Posts Tagged ‘LSE Comment’

The data behind mortality trends: explaining the recent improvement in mortality in England

One of the most important functions of a government is to ensure the health of its population, with the main indicator being measures of mortality such as life expectancy. Mike Murphy writes that, contrary to popular belief, current levels of mortality are the lowest ever-recorded by a substantial margin.

Recent reports of adverse mortality trends in Britain have attracted considerable media, academic, and political attention. The headline statement that life expectancy has fallen by six months has been widely disseminated: it was quoted in Parliament by the Leader of the Opposition, and in the valedictory report by the past-President of the Public Health Faculty in July 2019. If true, this was a major event – annual life expectancy at birth in Britain had never declined by such an amount since 1951.

In fact, the figure referred to a revision of forecasts by the Institute and Faculty of Actuaries of life expectancy of those now aged 65. Since about half of these people are likely to be alive around age 90 in a quarter of a century’s time, this figure will be determined by what happens over the next 40 years or so, especially in the latter part of that period.

The misinterpretation of such data had long been pointed out, but continues to be widely disseminated. The British Medical Journal has already published a number of items this year asserting that mortality is deteriorating or at best stalling, however none of these include the most recent data available.

Measuring trends in mortality

The most widely-cited source for life expectancy figures is the Office for National Statistics (ONS). Unlike other countries, UK presents life expectancy as three-year averages rather than as single-year estimates, to smooth out annual fluctuations largely arising from seasonal variations in diseases such as influenza. The release in September 2018 attracted wide attention, since it reported that life expectancy had stopped improving for the first time since the current series was started in 1980-2. The conclusion was based on the fact that there had been no change between the latest average for years 2015-17 and the previous one over the period 2014-16 (i.e. simply one third of the difference between the annual values for 2014 and 2017, so some years out-of-date).

So, although life expectancy is the most widely cited indicator for this claim, it provides estimates of change that are some years out of date. Indicators such as Standardised Death Rates (SDR) are alternative valid measures of overall age-adjusted mortality, which – like life expectancy – are also wholly derived from age-specific mortality rates. However, SDRs are more timely, being published quarterly within three months of the end of the reference period, but only for England. Annual SDR values (necessary to control for seasonal variation) for England may therefore be computed up to the period July 2018 to June 2019 from published statistics (Table 1), although movements in other parts of great Britain tend to be very similar.

Table 1. Mortality indicators in Britain, 2013-19

Sources: see here and here.

The role of winter fluctuations

The major determinant of year-to-year changes in mortality is fluctuations in the number of deaths in the winter months of the year. The number of excess winter deaths in England and Wales – the difference between the number of deaths that occurred between December and March and the average number in the preceding 4-month period (August to November) and the following April to July – was exceptionally low in 2014: the value for the winter of 2013-14 was, by a considerable margin, the lowest recorded in the series starting in 1950-51. Therefore, 2014 had a particularly high value for life expectancy at birth and changes computed from that baseline are correspondingly low (Table 1).

Figure 1 shows quarterly SDR values over this century. The stalling of the trend in mortality improvement around 2010 is apparent, although there is little evidence of actual deterioration apart from short-term temporary fluctuations largely concentrated in the winter period. Although the importance of winter mortality for year-to-year change is not disputed, the debate about the cause and implications of recent trends has become acrimonious, polarised and politicised. The main reason for increased variability on mortality has been identified as the impact of seasonal respiratory disease and the varying effectiveness of flu vaccines; yet this conclusion has been disputed, with government austerity policies identified as the major factor for mortality stalling.

Figure 1. Quarterly Standardised Death rates (SDR), England 2001-19.

Source: see here.

The most common way of estimating winter effects is by comparing winter mortality indicators with the average of surrounding less-volatile non-winter values, as mentioned above. Estimates for England and Wales based on deaths are routinely published and we also present alternative estimates for England based on SDRs. Table 2 shows an SDR measure of excess winter mortality, the difference between the winter value for months January to March and the average of preceding and succeeding three non-winter months, for the periods just before and just after the start of mortality stalling. Average levels for the earlier period were larger than in the later one, but the percentage excess values are very similar, although this conclusion depends on the precise limits selected.

Table 2. Mean and standard deviation of winter excess SDR, England 2005-16

Sources: see here and here. This compares the winter value as the SDR for January to March quarter with the average of the surrounding non-winter quarters of October to December and April to June as an absolute value and a percentage. A broader alternative measure based on October to March compared with preceding and succeeding values in the April to September period is also presented.

However, variability is considerably greater in the latter period, which contains both the highest and lowest values in England and Wales of Excess Winter Deaths in the period since 1976. A substantial increase in excess winter mortality levels would have been necessary for it to have a major impact on mortality stalling, but the magnitude of observed changes is insufficient to account for any but a small fraction of the secular reduction in mortality improvement since about 2010-11. Therefore, although clearly important for annual fluctuations around the trend, winter mortality/influenza has had a more limited impact on underlying trends than sometimes claimed, while making the identification of trends more difficult.

Recent trends in mortality

A striking feature of these recent results is that England has experienced rapid and unexpected overall mortality improvement in the last 12 months or so. Figure 2 shows actual 12-month annual values and a fitted smoothed underlying trend. Mortality is currently at an historic low level.

Figure 2. Annual 4-quarter SDR; original and trend values, England 2001-19

Source: see here. There are separate series published for each quarter. These were combined to produce a time series of 4-quarterly averages to remove seasonal variation. The smoothed trend values were obtained from a non-parametric generalised additive model SDR(t) = a + s(t) from 2001. The estimated 95% CI shown as shaded: these values are probably underestimated since the observations are not independent.

These recent values do not contradict the well-established fact that there was a period of mortality stalling from about 2011 and that it was more marked in the period 2011-16 than in many comparator countries. High mortality in the first quarter of 2015 attracted considerable concern, as did high mortality in early 2018, which has continued into 2019. In contrast, more recent values have attracted little attention. Mortality has been consistently low throughout the last set of quarterly values: Q3 and Q4 of 2018 were the lowest recorded quarterly values. The first quarter of 2019 was equal lowest with the exceptional value of Q1 2014, but 9% below the corresponding value in Q1 2018. Q2 2019 SDR was the lowest value since the series started in 2001, over 2% below the next lowest value even though it contained a heatwave in late June, although some of any additional deaths will occur in Quarter 3, when there was another heatwave in late July.

Data on short-term changes, especially those using atypical baselines, are a poor basis for drawing conclusions about underlying trends. Caution is required before over-interpreting a single year’s data as happened for 2015, although in that case, the English SDR values in the last two quarters of 2015 were actually lower than those in 2014. Annual fluctuations have been driven mainly by those in winter mortality, whereas longer-term changes are more influential in other seasons. While random events such as epidemics can produce temporary increases, the mechanisms that might be responsible for a short-term substantial decrease in non-winter mortality are less obvious. The magnitude and consistency of mortality improvement in each quarter in the past year suggests that more attention be given to monitoring recent figures.


England is unusual in providing up-to-date sub-annual, age-adjusted estimates. Provision and dissemination of more timely and relevant data across the UK, however, would improve understanding and monitoring of trends: for example, by publishing and publicising more timely annual life expectancy data across the UK. Even the rolling 4-quarter moving average trend data presented above require aggregation of four separate quarterly series.

The very recent substantial improvements in mortality in England were largely unexpected (unfortunately, similar data are not available elsewhere in the UK) as was the earlier stalling, but such events provide especially useful information for testing some of the hypothesised explanations for recent trends, such as UK Government austerity policies or reduced real health and social care expenditure, that have been largely based on correlations of relatively short runs of macro-level time series. Recent national mortality trends are likely to attract considerable attention in any forthcoming election campaign, and the need for understanding recent movements is clear.


About the Author

Mike Murphy is Professor of Demography at LSE and International Visiting Professor at the University of Helsinki.




All articles posted on this blog give the views of the author(s), and not the position of LSE British Politics and Policy, nor of the London School of Economics and Political Science. Featured image credit: Pixabay (Public Domain).

Empty Bills: The Queen’s Speech was an odd contribution to solving the UK’s problems

Artemis Photiadou and Alice Park draw on various strands of research to argue that unless a Conservative manifesto is more radical and relevant than this Queen’s Speech, then a future Johnson government will fail to address fundamental issues, many of which have been caused by other Conservative policies.

Though the likelihood remains the Queen’s Speech will be voted down in the House of Commons – a course which may be avoided if the DUP and the 23 former Conservatives vote with the government – most of its content has already been written off as empty because of the government’s non-majority, while the opposition has dismissed the whole event as an attempt to misdirect attention. Nevertheless, the Conservatives continue to appeal to a large proportion of the electorate, they continue to top the polls, and may win an election in the near future. For this reason, the Queen’s Speech has been largely seen as the basis of a future Conservative manifesto.


The legislative agenda of this parliament will, as with the last one, be dominated by one subject: Brexit. Indeed, most other promises are contingent on how Brexit is delivered.

A number of significant bills that did not complete their passage through parliament before the end of the 2017–19 session will need to be re-introduced. As part of the legislative preparations for departure, a new Immigration and Social Security Co-ordination (EU Withdrawal) Bill has been on the parliamentary agenda since 2017, and Johnson and Home Secretary Priti Patel have consistently emphasised their determination to end freedom of movement and reform the immigration system.

The proposals include an end to freedom of movement, and introduction of a long-touted ‘points-based’ immigration scheme for EU and non-EU citizens alike after the transition period in 2021. As Heather Rolfe explains, this government has indicated a move away from Cameron and May’s ‘numbers game’ approach, and focuses instead on an immigration agenda based on encouraging only high-skilled migration. While the details remain vague, such a system often polls favourably with the public, who, according to focus groups, interpret it as a system that would ensure controls were enforced. However, the focus on high skills is generally of less concern. Both the public, and in particular employers, consider that it is not so much skills, but a system that responds to labour shortages that is key. Especially in sectors such as hospitality, construction, agriculture, and social care, reliance on migrant workers is not because Britons lack the training but because there are not enough of them. The absence of any detail in the Queen’s Speech, combined with the fact that a points-based system contradicts Johnson’s vision for an outward-looking post-Brexit UK, mean the policy remains symbolic, useful for confirming the government’s opposition to free movement.

Furthermore, ending freedom of movement  will have other consequences for overall migration flows, as Jonathan Thomas outlined, drawing from past policy interventions. Paradoxically, the end of freedom of movement for EU citizens, and the effect of the settled status scheme, could lead to a rise in irregular migration patterns, with knock-on effects for enforcement and public attitudes – far from the control that is sought and promised.

The NHS 

Healthcare remains a top public concern. Yet the main feature of the NHS section of the Speech is that it promises more money, with little else beyond that. This is a problem for the simple reason that it is unclear how and where NHS money is currently going. According to CHPI research, whereas officially NHS England spends about 7% of its total expenditure purchasing healthcare from the independent sector, new calculations put the figure closer to 26%. In short, with a flawed understanding of spending, any attempt to increase resources without controlling expenditure will be in vain.

Second, there appears to be no intention to hasten the growing trend towards privatisation, for example through strengthening the rules governing financial incentives in the sector. As the system currently stands, NHS consultants can – and often do – own shares and equipment in private hospitals. The result is the creation of an incentive towards private referrals and possibly over-treatment – given they receive fees whenever equipment they own is used – a prospect which is particularly problematic when seen together with the lack of transparency on NHS spending.

Similar problems exist with adult social care pledges. While the government plans to provide more funds to councils, they do not complement this with efforts to stabilise the care home market through regulation: even with more money made available, a number of care homes will continue being run by hedge funds and private equity investors. The extraction of rent and profit will therefore continue being the market’s top aim, and scandals like Four Seasons are almost destined to be repeated.

Mental health is similarly covered in broad terms, promising that ‘patient choice and autonomy will be improved’. One way of reading this tallies with pre-existing self-care approaches, part of a Conservative tendency towards behaviour change and short-term self-action, rather than medical intervention, despite evidence that the former is ineffective. A more effective way of beginning to tackle problems in mental health would be the reversal of what appears to be the strategic downgrading of jobs and working conditions, with many employees on short-term NHS contracts.

In light of fundamental issues remaining unaddressed, it is questionable how successful the smaller proposals – on laws to establish a health service safety investigations body and to make it simpler for NHS to manufacture and trial new medicines – could hope to be.


While the NHS is afforded two bills, crime, which is a lesser public concern, has seven. The government’s focus on this increasingly seems to be, as Thomas Guiney explained, an attempt to disassociate ‘the Conservative brand from its cornerstone’ austerity policies, ‘while maintaining a reputation for fiscal prudence’. Yet here again the austerity of the past decade is what goes to the root of the problem, not insufficient sentencing regimes. With poorer men between 17–34 being most likely to be both victims and perpetrators of serious violent crime, research suggests that the economic changes that ‘gradually rendered considerable numbers of the British working class economically obsolete’ have contributed to this state of affairs, together of course with cuts to policing. Without addressing the causal factors, the government’s new policies, such as the Foreign National Offenders Bill, combining this attempt to appear tough with an overt aversion to immigrants, are unlikely to tackle crime.

Fairness and protection for individuals and families

There is also an attempt to tackle domestic abuse, with the Domestic Abuse Bill being reintroduced, as had been agreed prior to the unlawful prorogation. This is generally perceived as a positive step towards tackling domestic abuse, not least because it defines it. But, as experts have warned, the government’s own policies in other areas actually facilitate some of the kinds of abuse that the bill seeks to prevent. Concerns pertain particularly to economic abuse, including cases where victims cannot afford to leave their abusive partner. Marilyn Hayward has pointed out how these situations become almost inevitable due to the way Universal Credit works, since couples can only nominate one bank account for the household benefit to be paid into. As it stands, therefore, the government gives with one hand and takes away with the other.

Such cases clearly illustrate how intertwined issues of abuse, gender, class, and austerity are, with the above bill and the rest of the Queen’s Speech showing little intention to address the last three. On a similar point, despite the promise to bring much-needed oversight over pensions savings and ‘tackle irresponsible management of private pension schemes’, there will be no redress for the women who suffered from the change in the state pension age – of which they were unaware. Here again the persisting gender and class inequalities are evident: around 80% of women with low levels of education knew about the change compared to 92% of women with high levels of education; ‘those out of the labour market, ethnic minorities, and unmarried women were also less likely to be aware’.

Other legislative measures

One eye-catching, and largely unexpected new bill (until leaks appeared over the weekend), is a new electoral integrity bill. Its main proposal is to introduce a photographic ID requirement for all voters at polling stations for UK parliamentary elections in Great Britain (it is already required for Northern Ireland voters, where you can obtain a free ID card) and for English local elections. This follows on from voter ID trials that have taken place in several areas during the 2018 and 2019 English council elections, part of  a package of reforms Sir Eric Pickles proposed in a 2016 report on electoral integrity.

However, the number of reported cases of electoral fraud from personation at the polling station are exceedingly low in the UK – just one case was prosecuted in 2017 –  and the available evidence suggests it is rare: the trials in 2018 and 2019 suggests the percentage of voters turned away, who did not subsequently return, was very low. Yet the effect for general elections is not necessarily the same as for lower-turnout local elections. The Electoral Commission has noted that as many as 3.5 million voters may not have photo ID, with ownership differing between socioeconomic groups, with citizens from ethnic minorities in particular potentially disadvantaged. So, there are considerable concerns that the negative impact of voter ID laws on restricting access to voting by those without ID outweigh any potential impact on countering electoral fraud, or improving trust in the system.

Another question is whether, given the low numbers of cases of voter personation, this is the right legislative priority for improving elections – and whether further piecemeal reform like this is the right approach at all. Earlier this year, the Commons’ Public Administration and Constitutional Affairs Committee asked for evidence on reforming Electoral Law, noting that many rules dated back to the nineteenth century, and had been updated piecemeal, creating a fragmented and confusing set of laws. Proposals from the Law Commission in 2016 for systematic review and simplification had not been acted on. Much of the academic evidence the committee received reiterated this view that consolidation and simplification of fragmented law, particularly in the context of devolution, was a priority over voter ID, which would introduce further complexity. Comprehensive reform aimed at simplifying electoral law, tackling changes to electoral campaigns in the digital era, and improving the system of voter registration, were seen as more important.

Suggested measures that would improve voting in the UK, and be a better use of resources according to Toby James and Alistair Clark, include a single national electoral website. Unnecessary complication with electoral registration, and a lack of resources, were contributing factors to many EU citizens being unable to vote in 2019’s European Parliament elections. The legislation proposed in this Queen’s Speech includes other changes, including tightening the rules on postal vote collection, and requiring re-registration for postal votes every three years, improved access at polling stations for voters with disabilities, and introducing digital imprints for online campaign material. However, the disproportionate focus on voter ID overall could add another layer of administrative complication for little benefit, and distract from more important, systematic reform, as highlighted by the Law Commission and others since.

The environment

What may be seen as both a popular and positive inclusion is the three pages devoted to an Environment Bill, setting binding targets for reducing pollution, with the topic recently having reached its highest level of public concern on record. But the caveat that runs through the Queen’s Speech remains: even if it passes, and even if the bill itself passes, it is unlikely that this government will be in place long enough to be tested on it.

Empty bills

Put together, and examined against policy research in the relevant areas, the government’s intentions seem set to deal with peripheral issues. What is more, non-issues such as voter ID could be turned into real ones. A future Johnson government will need a more attractive and radical manifesto than this Queen’s Speech, if it is to meaningfully start solving the UK’s problems.


About the Authors

Artemis Photiadou is the Managing Editor of LSE British Politics and Policy and teaches in the LSE’s International History Department.



Alice Park is the Commissioning Editor the LSE British Politics and Policy blog and the Managing Editor of Democratic Audit.



All articles posted on this blog give the views of the author(s), and not the position of LSE British Politics and Policy, nor of the London School of Economics and Political Science. Featured image credit: UK Parliament, Copyright House of Lords 2019 / Photography by Roger Harris (BY-NC 2.0).

An asset to Boris Johnson: ideology in Brexit Britain

Despite all the comings and goings at Westminster, and the debatable qualities of the latest Brexit proposal, the underlying ideological dispositions of a large proportion of the electorate favour Boris Johnson and the Conservatives, write Joe Greenwood and Joe Twyman.

In the aftermath of the success of the Leave campaign in the 2016 referendum on the UK’s membership of the EU, and the ascendency of Donald Trump to the US presidency, there was an enormous amount of debate, discussion, and contemplation of what drove those outcomes. Was it economics? Or culture? Or, more likely, some combination of the two? There is now widespread agreement that a component of the explanation is the renewed prominence of a familiar but sometimes overlooked ideology: populism. In general, populism has been defined as a reduction of the world into an opposition between elites and masses, but this applies to varieties on both the left and right. Given political events, we were interested in the particular variety of that ideology that seemed to have played a part in driving them: authoritarian populism. In particular, we were interested in what beliefs constitute that ideology, and who the people who hold them are.

Using previous research on the beliefs of authoritarian populists as our starting point, we measured eight different beliefs: support for patriotism, support for a ‘strong and tough’ foreign policy, disapproval of the EU, distrust in EU institutions, scepticism regarding the need to protect human rights, a negative view of immigration, opposition to immigration from outside the EU and self-reported left-right political placement. To find out how widespread these beliefs are amongst the British public, we asked a representative sample of 14,923 adults from YouGov’s panel of respondents for their views. Our survey was fielded at various points from 30 May 2017 to 2 March 2018, and thus encompassed a range of political events including the 2017 general election. The resulting sample gave us an excellent opportunity to paint a picture of the ideological landscape in Brexit Britain.

Our first finding was that we live in a country of authoritarian populists. More than half of the public fall into an ideological group with notable authoritarian populist views. Overall, using a process in which we test which beliefs cluster together, we found five ideological groups in the population: mainstream populists (17%), centrist weak populists (20%), moderates (31%), left-wing progressives (18%), and right-wing populists (14%). Both of the first two groups are centrist in terms of left and right and display similarly strong levels of authoritarian populist beliefs. They differ only in the sense that centrist weak populists are less supportive of strong and tough foreign policy, and much less negative about immigration. Moderates are true to their name; centrist in left-right position, with minimal support for authoritarian populist beliefs except that, like the vast majority of the population, they are supportive of patriotism. Indeed, the only group that is notably less supportive of patriotism is the left-wing progressives (though they are not roundly opposed), and this group constitutes the most clear opponent of the authoritarian populist ideology overall.

Standing in stark opposition is the right-wing populists who, as the name suggests are firmly on the right of the left-right spectrum and strongly supportive of all the authoritarian beliefs that we asked about. Overall, then, the picture is one of widespread sympathy for many of the elements of authoritarian populist ideology, with only two of five ideological groups displaying marked opposition: moderates and left-wing populists (albeit, they constitute almost half of the population between them).

But what do these groups look like in terms of demographics? Well, right-wing populists are the only group that is a majority male (58%), whilst also having the oldest average age (56) in contrast to the much younger moderates (average: 44) and left-wing progressives (42). In terms of social grade, the only group that draws a majority from the lower categories (C2, D, and E) is mainstream populists, whilst moderates and left-wing progressives are marked by being majority higher social grade (A, B, and C1): 61% and 68% respectively.

This divide is also reflected in the educational qualifications of the groups: the two groups with the highest proportion of university educated people are moderates (42%) and left-wing progressives (58%). This all fits with the common observation that the adherents of the ideology that supported Brexit tend towards being older, male, from lower social grades, and with lower levels of education. However, it must be stressed that these are only tendencies: there are significant minorities of women, younger, higher social grade, and university educated people amongst right-wing populists, mainstream populists, and centrist weak populists. Indeed, this is why it is important to measure ideology in its own right: we cannot rely on demographics alone to understand why people behave the way they do in relation to politics.

One of the most important political behaviours is voting, and there was an important and dramatic realignment in party support amongst the ideological groups between the 2015 and 2017 general elections. Figures 1 and 2 show the proportions of each ideological group who cast their votes for each party in 2015 and 2017. In line with the tendency for the two largest parties to cannibalise the votes of smaller parties in 2017 (combined, they gained 82.4% of the votes cast) we observed a shift towards Labour and especially the Conservatives amongst mainstream populists and centrist weak populists. More dramatically, the already healthy 2015 levels of support for Labour (58%) amongst left-wing progressives grew to three quarters (76%) in 2017, whilst the two-thirds (67%) of right-wing populists who support the Conservatives in 2015 grew to nine in ten (91%) in 2017. We know the reason for these realignments: 2017 constituted ‘the Brexit general election,’ with voters largely moving to the party that advocated their preferred, or at least a tolerable, position on the UK’s exit from the EU.

But, dominant though it is, Brexit is only one manifestation of the underlying ideological cleavage with which we are concerned. It is important because it provides a ready shortcut for the population to understand the positions of politicians, and it has been exploited by politicians to activate and sustain this ideological divide. A key actor in that process of activation and sustenance has been Boris Johnson. In the days and weeks since ascending to the position of prime minister he has dedicated much time to deploying rhetoric and questionable constitutional manoeuvres designed to remind the swathe of the population with at least some authoritarian populist leanings that he is on their side. He no longer needs to worry about UKIP and has, to a significant extent, disarmed the threat from the Brexit Party, a trend that is likely to be magnified when voters decide where to place their cross on the ballot paper.

Given the UK’s electoral system, the newly salient divide between the Liberal Democrats and Labour over Brexit, and the apparent failure of Jeremy Corbyn’s attempts to re-activate left-right considerations amongst a large enough section of the electorate, Mr Johnson can afford to be cautiously optimistic the coming general election. If those conditions remain and he keeps banging the Brexit drum, there is a majority of the electorate with staunch or partial authoritarian populist beliefs who are likely to support him. Indeed, in addition to their genuine and heartfelt ideological positions and policy concerns, this might be part of the reason that many MPs have so far been disinclined to support holding another general election.


Note: the above draws on the authors’ chapter in a new edited volume in honour of Professor Anthony King, available here.

About the Authors

Joe Greenwood is a LSE Fellow in the LSE Department of Government, where he teaches on GV101 (Introduction to Political Science). He previously worked at YouGov and, before that, completed his PhD at the University of Essex. His research focuses on political participation, privilege, and perceptions in the British context. He tweets @niceonecombo.

Joe Twyman is co-founder and director of Deltapoll and a presenter of the Polling Politics Podcast. He was previously Head of Political and Social Research at YouGov, having been a director at the founding of the company back in 2000. He tweets @joetwyman.


All articles posted on this blog give the views of the author(s), and not the position of LSE British Politics and Policy, nor of the London School of Economics and Political Science.

Powered by WordPress | Designed by: index backlink | Thanks to insanity workout, car insurance and cyber security