[This post accompanies the text in p225-231 of the printed edition of the book, (The reconstitution of the working class in Chapter 8, ‘Neither bread nor freedom:’ conditions for popular revolt)]
One of the many challenges of producing a very long book is that there was not much space to include graphs and tables. I didn’t want to overwhelm the text with figures and didn’t want to put readers not used to academic-style texts off by a heavy dose of statistics. One of the purposes of this blog is to provide a space to elaborate on some of the points which were compressed in the printed text and to share some of the data I was working with. I’d welcome comments and criticisms from comrades who may have more experience in quantitative methods than me (my academic training, at least until recently, has been largely in qualitative methods). However I appreciate the clarity and insights that quantification and representation of arguments in graphical form can sometimes bring. It is always important to remember that tables, graphs, databases and so on are just that – arguments and attempts at persuasion – not ‘facts’. Trying to relate a Marxist approach to understanding class to the statistics which are produced by governments and collected by international bodies (primarily the International Labour Organisation – ILO) is challenging for a number of reasons. Building on the discussion about the role of the middle classes in revolution here, it is easy to see that the categorisations used in international standards related to status at work such as ISCE will potentially put people in different social classes in the same category (for example a highly paid manager will be an employee, just like the clerical or service worker in the same company but they have counterposed class interests and will almost certainly find themselves on opposing sides in the class struggle).
There are of course other problems with national statistics, particularly in countries in the Global South. Statistics are expensive to collect and in contexts where the institutions of the state are weak and underfunded they tend to be highly inaccurate. They are also biased by the intentions of the people and institutions collecting them. Governments and UN bodies are (understandably) not counting workers in order to help develop a strategy to overthrow the bourgeoisie, they want to understand what people do at work in order to manage industrial relations more efficiently, extract more taxes (or lower them) or plan capital investment and develop education and training programmes to raise productivity. The statistics they collect are skewed for ideological reasons, by the lure of convenience sampling, through changing estimations of the strategic importance to the ruling class of different kinds of work in a particular context.
Nevertheless, national statistics on status at work can provide a starting point for creating a very low-resolution picture of what the labour force looks like in a specific country or group of countries, and how that picture changes over time. I have therefore used them in my book as a starting point for a series of arguments which are crucial to the overall analysis.
The first two points I will cover here and illustrate with data which it was impossible to include in the printed book are:
- Employees form a substantial majority of the employed labour force in most of the countries discussed in the book.
- The proportion of employees in the employed labour has continued to grow since 1991 which corresponds roughly to a period of deepening neoliberal economic reforms.
These points are important for several reasons. Firstly, because although as noted above, employees (in other words people who work for wage under someone else’s direction) are going to be the heart of the working class according to a Marxist definition. Therefore if we want to make some back-the-envelope calculations about the rough size and social weight of the working class in a given context then looking at what proportion of the employed work force are taking home a wage as opposed to working for themselves or members of their families one starting. There are plenty of health warnings which need to be taken into account about the ILO’s statistics of course – including the fact counting someone as an ‘employee’ under a classification schema such as ISCE-93 doesn’t tell us anything about how regularly they are doing waged work, nor what they get paid, nor whether if they stop working it will hurt the profits of their boss or disrupt the functioning of the state. It doesn’t tell us what is going on in that person’s head either and how they see their relations with other people in the workplace.
But the snapshot of employment status which emerges from these statistics across the countries discussed in this book is one where in most cases, waged work structures what most people do to make a livelihood.
A second reason this is important is that despite the highly ideological presentation of neoliberal reforms as shifting employment relationships away from waged work towards being your own boss, or being the boss for others, the general trend has been towards the growth of the relative proportion of waged workers in the labour force.
This confirms the analysis which Chris Harman put forward in his article, The Workers of the World, written twenty years ago in response to the emergence of the anti-capitalist movement and the influence within it of Michael Hardt and Antonio Negri’s notion of the “multitude” as an alternative agent of revolutionary change to the working class.
Visualising the data by country
These pie charts show a snapshot of the different sections of the employed labour force in 1991 and 2020, according to the data in the ILOSTAT database. The classifications of status are:
- Employees (ISCE-93 1)
- Employers (ISCE-93 2)
- Own-account workers (ISCE-93 3)
- Contributing family workers (ISCE-93 5)
These classifications are taken from the International Classification of Status in Employment standards set by the ILO and conform to ISCE-93, which is the standard most widely used by national statistical systems in the production of labour statistics. NB in the data used here there were no incidences of “Members of producers’ cooperatives” (ISCE-93 4) or “Workers not classifiable by status” (ISCE-93 6).
I took snapshots of the data at ten-year intervals from 1991-2021.
It is also significant to note that these are the ILO’s “modelled estimates”, which essentially means that statisticians at the ILO have filled in gaps left by incomplete data collected by national statistical agencies by developing econometric models to allow the publication of data which be compared across countries.
This means that the data here is a picture of what the ILO’s statisticians think the basic structure of what they call the “labour market” looks like in any given country, not necessarily what the national government thinks it looks like (although it will draw heavily on national government data where available).
There are some small differences in the dataset presented here compared to those used in the book itself (where I was relying on the modelled estimates from 2019) as some of the figures have been adjusted by the ILO, but these differences are not significant and the overall trends are the same.
Download an edited spreadsheet of the data here as a csv file.
Go to the original source here.
Employment status by country, 1991 – 2021
Blue = employees
Orange = employers
Grey = own-account workers
Yellow = contributing family workers
Just for contrast here are the same figures for India and the UK
Several things stand out from looking at this data across countries in addition to the points made above. Firstly, this is a quick way to grasp the variegated nature of the societies discussed in this book, a point I will return to in later posts in this series when discussing changes in the distribution of the labour force across different types of economic activity. Secondly, although the general trend in almost every country presented here is towards a larger proportion of employees in the labour force, there are some cases where the opposite is true, such as Yemen. In this case, as in several other cases where the growth in the number of waged workers appears to stall (such as Algeria between 1991 and 2001, or Syria between 2011 and 2021), one likely reason is the devastation wreaked by civil war. Secondly it is worth saying again that these crude figures can conceal as much as they reveal. In Syria, for example, the period of the counter-revolution and civil war sometimes resulted in government employees continuing to receive payment for work they could no longer carry out because of the shifting frontlines of battle. Likewise, large numbers of public service jobs in Iraq are said to exist as salary payments to “ghost employees”, which functions as a stipend for the recipient but has no relationship to any actual work. (Recently the Kurdistan Regional Government eliminated more than 50,000 “ghost employees” according to US development company DAI).
However, even if they have to be taken with buckets of salt and read in conjunction with other sources of data, these figures provide ballast to the argument that what Chris Harman wrote 20 years ago still holds true:
“The working class is not disappearing. It is not becoming bourgoeisified. It is not turning into a privileged layer. It is not gaining somehow from the impoverishment of wide sections of the Third World, especially Africa. It is growing even while it is restructured globally.”
The next posts will explore this in more detail.