How important are large firms for aggregate productivity growth in the UK?

Marko Melolinna

Aggregate labour productivity growth has been low in the UK following the global financial crisis in 2008 (Chart 1). The average annual growth rate has been only 0.7% over the period 2008 to 2019, which is around a third of the growth rate seen during the decade preceding the crisis. There are many ways of analysing the reasons for this weakness, but in this blog post, I concentrate on examining the role that the largest firms in the UK have played in the story. Our analysis covering the past three decades from 1990 to 2017 suggests that firm-specific, or idiosyncratic, shocks to the 100 largest firms had a significant effect on aggregate productivity dynamics in the UK.

Chart 1: UK productivity and pre-crisis trend

Notes: Productivity is defined as UK GDP divided by number of employees in the UK economy. 1997Q1=100.

Source: ONS and author’s calculations.

Various explanations of the weak productivity dynamics in the UK in the aftermath of the financial crisis have been proposed in the literature. A wide range of approaches have been used, increasingly relying not just on aggregate, but also on firm- and industry-level data. A crucial question in the analysis has been the extent to which the productivity slowdown has been of a short-term cyclical or more persistently structural nature. Evidence suggests that during the initial phases of the recession, firms in the UK acted flexibly by holding on to labour and lowering factor utilisation in response to weak demand conditions, with negative cyclical effects on productivity. But since then, more structural factors, like reduced investment in both physical and intangible capital and impaired resource allocation from low to high productive uses, are likely to have played a more persistent role in the weakness (see eg Barnett et al (2014) and Bank of England Quarterly Bulletin (2014Q2)). Purely empirical characteristics of the productivity data also suggest an important role for structural factors (see Melolinna and Tóth (2019)). More recently, uncertainties related to the EU referendum have had a negative effect on productivity growth (see Bloom et al (2019)).

In terms of industry-level data, Tenreyro (2018) offered evidence pointing to the importance of very few sectors in driving the weakness in productivity growth, with finance and manufacturing being the largest contributors to the slowdown. In a recent example of analysis with the universe of UK firm-level data, Schneider (2018) argued that the slowdown has been driven by shortfalls in the top quartile of the firms’ productivity distribution.

In a recent Working Paper, my co-author and I looked at UK productivity dynamics from a slightly different angle compared to previous studies. Our interest was not on the drivers of the weakness, nor on the industry mix or universe of all firms. Instead, our focus was on establishing whether idiosyncratic shocks to the largest UK firms, rather than common shocks, have had a significant effect on the dynamics of aggregate productivity in the UK over the past three decades. To do this, we introduced a unique way of separating firm-specific idiosyncratic shocks from common shocks.

Traditionally, macroeconomists are not very interested in the economic activity of a small number of firms (or households), since their effect on the aggregate should be negligible. However, an important paper by Gabaix (2011) showed that under certain conditions, a small number of firms can have significant aggregate effects. In particular, he showed that if the distribution of firm size is sufficiently skewed and certain other statistical properties of the firm population are met, then a small number of large firms may account for a disproportionally large fraction of aggregate fluctuations. To measure this, the author introduced the concept of a ‘granular residual’ (GR), which is a composite of idiosyncratic, firm-level shocks to a selected number of the largest firms in the economy. Gabaix (2011)  finds that the GR is a significant driver of aggregate GDP and productivity dynamics in the US.

In our analysis, we first showed that the population of firms in the UK is indeed sufficiently skewed for large firms to potentially have significant aggregate effects. We then took Gabaix (2011) (and several papers that have built on it) as a starting point for calculating a firm-level GR. As Gabaix (2011) shows, in its most general form, the GR can be computed with a simple formula that adds up the idiosyncratic firm-level technology shocks for the 100 largest UK firms each year. These shocks are weighted by the size of each firm relative to GDP, so the larger the firm, the larger the contribution of its technology shock to the GR measure.

We also make an effort to estimate the idiosyncratic technology shocks somewhat differently from previous literature. In our approach, we control for the unobserved, firm-specific differences in technology that do not vary over time, and — unlike Gabaix (2011) — we also allow for the possible existence of common shocks (which can be correlated over time) affecting many firms. We do this by estimating a firm-level production function, where firm production depends on its fixed (capital) and variable (labour) inputs, as well as the efficiency with which these inputs are combined to produce output. We use an established procedure introduced by De Loecker and Warzynksi (2012) to estimate firm-level weights on the fixed and variable inputs, which allows us to back out the firm-specific technology shocks. We then go on to use these firm-specific shocks in simple regressions explaining aggregate UK productivity dynamics over time.

We took our firm-level production function model to financial accounts data on the 100 largest UK-based firms each year (excluding finance and oil sectors). The resulting GR contribution to productivity growth is depicted in Chart 2, together with aggregate UK productivity growth. The GR contribution in the chart shows how much the sum of the idiosyncratic firm-level technology shocks of the 100 largest UK firms contributes to aggregate productivity growth each year, as described above. So for example in 2005, the GR contributed around 0.4pp to aggregate productivity growth of around 2%.

By properly accounting for firm-level technology shocks in this way and regressing aggregate productivity growth on the GR contribution, we find that the shocks to the largest 100 UK firms can explain around 30% of aggregate productivity dynamics in the UK, on average, over the past three decades. Our analysis also suggests that it is important to allow for common shocks as well as firm-specific idiosyncratic shocks in the model. In the paper, we show that simplifications of our approach, which omit controlling for firm-level idiosyncratic shocks or do not account for common shocks, do not work well, highlighting the importance of identifying firm-specific shocks correctly.

Chart 2: UK productivity growth and ‘Granular residual’ of large firms

Notes: Productivity is defined as UK GDP divided by number of employees in the UK economy. Granular residual refers to the contribution of the largest 100 firms to aggregate productivity growth, as defined in the text.

Source: ONS and author’s calculations.

The main takeaway, based on our empirical application to UK productivity data, is that technology shocks to large firms can be important drivers of aggregate productivity dynamics. So while commentators have not generally looked at idiosyncratic shocks when explaining the UK productivity puzzle, the evidence in my post suggests that they should. However, as Chart 2 also shows, the correlation between the GR and aggregate productivity growth is far from perfect, especially since the financial crisis. So there are plenty of other factors taking place outside the largest firms that can account for the majority of productivity dynamics. The on-going challenge for policymakers is to figure out what factors, common and idiosyncratic, drive these dynamics, both at the aggregate as well as the firm level.

Marko Melolinna works in the Structural Economics Division.

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.