Smoking in United Kingdom
Analysis of trend in smoking and the effect of smoking ban on smokers in United Kingdom
Hong Ngoc Do
2024-06-23
1 Introduction
Tobacco epidemic is one of the most major public health concerns that
the world has ever faced. Each year, tobacco causes over 8 million
deaths globally, among which more than 7 million are the result of
direct tobacco use while around 1.3 million are the result of
non-smokers being exposed to second-hand smoke. (WHO Global Report,
2023). The share of smoking declined worldwide, however, the number of
annual deaths can be projected to keep growing, as tobacco kills its
direct users and second-hand smokers gradually. Of all tobacco use type,
smoking is the most common form.
The United Kingdom was one of the
heaviest sufferers from tobacco threats. Smoking causes 52% of cancers,
25% of all cancer deaths and takes around 80,000 lives in the UK every
year (Foster, 2024). It was the leading cause of preventable deaths and
diseases in the UK. In response to the rising number of related-smoking
deaths, the UK Government decided to take robust action to protect
future generations from these harmful products. Many positive outcomes
are evident, such as the current smoking rate in 2022 being 14.2%, which
is half of what it was in the early 2000s. Many reasons are argued,
however, in the scope of this working paper, I focus on the influence of
the Smoking laws by UK Government on the change in smokers’ number.
The following therefore takes a closer look at the current status of
tobacco use in the UK. The aim of this paper is to evaluate this
epidemic and analyze the intensity of Tobacco Ban implementation and its
influence in the future on smokers, finding out whether or not a trend
is formed.
In the Methodology section, I will explain in detail how
the study was conducted, the terms that are commonly used, as well as
the data collection methods and analysis. Chapter 3 focuses on the
comparison between the UK versus the rest of the world and some
countries in Europe in terms of tobacco prevalence, hence, a
relationship between tobacco consumption in UK and Tobacco Control
regulations is represented, along with that a prognosis for the future,
based on the basis of past developments in Chapter 4. Finally, the
findings of the analysis and the forecast are summarized.
2 Methodology
2.1 Overview and definitions
“Any tobacco use” is defined in this report as the use of any type of
tobacco – smoked and/or smokeless. “Any tobacco use” excludes use of
products that do not contain tobacco, such as electronic nicotine
delivery systems. Current tobacco use prevalence is defined as the
proportion of the population aged 15 years and older who use one or more
smoked or smokeless tobacco products on a daily or non-daily basis.
Tobacco products include cigarettes, pipes, cigars, cigarillos,
waterpipes (hookah, shisha), bidis, kretek, heated tobacco products, and
all forms of smokeless (oral and nasal) tobacco. Tobacco products
exclude e-cigarettes (which do not contain tobacco), “e-cigars”,
“e-hookahs”, JUUL and “e-pipes”. The rates are age-standardized to the
WHO Standard Population.
2.2 Data sources
This working paper obtaining data on the prevalence of tobacco use is
conducted by World Health Organization (via World Bank) – World
Development Indicators (WDI). WDI is the primary World Bank
collection of development indicators, compiled from
officially-recognized international sources. It presents the most
current and accurate global development data available, and includes
national, regional and global estimates. To allow comparisons between
countries, prevalence rates are standardized to the WHO Standard
Population. According to World Bank’s Metadata description, a
statistical model based on a Bayesian negative binomial meta-regression
is used to model the prevalence of current tobacco use for each country,
separately for men and women between 1960 and 2023. However, due to a
lack of time series between 1960 and 1999, in this paper, I chose to
observe how the trend fluctuates from 2000 to 2023. It contains 6118
samples, including 266 countries and continents and 23 reported rates
yearly.
In comparison with EU countries, the criteria chosen here
are GDP per capita (current US$) and gender ratio. GDP per capita is the
gross domestic product divided by the midyear population, which is often
used to measure economic prosperity. It is found that richer countries
tend to smoke more (Ritchie and Roser (2023)). By examining the GDP per
capita of each country in Europe, it is possible to rank countries based
on wealth and then compare relative ones to the UK in order to find out
the relationship between GDP per capita and the smoking rate. The data
source is generated from from World Bank Open Data. [3]
Similarly, it is believed that having an equivalent number of males and
females should lead to corresponding percentage of smokers. Hence, data
on gender ratio in Europe is investigated to compare with one in the UK.
Data sources are collected from Eurostat
and Statisticstimes.
In the Forecast part, Time series analysis and Forecast model with ARIMA
will be conducted. Time series analysis accounts for the fact that data
points taken over time may have an internal structure (such as
autocorrelation, trend or seasonal variation) that should be accounted
for (“6.4. Introduction to
Time Series Analysis”
n.d.). The key idea of time series is that a time series is
stationary, meaning that the values always tend to vary about the same
level and that their variability is constant over time (M.Charlton and
A.Caimo, p.3, (2012)). Therefore, an autocorrelation test, a partial
correlation test and an Augmented Dickey-Fuller Test will be conducted
to examine whether the data is stationary. Thus, a forecast model with
ARIMA will be used. ARIMA(p.d,q), where p stands for the number of
autoregressive terms, d is the number of non seasonal differences
integrated and q describes the number of lagged forecast errors in the
equation (moving average). This model takes into account both trend and
autocorrelation to produce more accurate forecasts (Hyndman &
Athanasoppulos (2021) Chap.9) (Forecasting:
Principles and Practice (3rd Ed)
n.d.) [5].
Finally, the model that minimizes the Akaike
information criterion (AIC) is chosen (Hyndman & Athanasoppulos
(2021) Chap. 8.6) (Forecasting:
Principles and Practice (3rd Ed)
n.d.), with the confidence interval chosen being 95%.
2.3 Limitation and exceptions
The limited availability of data on health status is a major
constraint in assessing the health situation in developing countries.
When a country has fewer than two nationally representative
population-based surveys in different years, no attempt is made to fill
data gaps and no estimates are calculated. To fill data gaps,
information is “borrowed” from countries in the same UN subregion.
There are limited criteria being used to compare UK with EU countries,
otherwise, several prosperous factors that influence the prevalent of
smoking, such as education rate, socio status, income, psychological
problems, etc., have not been considered. Moreover, GDP is easily
influenced by population—as a rule, the bigger the population, the
higher the GDP—and can be artificially inflated in tax-haven countries,
that’s why the data should be treated with such caution.
Within the
Forecast part, there are limited observations. Chatfield (1996) (“6.4.4.5. Box-Jenkins
Models” n.d.) recommends at least 50
observations . Many others would recommend at least 100 observations. In
the scope of this paper, the observed data of smoking rate is 23
observations. Furthermore, other methods to forecast should also be
considered, i.e. Exponential Smoothing, to compare their fits and
release the best one.
3 Analysis
3.1 Prevalence of smoking in the World vs UK
With 1.1 billion smokers and 200 million additional users of other
tobacco products worldwide, smoking continues to pose a hazard to public
health. According to WHO Global Report , the total number of tobacco
users worldwide has decreased over the last 20 years, from 1.397 billion
in 2000 to 1.3 billion in 2020, or by around 100 million individuals.
Male tobacco consumption increased by around 12.6% throughout that time,
from 49.3% in 2000 to 36.7% in 2020. Positively, however, the new data
indicates that the number of men who smoke has ceased increasing and is
expected to drop by more than 2% between 2020 and 2025.
Figure 1
shows the map of change in the proportion of people who use tobacco from
2000 to 2023. It gives us an overview of the differences the world has
made. A decline is shown as an effort of fighting with tobacco epidemic.
The highest smoking rate is in the South-East Asia Region with a range
of 66.9% in 2000 and the lowest part is in the African Region with 3.5%
of the population.
Figure 2: Trends in current tobacco use by constinents
The Region of the Americas is predicted to have the fastest fall,
with an average relative reduction of 34% between 2000 (50.4%) and 2025
(25.7%). Both the South-East Asia Region and the African Region are
expected to see an average decline of about 32%. The pace of decline in
the European Region is very gradual; it is currently projected to
decrease by 19% relative between 2000 and 2025. The Eastern
Mediterranean Region is now forecasting a 22% relative decline by 2025,
which is slower than the global average reduction rate of 24%. The
Western Pacific Region is experiencing the slowest fall of any WHO
region; over the same period, an average relative reduction of barely 8%
is anticipated.
Following up on the decreasing pattern in the world,
the UK has also shown a negative trend in the number of smokers. Figure
3 and Figure 4 have been used to represent the trend. The data is based
on estimates from the World Bank and the Annual
Population Survey (APS).
In the map, we clearly see that there is a striking reduction between
2000 and 2020, falling from 40% of the population in 2000, which equates
to 24 million people, to 15.4% in 2020 (11 million people). After 20
years, the UK recorded the lowest percentage of smoking ever.
In the
graph, the proportion of smoking rates in persons, males and females
shared a similar pattern. In total, the number of smokers fell from
nearly half of the population in 1974 to about 14% in 2022, which
equates to around 7.1 million people in the population. This is the
lowest proportion of current smokers since records started in 1974.
The trend in men smokers decreased more sharply than in women. The
number of male smokers has reduced from 51% in 1974 to 14.2% in 2022.
For women it was 41% in 1974, compared with 12.4% in 2022.
There is
a slight average decrease of 0.5% during the Covid-19 period
(2020–2022). It is not mentioned that the rate of male smokers has an
upward trend, which is three times higher than the average decrease
previously mentioned. During this pandemic, people are also turning to
cigarettes as a stress-reduction strategy.
3.2 Prevalence of smoking in the EU vs UK
As a hypothesis was mentioned in the Methodology section, richer
countries tend to have a higher number of smokers. Among the 27 nations
in the European Union, 4 countries- Germany, Belgium, Finland and France
were chosen and their smoking percentage is described in Table
1.
| Countries | GDP per capita (US$) | Smoking rate (%) | ||
|---|---|---|---|---|
| France | 40886.3 | 34.6 | ||
| United Kingdom | 46125.3 | 14.2 | ||
| Germany | 48718.0 | 21.3 | ||
| Belgium | 49926.8 | 26.7 | ||
| Finland | 50871.9 | 22.3 |
Figure 5 combines the correspondingly equivalent countries with the
UK in GDP per capita on the scale of +/- 5000 (current US$) and smoking
rate (%).
An interesting result shows that the UK has the lowest rate with
14.2%, far lower than the others. Moreover, France has the lowest GDP
per capita but the highest smoking rate, almost 3 times higher than the
UK. An increasing relationship has been seen between smoking rate and
GDP per capita, except for Finland, which is the richest country on the
list but only the third highest country in smoking prevalence, with
22.3%.
With this result, the hypothesis mentioned above is not
significant.
A second hypothesis states that countries with similar
gender ratios should have a similar percentage of smoking prevalence.
The following Figure 6 takes a closer look at the proportion of smokers
in the European Union from the perspective of a similar gender ratio
with the UK In general, there are more women than men in both UK and EU
(Table 2), yet, men smoke more than women do. Still, the lowest
percentage of smoking belongs to the UK with 14.2% in total, 16.1% men
and 12.4% women smoking respectively. With a completely similar sex
ratio with the UK (102.4), Ireland is higher than UK in both gender. A
striking observation is that Belgium has the highest smoking prevalence,
nearly double that of the UK, despite sharing an approximate sex ratio
(102.6). Germany has a quite high percentage of women who smoke,
however, this can be affected by its highest gender ratio compared with
others (103.0).
Under this case, the second hypothesis also does not
show a significant value.
| Countries | Gender ratio | Smoking rate in Persons (%) | Males | Females | |
|---|---|---|---|---|---|
| United Kingdom | 102.4 | 14.2 | 16.1 | 12.4 | |
| Belgium | 102.6 | 26.7 | 29.0 | 24.5 | |
| Germany | 103.0 | 21.3 | 23.2 | 19.3 | |
| Finland | 102.1 | 22.3 | 26.1 | 18.5 | |
| Ireland | 102.4 | 19.3 | 21.5 | 17.0 |
Many reasons can be drawn from this, however, in terms of smoking
bans, there are some differences in the enactment of laws among these
countries.
In France, starting in 2007, smoking is now not allowed
in public places, objectively prevent second-hand smoking. However,
French smokers are often allowed to smoke indoors (Global Affairs
Explained, 2022). This lax enforcement of smoking laws in France makes
it easier for all people to smoke. It removes the need to go outside,
often in cold weather, to smoke. It also increases social smoking and
makes it harder for people who already use tobacco to quit. New tougher
measures are introduced by January 2024, including the price of a
tobacco pack will rise by €0.40 to €0.50, but it seems like a 10 cent
increase does not make an outstanding difference (Rodriguez n.d.).
Finland has had a long
history of smoking restrictions since 1975. Many strong legislations
came into force, heading to completely abolish smoking by 2030. The
collaboration between the health authorities of Finland,
non-governmental organizations and intensive health promotion has
contributed to the successful tobacco policy, whereas in France the
power is held by local authorities until 2024. For instance, Finnish
municipalities such as Nokia, Ylöjärvi and Espoo have banned smoking at
their campuses and some companies have offered cash bonuses of up to
€1,000 to employees who quit smoking (Nordal
2010).
In Belgium, an Inter-federal Strategy for a
Tobacco-free Generation was introduced in 2022, setting ambitious
targets to reduce daily smoking. There is greater progress in reducing
smoking among young people, but targets for daily smoking will not be
achieved with the lack of new policies (admin-sciensano 2023). A new and stricter
smoking ban in certain public places will be extended from 2025 (Times n.d.), so studies on Belgium’s smoking
rate can be further conducted.
In both scenarios, the UK holds the
first place as the country with the lowest smoking rate, which will be
discussed in the following chapter.
3.3 Smoking ban and its effects in UK
Smoking is considered the single biggest cause of preventable illness
and death, causing about 80,000 deaths per year across the UK. This
chapter aims to show the impact of laws on smoking prevalence in the
UK.
Looking at Figure 7 gives the impression that a downward trend occurs
in all age-groups starting from 18 years old among the four constituent
countries, namely England, Scotland, Northern Ireland and Wales.
Overall, 2022 witnessed the lowest proportion of current smokers since
records began in 2011. The 25-34 age group had the highest proportion of
smokers, while the lowest was among those aged 65 and over. The most
fluctuation in the percentage of current smokers was in Northern
Ireland, whereas England had a gradual drop. Interestingly, smoking
percentages in all of the age groups in Scotland decreased, among those
the 18-24 group saw a noticeable drop in 2017, which was 16.3%. Another
observation in the age group of 25-34 years old shows that there was a
slight surge in all mentioned countries during 2020-2022, which is also
considered the pandemic time period. It is believed that people turn to
cigarettes in response to stress (Robert, BBC News, 2021). There was a
significant different in number of smokers in 2011 among 4 constituent
countries, yet, in 2022, they shared a quite similarities, which was
approximately 14% on average.
Summarily, it can be discussed here
that there is an influence that should be strong enough to be a game
changer, forcing teenagers, adults, the elderly to give up smoking, thus
reducing the smoking rates in these four nations.
The following
Figure 8 shows the number of laws on smoking from 2011 to 2022. In
general, every reported country increased the number of laws through
years. Scotland has the most smoking laws, while Wales has the least,
with 48 and 36 in total, respectively. The number of smoking laws in
England increased the most, from 28 laws in 2011 to 44 ones in 2022,
compared to Northern Ireland having the least updated laws, with 12 more
during 12 years. According to Tobacco Control Laws,
England and Wales were the first countries to have smoking laws come
into effect (1933), others were Northern Ireland (1978) and Scotland
(1980). The number of laws in these countries usually stays unchanged
every 3 years. The 2020 – 2022 period saw the most number of laws in all
4 countries compared to previous years. This result possibly links to
the increasing smoking rate during this time, which was shown in Figure
7, meaning that more laws came to enforcement to deal with this upward
trend.
To explain clearer the correlation between the number of laws and the
prevalence of smoking in individual countries clearer, a scatter plot
containing a regression line has been conducted. The database is
generated from Office
for National Statistics.
First, Figure 9 shows that in 2022, England had the lowest proportion
of current smokers with 12.7%, which equates to around 5.3 million
people. This is a statistically significant decrease in the proportion
of smokers from 2021, at 13.0% (around 5.4 million people).
For the
other constituent countries of the UK, Wales had the highest proportion
of current smokers (14.1%, around 341,000 people). In Northern Ireland
and Scotland, the proportion of current smokers was 14.0% (around
199,000 people) and 13.9% (around 587,000 people), respectively. Since
2011, there has been a statistically significant decline in the
proportion of current smokers in England, Scotland, Wales and Northern
Ireland. For Northern Ireland, the estimate over time has been more
variable because of the smaller sample size.
Taking a closer look at
Figure 9, it is seen that the regression line does not go straight down
strictly, but clearly has a downward trend in all four countries. This
would mean that the more laws there are, the lower the smoking rate it
is. England gradually decreased in terms of smoking rate as it has the
most laws increasing gradually by years. Scotland is the latest country
to apply laws, which could be the reason for the highest smoking rate in
2011. However, as soon as it enforced more laws, a decrease was clearly
to be seen. Wales and Northern Ireland shared a similar pattern in the
number of laws, but Wales had a longer history of making laws than
Northern Ireland, which can be a good explanation for the straight
downward of the smoking rate in Wales occurring compared to Northern
Ireland.
Overall, it is possible to determine that more laws help
reduce the percentage of smokers in each constituent country of Great
Britain. A more in-depth analysis would be needed here to find out
whether the stricter laws adopted will be effective or
counterproductive, as measured by the behaviors of citizens. For
example, if a new law is phased out, 14-year-olds today will never
legally be able to buy a cigarette . If the ban is passed into law in
England, the country will have some of the strictest smoking laws in the
world (Franks, Sky News (2024)).
4 Forecast
Considering the data analyzed above, a forecast of future development
follows, finding out whether a trend of decreasing smoking rates
continues, assuming other factors are constant. In this section, two
methods are applied, including Time series regression model (TSRM), thus
forecasting with ARIMA model. Both of these methods were described in
the Methodology part.
4.1 Time series regression model (TSRM)
First of all, Figure 10 performs the smoking rate in total in UK
changing through a period of time from 2000 to 2022 using TSRM with a
frequency = 1. At a first glance, there is a noticeable reduction in the
observed data. This is also supported by the regression line. Looking
closer gives us the information that the highest and the lowest points
often occur once every 5 years.
## NULL
Following the figure are three tests being conducted, namely the acf,
pacf and adf tests. The function acf estimates the
autocovariance or autocorrelation function (Venables, W. N. and Ripley,
B. D. (2002)). If the lines cross the blue line, it means that the
examined data has high autocorrelation. In this case, when lag is
between 0 and 4, the lines cross the blue line, meaning that the data
has a correlation itself. Therefore, the data is not stationary.
The function pacf is the function used for the partial
autocorrelations. It is estimated by fitting autoregressive models of
successively higher orders up to lag.max (Ripley, B. D. (2002)). In this
case, the partial correlation is not significant.
Finally, the function adf.test is again used to test whether
the data is stationary or not using a p-value of 0.05. In this case,
p-value is 0.5078, which is larger than 0.05, meaning that the data is
not stationary.
##
## Augmented Dickey-Fuller Test
##
## data: smoking_uk_fc_time
## Dickey-Fuller = -2.1695, Lag order = 2, p-value = 0.5078
## alternative hypothesis: stationaryThese results above require changing the data to stationary form
using the auto.arima function, which returns the best ARIMA model
according to either the AIC, AICc or BIC value (Hyndman, RJ and
Khandakar, Y (2008)) and taking the acf, and pacf test again for final
confirmation.
4.2 Forecast model with ARIMA
By now, as the data is stationary, it is reliable to perform a
Forecast model with ARIMA. Here, I would like to see the trend of the
smoking rate and whether it continues to reduce in the next 10 years.
The observed values are good within the confident level = 95%. Thus, we
continue to see a decreasing trend with few fluctuations, predicting
that the smoking rate will continue to reduce, given that the historical
data is shown (Figure 11).
## NULL5 Conclusion
In conclusion, the tobacco epidemic is one of the biggest opponents
to not only the UK but also the world . Although long-term progress to
reduce the smoking rate has been made, citizens globally are still
facing preventable deaths caused by tobacco use.
This study aims to analyze the trend of tobacco use/smoking prevalence
in the UK in the present and in the future while applying a considerable
number of laws, thus, giving appropriate recommendations. It is shown
that the UK is outstanding as it is one of the countries in the world
decreasing the fastest in percentage of people smoking. Also, when
compared with other European Union countries having similar backgrounds,
the UK gives the best impression as the country with the lowest smoking
rate. Clearly, the proportion of smokers in the four constituent
countries of Great Britain, namely England, Scotland, Wales and Northern
Ireland follows a downward trend in all age groups from 18 years old to
65 and over, especially the young adult groups. A forecast based on past
data also indicates a continued decrease in smoking statistics until
2032.
One contributor to these positive results in the UK is the
number of smoking laws. The more laws there are, the lower the smoking
rate. The UK Government should increasingly put a comprehensive set of
bans and regulations in full enforcement and implemented them in full to
be a smoke-free nation by 2030, meaning only 5% of the population would
smoke by then (Khan, 2022).
This study only focuses on the effects
of laws on smoking rates in the UK. A deeper analysis should be
conducted, concerning various factors, i.e. income, socio-economics,
ethnicity, etc. Examining these and other factors can provide a better
understanding of the smokers’ perspectives, helping to fulfill the
smoking-free achievement of this nation.