Blog

​​The UK’s productivity problem is well known – but it is not evenly shared. To make sense of the UK’s productivity landscape, we need to understand what is happening at the local level. Subregional productivity differences result from a mix of factors, including variations in industrial and occupational structures as well as genuine differences in ‘production efficiency’. A typical starting point in such analyses is usually the official productivity statistics, which estimate labour productivity as gross value added (GVA) per hour worked or per job. Having good data is therefore crucial – yet the inclusion of a concept known as ‘imputed rent’ within standard productivity metrics can distort our understanding of how local economies actually perform.

Imputed rent and productivity

One component of output metrics such as GDP or GVA is the so-called imputed rent. The ONS defines imputed rent as “an estimate of the housing services consumed by households who are not actually renting their residence”, noting it can be understood as “the amount that non-renters pay themselves for the housing services that they produce.” In other words, imputed rent is a concept that treats owner-occupiers (people who own a property and live in it as their main home) as if they were both landlords and tenants of their own homes, assigning a monetary value to the housing services they consume even though no actual transaction takes place (which is why it needs to be ‘imputed’). It is included as part of UK output alongside industrial and service-sector production, which the ONS explains is necessary to ensure GDP statistics remain consistent and comparable over time and across countries.  

However, the inclusion of imputed rent in output metrics has drawn some criticism. For example, Policy Research in Macroeconomics (PRIME) argues that imputed rent is a “purely fictional” component that makes up an inordinately large share of GDP and GVA, artificially supporting measured output even when real economic activity is weak. Indeed, ONS data indicates that imputed rent accounted for between 9% and 11% of UK GVA over 1998-2023, though for many local areas the share is much higher – for example, imputed rent made up nearly 32% of East Renfrewshire’s GVA in 2023. Similar critiques have been raised by others.

What does this mean for labour productivity data and analyses of productivity at a local level? Because imputed rent involves no labour input, its inclusion boosts GVA while leaving hours worked or job numbers unchanged. As a result, measured labour productivity rises even though no additional labour‑based output has been generated. This is true for the UK as a whole as well as all local areas. However, the scale of this effect is not uniform: areas where imputed rent makes up a relatively large share of GVA will see a greater‑than‑average boost to productivity (and vice versa).

In broad terms, each area’s share of GVA attributable to imputed rent is shaped by three main factors: local rental levels, the number of owner-occupied homes, and the scale of other economic activity included in GVA (such as industrial production). As a result, local areas with high homeownership, high rents and relatively limited economic activity see the largest productivity uplift from including imputed rent in GVA figures.

A report from the Centre for Regional Economic and Social Research at Sheffield Hallam University notes that imputed rent “distorts comparisons of productivity between local areas within the UK because of variations in housing tenure and costs.” At SQW, we have also encountered this issue in our own project work, including on a study of Buckinghamshire’s productivity performance, the findings of which can be accessed here.

Effect of imputed rent on local productivity comparisons

To test how imputed rent affects each area’s productivity position relative to the UK average, we have produced an alternative productivity measure for all lower-level local authorities that excludes imputed rent.

Specifically, we calculated GVA per hour worked and GVA per job for every local area and expressed each result relative to the UK average, where the national level equals 100 (this is what is referred to as relative productivity). We then repeated the same calculations using a version of GVA with imputed rent removed. By comparing the original (unadjusted) and adjusted figures, we derived a simple difference measure that shows how much the inclusion of imputed rent shifts each area’s relative productivity.         

The results are shown in the maps below. The analysis highlights which places appear more or less productive simply because of the way housing services are treated in the national accounts. Areas shaded blue are places that benefit from the inclusion of imputed rent in GVA – meaning their relative productivity looks higher when imputed rent is counted. Areas shaded red appear less productive once imputed rent is included, i.e. removing imputed rent from GVA would raise their relative position. The darker the shading, the larger the effect. Areas shown in light grey are those where the difference between the two measures is very small, meaning imputed rent has little impact on how productive they appear relative to the UK average.

The results reveal some interesting patterns. Broadly speaking, areas in the south of England – particularly along the south coast – show some of the largest gains from the inclusion of imputed rent in GVA. Conversely, northern England and parts of the Midlands have a higher prevalence of areas experiencing negative effects.

A second pattern emerges when looking through a different lens: urban areas feature prominently among those seeing the largest negative effects, likely reflecting their lower levels of homeownership and relatively high volumes of other types of economic activity. For example, the City of London is the single most adversely affected area, though similar patterns can be observed across every UK region and nation.

The tables below highlight the local authority areas most affected by the inclusion of imputed rent in official productivity statistics, showing the top ten areas with the largest declines and the largest uplifts in relative productivity. In both cases, the figures compare relative productivity with and without imputed rent, illustrating how the inclusion of imputed rent can materially alter the apparent productivity performance of different places.

One finding is that several areas that already appear highly productive in the official statistics – including London boroughs such as the City of London, Tower Hamlets or Westminster – would display an even greater productivity advantage if imputed rent was excluded. Conversely, many of the areas with below-average productivity in the official statistics – such as Castle Point, Wandsworth or East Dunbartonshire – would see their gap to the UK average widen even further. More broadly, once imputed rent is removed, the spread of local productivity performance (measured using the standard deviation) becomes wider, meaning the differences between places are larger than in the official statistics.

Implications for productivity policy and analysis

The findings underscore an important point for productivity policy and analysis: measurement choices matter. If imputed rent was stripped out from sub-national analyses, we would see a wider spread of local productivity performance and a reshuffling of area rankings.  

The ONS includes imputed rent in measured output to ensure consistency with international standards and comparability across countries. At the local level, however, the case for its inclusion is less clear‑cut. As a sizeable but labour‑free component of GVA, imputed rent raises questions about what measured productivity is really capturing in subnational comparisons. If nothing else, it reinforces the importance of relying on a basket of indicators in analysing local economic performance, and using these critically and judiciously.  It also, perhaps, suggests that sub-national variations in the UK’s overall productivity challenge are even greater than conventionally assumed (though recognising that these differences reflect industrial composition as well as production efficiency). The implications are important in formulating effective national and local policy responses.