The conventional wisdom in the mainstream poverty literature is that Margaret Thatcher’s policies caused an explosion of poverty. As one standard work puts it:

‘During the 1960s, just over 10% of the population lived in a low-income household. This rose slightly under the Conservative administration and following the oil shock in the 1970s, and then declined to about 8% during the mid-1970s. In 1979 […] changes in economic and social policy resulted in a trebling of the proportion of people living in low-income households from 8% to 25% – clearly showing that governments do have an effect on the amount of poverty in a country and that social policy does make a difference.’

A look at the lower half of the income distribution (the part which is relevant for relative poverty rates) shows where this impression comes from. Over the 1980s, real incomes in the middle of the distribution increased by about a quarter, while incomes at the lower end recorded only modest increases. At the very bottom, they even fell slightly. The change in the income gradient over the decade therefore looks more like an outward rotation rather than a parallel upward shift. Even at the end of the 1990s, those at the lower end of the income spectrum did not seem massively better off than they had been in 1979, while most of the rest of the distribution had pulled away. For most poverty researchers, this settled the case against ‘Thatcherism’ once and for all.

Figure 1: The income distribution (lower half) in 1979, 1989 and 1999, equivalised annual figures in today’s prices

 -based on data from IFS

However, in the mid-1990s researchers found that there was something wrong with the income data. They checked how closely income was related to expenditure, and found a fairly close relationship over most of the distribution – but not at the tails. In the bottom decile, expenditure levels were far above recorded income levels; and indeed they had to be, because in the very bottom percentiles, reported incomes were around zero. The authors repeated the same exercise for older data from 1979, and found that this problem had not existed back then. It must have been during the 1980s that in the bottom decile, income became decoupled from expenditure. This finding has subsequently been confirmed by numerous other studies (see here, pp. 123-133).

If we express living standards by consumption data rather than income data, a rather different picture emerges. There is no arguing with the fact that the UK became a more unequal society during and in the aftermath of Thatcherism. The gradient of consumption, too, became steeper over time. But the notion that the poorest have been cut adrift from the rest of society is a peculiarity of the income data. When Margaret Thatcher left office, consumption levels of the least well-off were about a quarter above their 1979-level.

Figure 2: The consumption distribution (lower half) in 1979, 1989 and 1999, equivalised annual figures in today’s prices

-based on data from IFS

This does not mean that the Thatcher/Major years were a wonderful time for the poor. In the early 1980s, the employment rate among low-skilled men dropped by about ten percentage points, and never fully recovered. There was a basic problem with the sequence of economic reforms, even if these reforms themselves were entirely reasonable. Under Thatcherism, attempts to subsidise economically unviable industries forever were abandoned, and so was the use of monetary policy as a tool for employment creation. From then on, job creation was to be achieved through freeing the labour market from trade union power and from state interference with the wage-setting process.

Very well, but unfortunately, the former happened before the latter was completed. Thus, when the major industry shake-outs occurred, people were left to look for work in a still largely unreformed labour market, which was not fit for reintegrating them.

So yes, there were things that could have been handled differently. When it comes to poverty, the 1980s were not an especially glorious decade. But poverty did not jolt up during this period. The received wisdom is the product of a statistical artefact.


Head of Health and Welfare

Dr Kristian Niemietz joined the IEA in 2008 as Poverty Research Fellow, becoming its Senior Research Fellow in 2013 and Head of Health and Welfare in 2015. Kristian is also a Fellow of the Age Endeavour Fellowship. He studied Economics at the Humboldt Universität zu Berlin and the Universidad de Salamanca, graduating in 2007 as Diplom-Volkswirt (≈MSc in Economics). During his studies, he interned at the Central Bank of Bolivia (2004), the National Statistics Office of Paraguay (2005), and at the IEA (2006). In 2013, he completed a PhD in Political Economy at King’s College London. Kristian previously worked as a Research Fellow at the Berlin-based Institute for Free Enterprise (IUF), and at King's College London, where he taught Economics throughout his postgraduate studies. He is a regular contributor to various journals in the UK, Germany and Switzerland.

11 thoughts on “Thatcherism and poverty: did the poor get poorer?”

  1. Posted 15/04/2013 at 10:36 | Permalink

    Do consumption levels reflect income, though? Did larger amounts of the poor use credit and remortgaging etc than before? This just creates debt, not real income. I could be wrong, but it might be worth checking.

  2. Posted 15/04/2013 at 11:58 | Permalink

    Great blog post

  3. Posted 15/04/2013 at 13:31 | Permalink

    Tanya, good point, but while the use of consumer credit did increase during the 80s, it did not increase disproportionally among low-income groups. So credit can explain why the variation in expenditure increased at any given level of income, but it cannot explain why this phenomenon is much more pronounced at the bottom of the income distribution. It is the bottom of the distribution which shows the most atypical figures.

  4. Posted 15/04/2013 at 16:05 | Permalink

    A similar phenomenon is observed looking at VAT as function of income and expenditure. It is often claimed that VAT is regressive, which always seemed odd to me since those on the lowest incomes will tend to spend more proportionately on VAT-exempt or low rated items such as food, housing and domestic fuel, yet the figures show a big upturn in the proportion of income spent on VAT specifically amongst the ‘poorest’ 10% compared to the deciles immediately above. The IFS, however, looked at this and found that VAT is not regressive as a function of expenditure, since those on lowest incomes spend far more than their income. The IFS concluded that this is because many of those with very low incomes are not poor at all – they may be temporarily between jobs (but living off savings) or they may be simply spending saved money once retired. Consequently, the composition of the supposedly ‘poorest’ 10% is highly distorted – there will be some who are, indeed, poor, but a good proportion will be very far from poor.

  5. Posted 15/04/2013 at 20:19 | Permalink

    The obvious explanation is that the bottom of the income distribution represents different people in different years. So if the year to year variation of income goes up, the person at the bottom today is more likely to be someone who was doing better last year and will be doing better next year, and he average his consumption over time by either saving when his income is higher or borrowing when it is lower.

  6. Posted 16/04/2013 at 12:00 | Permalink

    David, yes, that’s part of the explanation; we can see that the composition of the bottom decile has changed over this period. In the 1970s, these were mostly pensioners. By the 1990s, it contained more self-employed people, and more working-age unemployed. So we would expect an increase in the share of people whose momentary income deviates from their permanent income, and therefore a looser relationship between income and spending in the bottom decile.
    But that’s not the whole story. We also have data on assets and savings, which show that most low-earners do not have the reserves to engage in consumption smoothing. The most important explanation seems to be a misreporting of benefit income.

  7. Posted 18/04/2013 at 11:42 | Permalink

    Although expenditure is a marked improvement on income, taking into account cash benefits that are not reported as income, consumption figures would be far better. Thats what we mean when we talk about living standards. The value of free-to-user education and health services in particular has been increasing over time, and are disproportionally used by the poor (the rich go private)

  8. Posted 18/04/2013 at 15:50 | Permalink

    Hmmm. It’s a tricky one though, because if we go down the road of trying to put prices and values on universal education and health services, would we need to start factoring in who generates the wealth and how they are rewarded for it? There are more of the ‘poor’ and ‘middle incomes’ and they generate the vast majority of the wealth, but get paid relatively little for it. A call centre worker is barely on a living wage, say, but generates the wealth that allows his/her boss to have private education/healthcare etc….theres potentially a whole spiral of calculations to factor in if we start thinking more deeply about wealth and spending. I’m no expert, but I think NEF have tried to calculate things like this.

  9. Posted 18/04/2013 at 17:00 | Permalink

    Nope, doesn’t matter “who” paid for it. All that the individual cares about (and since we are assessing their well-being, all we should care about) is their utility from consumption, which we approximate by the value of their consumption (and deferred consumption from savings).

    They “pay” for their access to health etc. via taxes. But since we don’t count taxes as part of their income or their expenditure, they are irrelevant. Assigning a monetary value to entitlement goods isn’t too hard – just look for a free market proxy, and assign the market price.


    Under the line, because it doesn’t matter for the standard of living calculation, but how do you figure that those on “poor and middle incomes” produce the vast majority of wealth? That’s true, but only because there are a greater number of them. People tend to get paid the marginal value they create, whether that’s through the application of their labour or of their capital. And thats before all the things we do to smooth consumption inequality. Almost certainly the very poorest in society consume more than the value they create, since their earnings are increased by state provision and benefit entitlements. Its the middle classes and the rich who end up with less than they created. (50% tax rate, for example)

  10. Posted 18/04/2013 at 17:10 | Permalink

    Bill, Tanya-

    Valuing consumption makes sense for near-market services when we have good reason to assume that most people would buy something similar on the market anyway. Example: Medicare and Medicaid in the US. Here, we can check how much an equivalent private health insurance policy would cost, and add that cost to consumption. In the UK, one could also treat implicit rent subsidies in social housing in this way: We would have to estimate what it would cost to rent an equivalent flat privately, and add the difference to the income or expenditure of social housing residents.
    But that is about as far as this can be taken. We all ‘consume’ public services, but their cost tells us little about what they’re worth to us. The problem is that they are not fungible. We cannot tell the state ‘Actually, I’d rather have a bit less of that, can you refund me the tax money so that I can buy something else with it?’

  11. Posted 19/04/2013 at 16:12 | Permalink

    “We cannot tell the state ‘Actually, I’d rather have a bit less of that, can you refund me the tax money so that I can buy something else with it?’ ”

    Now that would be interesting.

    Many people object to their tax being spent on things they do not agree with. How many people would agree to the private funding of the police, the armed forces, education, healthcare and social services out of their income? As in no tax for these being taken at source.

    I think if people were refunded for the services they did not use and people who over used the services were charged the extra it would be a very interesting situation indeed.

    Very informative blog by the bye.

Comments are closed.