectations affect spending and investment, where productivity gains occur across sectors, and whether monetary policy adjusts in line with changes in the natural rate of interest. The task for policymakers is therefore to assess in real time, whether productivity gains are generating demand pressures and shifting the natural rate of interest, while looking through temporary relative price movements that do not affect medium-term inflation dynamics.
Ludovica Ambrosino is a PhD student at London Business School, Jenny Chan works in the Bank’s External MPC Unit and Silvana Tenreyro is the James E. Meade Professor of Economics at the LSE.
If you want to get in touch, please email us at bankunderground@bankofengland.co.uk or leave a comment below.
Comments will only appear once approved by a moderator, and are only published where a full name is supplied. Bank Underground is a blog for Bank of England staff to share views that challenge – or support – prevailing policy orthodoxies. The views expressed here are those of the authors, and are not necessarily those of the Bank of England, or its policy committees.