ONS experimental data on green jobs
Over reliance on survey data, too much use of modelling and writing up research bulletins in an opaque style that yields texts presented in treacle and analysis obscured by fug.
On 14 March 2024 the ONS published its latest estimate for the number of green jobs in the UK economy. This is experimental data. It seems odd at a time when the ONS is having difficulty in constructing its regular statistics on employment, inactivity and unemployment that it allows its resources to be distracted into constructing experimental data that involves an opportunity cost at a time when the organisation probably cannot afford such distraction with interesting novelty. The experimental data based on survey evidence is a further illustration of the emphasis that official statisticians at the UK’s ONS place on the use of survey data rather than making greater use of administrative data in general. This is particularly pertinent in relation to the labour market where there are potentially rich sources of administrative data to track unemployment in a timely manner.
Employment in green industries by major activity
Source: ONS
Over reliance on surveys too little use of administrative data
Developing estimates of green jobs in the UK, however, is interesting in that it exemplifies several features of UK official statistics produced by the ONS and habits of analysis and presentation of research that can be found in the published work of other UK public sector organisations. This ONS research bulletin has four features that stand out and they are not confined to its work.
First the difficulty of using and relying on survey data rather than administrative data that is collected as part of everyday activities, such as the payment of social security benefits and tax receipts. In the instance of the experiment in identifying green jobs difficulty gathering data in this area is shown by the fact they rely on self-reported Opinion and Lifestyle Survey to gather self-reported data about people working in green jobs. They appear to rely heavily on User Engagement exercises to identify green data.
Too much mathematical modelling of data rather than use of hard recorded data
Second, the difficulty of data and statistical series constructed by the extensive use of mathematical modelling that marshals the available data from whatever source. The process of modelling in itself represents the construction of information that is several times removed from the raw data that lies behind it. In fairness to the statisticians engaged in this work the reference to ‘modelling’ appears to be not much more than an extrapolation of previous trends that will presumably be revised later as more data become available. It would be better to spell this out in a straightforward manner without the suggested ornamentation of the suggestion of modelling
Third, is the way that ONS statistical bulletins even when they are attempting to exemplify the underlying data they are reporting, have in their drafting a strange opaque quality where the strict grammatical meaning of the draft is elusive for a normal reader with a conventional command of grammatical written English. There is frequently an elision of meaning in the drafting. What is supposed to be a lucid and granular account turns into a succession of sentences and paragraphs that leave the reader who is interrogating them hanging in the air. Often a needed sentence to explain a point is missing or dealt with by referring the reader to a further part of the bulletin or to a previously published bulletin.
Opacity and elision of meaning
The helpful things that the reader is advised to refer to for further guidance is often almost identical in its drafting to the sentence pointing the reader forward to the reference for further information. Official publishers of statistics and data have to be careful not to gloss over the data or start to undertake the full analytical research and conclusions from it that are rightly the preserve of the users of data and will often be appropriately contentious. An element of reticence is understandable and desirable. Yet other public bodies in other countries manage to publish and explain their data without going beyond that important boundary. They do so, moreover, in a manner that is clear and cogent. This drafting with treacle, exemplified by the fug of self-referential bureaucratic process and is not confined to the UK’s ONS, it colours much of the drafting of publications from UK public sector bodies – the Bank of England, the Office of Debt Management, the Treasury and OBR.
All drafts and published texts are capable of improvement. The explanation of complex technical matters in a concise manner can easily lead all of us to be less than crystal clear. Yet the need for clarity and the neglect of drafting has become a feature of UK official publications. At first step would be to reacquaint civil servants with Sir Ernest Gowers’s Plain Words, which was first published in 1954 revised by Her Majesty’s Stationary Office in 1986. Although much of the problem goes beyond good drafting and turns on an increasing apparent reticence about engaging in clear thinking that cogently explores or exposes an issue. Hence the treacle like fug.
The danger of normative judgement eroding positive analysis
Fourth, there is also a substantive subjective issue that is creeping into the work of the ONS and its publications. This is offering intellectually and politically contentious judgement – by inference. The latest bulletin for example the nuclear industry is scored as ‘green’ with the statisticians in their draft noting that ‘stakeholder’ find that scoring nuclear as part of the green taxonomy is controversial.
Below at the end of this Substack are a series of gobbets taken from the latest ONS bulletin on Green Jobs and a series of examples of references that readers advised to turn to for further information. They are all available on the ONS website. They are presented at the end of this Substack so the reader can clearly see the ONS prose.
I simply turned to them last week as part of an almost daily routine of seeing what the ONS has published. I have a continuing interest in the climate change debate and what the social and economic consequences of an ambitious programme of de-carbonisation may involve. I have copied extensive chunks so that readers of this Substack can judge whether my allegation of analysis through treacle is reasonable or exaggerated.
The need for context
What appears to be missing from these bulletins from my perspective, is any framing of the data on green jobs in relation to context. An obvious question to explore is what proportion of all employment is involved in the green supply of goods and services, how many jobs are involved and what proportion of total employment they represent. In 2002 ONS experimental data estimate that there were 639,400 full-time equivalents jobs. Yet offers no context for those jobs in relation to total employment. In 2022 there were, for example, around 33 million jobs in the UK.
New green jobs and jobs displaced and lost through decarbonisation?
It would also be helpful for the bulletin to offer guidance to the reader on data that may illustrate the extent that green jobs may be displacing carbon emitting jobs. ONS research should be able as part of its experiments to publish some analysis of the extent that green measures that raise the cost of carbon have resulted in jobs being removed from the UK labour market. The loss of employment through decarbonisation should be a matter of intellectual curiosity in any experimental official data exploring the impact of the green agenda on the labour market. Indeed the loss of jobs so far from carbon emitting employers is of much greater analytical interest than the green jobs scored from a self-reporting and defining cohort taken from a survey. Not least given the implications that the changes have for the structure of the labour market, the composition of employment by gender, the structure of pay and the distributional and regional issues involved.
The UK has made decisive progress in reducing carbon emissions over the last thirty-five years. Much of it has been because of conventional orthodox Treasury advice. These include creating a more neutral tax system, applying expenditure taxes and withdrawing production subsidies. In fact, the broad OECD advice laid out at that start of the climate change debate in the 1990s. In the UK there were several policies that were special to its circumstances and political economy that related to the audacious programme of structural reform and improvements in the economy’s supply performance. These included a radical reform of industrial relations law; accompanied by a non-accommodating monetary policy and a higher exchange rate that expedited de-industrialisation after 1979; and the imposition of hard budget constraints on the nationalised industries, limited their scope to be supported by external finance. These changes to industrial relations institutions and the financing and privatisation of the nationalised industries had profound implications for carbon emitting sectors, such as steel, coal and electricity. The privatisation of electricity and the withdrawal of production subsidies from coal in 1992, resulted in a progressive reduction in the use of coal in electricity generation and much greater use of gas to generate electricity in the privatised industry in the 1990s. During the 1990s VAT at a reduced rate was applied to domestic supplies of fuel and power when its zero rating was ended in 1993.
In addition, while domestic households have been significantly sheltered from the full cost of successive governments' de-carbonisation measures since the 1990s, the full economic costs have been almost completely passed on to businesses involved in the generation of carbon emissions such as manufacturing companies and specialist sectors such as steel. It is impossible to read these bulletins published to explore green employment to obtain any sense of these dynamics.
Warwick Lightfoot
18 March 2024
Warwick Lightfoot is an economist and was Special Adviser to the Chancellor of the Exchequer 1989-92
Gobbets from ONS bulletins on Green Jobs
Developing estimates of green jobs in the UK
Experimental estimates of green jobs in the UK, detailing the methods used, strengths and limitations of the data, and future developments. These are official statistics in development.
Green Jobs team
Last revised:
14 March 2024
This methodology article is part of our project to define and measure green jobs. We detail how the definition of green jobs was reached and the activities that make up the framework which underpins the overall definition of a green job. We summarise the methods used to produce experimental estimates of green jobs using the industry, occupational and firm approaches. We also identify areas for future development. These are official statistics in development.
We identified three approaches to measuring green jobs, with different measures supporting different user needs and types of analysis. These include:
an industry-based approach, including all jobs in a green industry or sector, with industries classified according to the activities they carry out
an occupation-based approach, including all jobs that are green regardless of the industry they are in, based on the activities carried out by workers or the objectives of their work
a firm-based approach, including all jobs in a "green" firm, with such firms being classified based on, for example, their level of emissions
Opinions and Lifestyle Survey data
In the periods from 4 to 14 January 2024 and 17 to 28 January 2024, we sampled 4,985 adults and 4,980 adults, respectively, through the OPN. These samples were randomly selected from people who had previously completed the Labour Market Survey (LMS) or OPN. The responding sample for 4 to 14 January 2024 contained 2,594 individuals, representing a 52% response rate. For 17 to 29 January 2024 the responding sample contained 2,764 individuals, representing a 56% response rate.
The green jobs questions were asked of working adults only, resulting in a pooled responding sample size of 2,557 working adults. Our analysis included those who did not respond to the green jobs questions.
Survey weights were applied to make estimates representative of the population (based on ONS population estimates). Further information on the survey design and quality can be found in our Opinions and Lifestyle Survey QMI. Because of the limitations of the survey, we advise against drawing comparisons of green jobs estimates between waves at different time periods.
As part of the ongoing development of our data collection, we updated the questions used for the 2024 wave. The data we collected from OPN waves in 2023 were not edited prior to the production of estimates, resulting in estimates which represent respondent opinion of whether their job is green or not.
Reviewing this process, it was identified that not validating responses could result in inconsistencies in the dataset. For example, 23% of respondents who reported that they would describe part of their job as green, also reported that they spend no time on green activities. At the same time, 18% reported that they would not describe part of their job as green, but they also said they spend some time on green activities.
OPN questions asked in the 2024 waves were developed to enable validation of responses about whether a respondent's job is green, using information about their occupation. This validation reduced the number of people that appear to have a green job under the occupation approach by 7 percentage points.
Sample sizes on the OPN mean that we have not been able to carry out analysis of characteristics of those who describe any part of their job as green within regions. Further work would be required to explore how we can develop estimates and analysis to support understanding of green jobs across regions.