Productivity and the allocation of capital following the great recession

Lead Research Organisation: Institute for Fiscal Studies
Department Name: IFS Research Team

Abstract

Labour productivity fell sharply at the start of the 2008 financial crisis. Output produced per worker remains around 14 percentage points below the pre-crisis trend. The scale of the productivity fall and the continued stagnation has been the defining feature of the UK's Great Recession and is commonly termed the 'productivity puzzle'.

Headway has been made in solving the puzzle, much of it by work at the Bank of England and the Institute for Fiscal Studies. We know that the solution does not lie in the changing composition of the work force or in an industrial shift in the economy. At the start of the recession productivity was depressed as employers with sufficient resources retained skilled and experienced workers even though they were less productive at that time (this is known as 'labour hoarding'). Since then, multiple factors have likely contributed to the continued weakness, including: a fall in real wages (that may have facilitated labour hoarding and led firms to substitute labour for capital, but that may also be a result of lower productivity); a fall in investment (such that workers have less, and less good capital to work with); higher firm survival than in previous recessions. However, it has been difficult to account for the entire puzzle by appealing to these factors alone.

In this project we will explore the role of capital allocation. Previous evidence shows that the overall productivity of the economy - Total Factor Productivity (TFP) - is improved substantially when resources flow to the uses that have the highest return. We have found preliminary evidence that the link between investment and rates of return in the UK has broken down since 2008. The recession was characterised by relative demand shocks and an increasing dispersion in rates of return to capital across firms. Despite this, there has been very little adjustment in the allocation of capital; capital is not flowing to the projects with the highest returns. This is perhaps not surprising given that the key channels through which capital flows across the economy (new investment and the entry and exit of firms) have been notably weak since 2008. The UK experience stands in contrast to previous UK recessions, and to the current experience of the US.

An increase in uncertainty and financial market impairment are important candidates for why capital has not been adjusted in response to shocks. In this project we view such factors as distorting firms' choices over how much capital to invest in. Our objective is to quantify the scale of such distortions and to estimate how important they are in impeding the allocation of capital and thereby lowering aggregate TFP (and therefore labour productivity).
This project will be novel in applying previously developed methods from the academic literature to the context of an advanced economy recession; most of existing literature on resource misallocation focuses on developing countries and seeks to explain their relatively poor productivity performance. In addition, we plan to extend the existing modelling framework to explicitly account for a process of dynamic capital adjustment. That is, to account for the presence of costs that inhibit the adjustment of capital even in normal times.

Productivity growth is the source of rising living standards and economic growth. Understanding the causes of weak productivity is therefore central to identifying the policy challenges and designing the solutions. For example, the extent to which low productivity is a temporary or permanent problem is central to judging the UK's supply capacity, and whether expansionary policies are called for. This is of particular importance to monetary policy, which must consider the inflationary consequences of economic expansion. This project will bring together expertise from the Bank of England and the IFS with a view to advancing the solution to the productivity puzzle and directly informing policy makers.

Planned Impact

Productivity growth is the key driver of living standards and economic growth. As such, understanding the drivers of productivity is of interest to a wide group. More specifically, understanding the causes of weak productivity following the recent recession (including how persistent the problem is and where distortions are arising) is central to identifying the policy challenges and designing the solutions. The results of this research will inform an area of ongoing concern, and therefore be of direct benefit to a wide range of policy makers. It will do this in a timely fashion, such that it is relevant to the current policy environment, as well as to any future crises.

Benefits to monetary policy makers

Understanding productivity has been one of the key research challenges at the Bank of England since 2008. In the broadest terms, the outlook for productivity maps directly onto the question of how much supply capacity the UK has. This in turn is crucial for deciding whether expansionary policies are called for and what the inflationary consequences would be. To the extent that there are specific financial market frictions impeding productivity growth, understanding these factors is directly relevant when considering other forms of central bank policy (such as the 'funding for lending' scheme).

The direct collaboration embedded in this project will ensure that the views of monetary policy makers feed directly into the design and development of the research. It will also provide a direct link between the output of the research and a key group of beneficiaries.

Benefits to other policy makers and experts

The research is also relevant in providing important context for fiscal policy. For example, there is currently a large debate about stagnating living standards and the appropriate policy response. Understanding weak wage growth, which is likely to be directly related to weak productivity, is central to this debate.

Despite the UK specific context of this research, it will speak more broadly to the literature on recoveries from economic crises. This will be relevant for a wide group of policy experts, including those working in other central banks and governments. We would expect the research to be of particular interest to those European countries that have also experienced weak productivity since 2008.

We expect international organisations (e.g. the IMF and OECD) to be interested in our research to the extent that it helps to understand the differing international experiences in the post crisis period. The research will also benefit the business, professional and media communities by providing them with better information, and thereby help them to engage in an evidence-based debate with policymakers on an important policy issue.

Finally, this project will have a clear impact on the PI, for whom this project will be a pathway to further developing skills in the design, conduct and management of research. Both parties in this collaboration bring different strengths to the project. This will be beneficial to the overall research outcome, but also to the researchers involved.
 
Description This project aimed to both have an immediate impact on the UK policy debate surrounding weak labour productivity and to contribute to the academic literature on measuring distortions to firms' capital choices.

The grant directly informed the UK policy debate. Most noticeably, the grant led to a large, very successful, conference on understanding the UK productivity puzzle. This was a one and a half day event, run by the IFS and the Bank of England, which bought together a significant number of stakeholders (there was an audience of 180) with 17 speakers, including academics, members of the policy community and policy makers. This injected insights from research into important policy debates. The event we ran shed light on (i) the importance of considering macroeconomic issues (including aggregate UK productivity) from a microeconomic perspective (e.g. by considering the behaviours of individual firms) and (ii) the merits of considering changes in different parts of the economy (e.g. changes in the labour market and firms' investment choices) simultaneously. (Further details are given in the 'engagement activities' section and all materials are available here: https://www.ifs.org.uk/events/1188). In addition, our work has directly fed into the discussions with policy makers due to the collaboration with the Bank of England (Alina Barnett, collaborator on the grant, is the chief economist to Ben Broadbent, Deputy Governor, such that her knowledge of our findings is directly used). Helen Miller has additionally had various meetings discussions with policy makers from HM Treasury about productivity and the allocation of resources.

The project has resulted in an ongoing collaboration between the IFS and the Bank of England. The collaboration is now being taken forward, with Helen Miller as a registered academic visitor at the Bank of England.

Our original idea was to exploit UK micro data (specifically, the Annual Respondents Database) to measure distortions to firms' capital choices. We were successful in conducting the analysis as planned on the data. Unfortunately, there were changes in the survey methods in 2008 which meant that it was not possible to distinguish between real changes around the recession and changes driven by survey changes. We additionally conducted analysis on an alternative source of UK micro data: company financial accounts (as provided by Bureau van Dijk via the AMADEUS database). These data raised a different set of challenges, most notably including that data is often missing for certain types of firms and the measure of firms' capital stick is based on accounting standards (which differ from the concept of capital used in economic models).

In both cases we found that the method we were using, which was set out by C. Hsieh P. Klenow in The Quarterly Journal of Economics in 2009, was not robust to small changes in how variables were measured or which sample of firms were used. This was a major limitation given the data problems we faced. This is important for policy because changes in the estimated magnitude of the problem (i.e. of capital misallocation) would lead to differing conclusions as to whether this was a first order problem that required a policy response, or a small problem that was of relatively little importance. Unlike in some other research studies, it is not appropriate to simply focus on large firms (for who data is more trustworthy) because the question of interest relates to capital allocation across the whole economy; ignoring small firms would ignore a major channel of reallocation. We have written this work up as a report that: discusses the indicative evidence for capital misallocation in the UK; sets out the method used to quantify frictions and the associated problems; describes the issues involved with taking the method to UK micro data. Our results will inform other academics by highlighting the pitfalls of the specific method we applied and guiding future efforts towards alternate approaches.
We have been collaborating with a user group that contains the ONS, HM Treasury, Bank of England, NIESR, IFS and others to discuss lessons from various research projects and the issues involved with using the Annual Respondents Database. As such, one of the key outputs of this project has been to feed into this knowledge pool.

This grant has lead us to approach the question of why UK productivity is weak in a different way going forward. Specifically, the outcome of this grant has been to move our focus away from models that reply on being able to accurately measure outcomes for all firms, towards those that are robust to only observing a subset of firms. We have secured research funding to model the propensity of firms to undertake large, capacity enhancing investments before and after the recession. This is very relevant to the allocation of capital, since changes in allocation are often driven by large investments. Rather than try to quantify the effect of all possible frictions to capital allocation, we are exploring the specific effect of firms' financial position on investment.

The project was a new collaboration between Helen Miller at IFS and Alina Barnett at the Bank of England. The aim was to bring together the complementary expertise that was held in the two institutions, importantly including cleaned data that was held at the Bank and been developed over many years. The collaboration was a success in that we advanced our understanding and are continuing to collaborate. However, we faced the unexpected difficulty of sharing data between the IFS and Bank of England; it took almost a year to find a way to share the data between the two institutions (further details are provided under 'collaborations and partnerships'). We have worked to resolve that issue, such that the work in this grant has led to a precedent for data sharing between UK institutions that access UK micro data through different systems.
Exploitation Route Our findings have already informed our own subsequent research, and the research at the Bank of England.
The approach we sought to use was attractive because it provides a simple, tractable approach to quantifying the effect of frictions in capital choices. As discussed above, our research found that the method is not particularly robust. This finding will be of use to future generations of researchers.
We have learned valuable lessons on how best to manage collaborations between the IFS and the Bank of England (and more generally between institutions that do not access ONS micro data through the same system). There is now a precedent for data sharing between institutions using the virtual microdata laboratory and Secure Data Service. Going forward, the IFS and the Bank of England are pursuing their collaboration in a different way. In a new research collaboration (on investment and uncertainty), all data work will happen at the Bank of England, and Helen Miller will provide input as a registered academic visitor to the Bank.
Sectors Other

 
Description As highlighted in the key findings, we ran a large conference as part of this grant. Policy makers (including civil servants and members of the Monetary Policy Committee) attended and spoke on panels. Following the event there were further internal discussions between the research team (particularly through Alina Barnett) and staff at the Bank of England. Helen Miller had numerous private meetings with civil servants at the Treasury. Our results, including those relating to data quality and the robustness of methods, thereby informed the thinking and subsequent research of policy institutions. Our efforts to establish research collaboration between the Bank of England and Institute for Fiscal Studies led directly to a precedent for data sharing between the Secure Data Service (SDS) and the Virtual Microdata Laboratory (VML). This will impact other researchers in government institutions (notably the Bank of England, government departments and the ONS ) who access data via the VML and wish to enter research collaborations with academics (who access data via the SDS).
First Year Of Impact 2015
Sector Government, Democracy and Justice,Other
Impact Types Economic

 
Description IFS and Bank of England 
Organisation Bank of England
Country United Kingdom 
Sector Public 
PI Contribution This project was a collaboration between Helen Miller at IFS and Alina Barnett at the Bank of England. The aim of the collaboration was to bring together the complementary expertise that was held in the two institutions, including the access to networks of researchers working on related questions. Most of the research was conducted by Helen Miller at the Institute for Fiscal Studies.
Collaborator Contribution The aim of the project was to make use of the Annual Respondents Database (ARD), which is derived from compulsory business surveys and collated by the ONS. One of the key benefits of working with e Bank was that they had undertaken substantial amounts of work over multiple years to produce a dataset that is usable for research. (This data is not straightforward to use. Importantly, because the underlying surveys were changed in 2008, a lot of work is required to create consistent data. The 2008 break is of particular concern for us because we are interested in the effect of the recession.) This project built from that effort by using the cleaned ARD data, at IFS, in a specific application. Alina Barnett also contributed to the research undertaken, and was important in communicating findings to policy makers at the Bank of England. The Bank hosted and contributed speakers to the large conference that took place as part of this grant.
Impact Alina Barnett and Helen Miller, "Measuring the effect of capital allocation on productivity", Report, 2017 Bank of England-IFS conference, "Understanding the great recession: From micro to macro", September 2015 Helen Miller, 2015, "Capital allocation and productivity", presentation, Bank of England The collaboration was a success in that we advanced our understanding and are continuing to collaborate. However, the collaboration was also difficult in an unexpected way. The Bank of England accesses this data through the ONS virtual microdata laboratory. IFS access the data via the Secure Data Service. We were unable to find a way for both institutions to use the same system. Instead, we have worked with ONS and the SDS to find ways to allow researchers to transfer cleaned data from the VML to the SDS. It took us almost a year to reach this solution, which is far from the perfect solution since it still requires running the project in parallel on the two systems. However, it does represent an output of this project in that there is now a way forward for collaborations and a precedent for data sharing between institutions using the VML and SDS.
Start Year 2014
 
Description 'Understanding the Great Recession: from micro to macro', a conference jointly organised by IFS and Bank of England 
Form Of Engagement Activity Participation in an activity, workshop or similar
Part Of Official Scheme? No
Geographic Reach National
Primary Audience Policymakers/politicians
Results and Impact The event was held at the Bank of England and organised by: Helen Miller (IFS), Richard Blundell (IFS) and Alina Barnett (Bank of England)

Participants:
- 17 speakers (including a Keynote address by Professor Robert E. Hall, a speech by Deputy Governor Ben Broadbent and a policy focused panel discussion)
- 160 attendees (excluding speakers) signed up

We were able to attract high quality speakers that are working on research that uses the study of micro economic agents (e.g. firms and consumers) to understand aggregate trends (e.g. productivity and incomes). This included speakers from Stanford University, University College London, the London School of Economics, Cambridge University, the OECD and the Federal Reserve Bank of San Francisco. The presentations were mainly focused on the UK, but there were also contributions that covered the experience of other countries (including the US and Germany).

We ran 3 main sessions, which looked at: labour markets since the crisis; trends around the recession; productivity and the allocation of resources. In addition Ben Broadbent, the Deputy Governor of the Bank of England, gave a policy speech on the UK labour market and Professor Robert E. Hall (Stanford) gave an hour keynote lecture on how we can understand the US recession (in a later panel he also spoke on the UK). We were keen to draw out implications for UK policy. To this end we organised a panel discussion at the end of the event that included academics and policy (fiscal and monetary) makers.

Throughout the event we scheduled time so that audience members could ask questions and participate in debates. Many people commented that they found this very useful.

The event was very popular. We reached the capacity for the venue (the Bank of England conference centre) within just a few days of advertising the event and were running a large waiting list. At around 180 our attendees list was much larger than we had initially anticipated (we had planned for 70). The attendees included academics (including PhD students), policy researchers (largely from the Bank and HM Treasury but including HMRC, BIS, OBR, ONS), and professionals (including from commercial banks and professional bodies). We got great feedback, both from speakers and attendees about the quality of the presentations and the surrounding discussions.

Co-funding for the event was raised from the Royal Economic Society; the venue and a contribution to other event costs were provided by the Bank of England.
Year(s) Of Engagement Activity 2015
URL http://www.ifs.org.uk/events/1188