Saving Behaviour, Expectations and Future Financial Hardship
1. Sarah Brown and Karl Taylor
University of Sheffield, IZA
April 2017
2. Since 2008, concern amongst policy-makers
over household financial vulnerability.
Many households hold low levels of savings to
fall back on during times of financial adversity.
Issues for both the short-term and the long-
term.
In the UK, the household saving rate has halved
since the middle of 2010 from 11.5% to 5.8% in
the fourth quarter of 2015 (ONS, 2016).
3. A commonly held view is that individuals are
not saving enough.
Garon (2012) comments that, in the U.S., ‘it
has become painfully clear that millions lack
the savings to protect themselves against
foreclosures, unemployment, medical
emergencies, and impoverished retirements.’
UK policies such as automatic pension
enrolment and schemes like ‘Help to Save’
announced in the 2016 budget, are designed
to encourage saving.
4. Evidence from the Money Advice Service
indicates that 4/10 working-age individuals
in the UK <£100 in savings at a given point in
time.
In NI, WM, Yorkshire and Humberside, NE and
Wales, more than half the adult population
has savings below that level.
The research also showed that some people
on low incomes do save money.
Roughly 1/4 adults with household incomes
<£13,500 have more than £1,000 in savings.
5. Saving is clearly constrained by income and
financial commitments, including debt servicing.
Understanding the saving behaviour of
individuals and, in particular, that of the low-
paid, is important from a societal perspective.
It is important to understand the consequences
of a lack of saving for future financial wellbeing,
as a lack of savings may, for example, lead to
debt accumulation if households are forced to
borrow to deal with unforeseen events.
6. Investigate the determinants of saving and
private pension contributions over two
decades:
◦ Static and dynamic modelling;
◦ Incidence and amounts;
◦ Role of financial expectations; and
◦ Role of low pay (limited attention so far)
Precautionary saving: households hold a contingency
fund in case of adverse future events, Keynes (1936).
7. Consider whether saving or employee pension
contributions are a buffer against future financial
hardship.
Gjertson (2016) presents evidence, based on a
small non-representative sample of low-paid US
households, supporting a protective role for
small amounts of saving for future financial
hardship.
◦ RE and FE count models;
◦ Types of financial problem incurred.
8. The BHPS is a random sample survey, carried out by the
Institute for Social and Economic Research, of each adult
member from a nationally representative sample of more
than 5,000 private households (yielding approximately
10,000 individual interviews).
Participants live in Scotland, Wales, Northern Ireland and
England and the BHPS covered the period 1991 to 2008.
Understanding Society replaced the BHPS after 2008.
In wave 1 of Understanding Society, over 50,000 individuals
were interviewed between 2009 and 2011, correspondingly
in wave 4 (last one we use) over 47,000 individuals were
interviewed between 2012 and 2014.
Both surveys contain information about people’s social and
economic circumstances, attitudes, behaviours and health.
9. Our sample is aged from 25 to 59;
Focus upon employed individuals only;
NT = 85,994; N = 14,071 (individuals);
T = 1991 to 2013/14.
Information on:
◦ Savings (including amount);
◦ Employee private pension contributions (including
amount);
◦ Financial expectations;
◦ Detailed socio-economic characteristics.
10. Our measure of monthly saving is based on
the responses to the following question:
“Do you save any amount of your income, for
example, by putting something away now and
then in a bank, building society, or Post
Office account other than to meet regular
bills? About how much, on average, do you
manage to save a month?”
11. Information is also available on monthly
private pension contributions:
“Other than your main employer or
occupational pension scheme, are you
currently a member of any personal pension
scheme or do you currently contribute to any
personal pension scheme? Please include any
Additional Voluntary Contribution scheme
you may belong to.”
A monthly amount of private pension
contributions is calculated.
12. 102030405060
%
1990 1995 2000 2005 2010
year
% save % pension
% save and pension % no save or pension
Types of saving
13. 2468
%
1990 1995 2000 2005 2010
year
save/income pension/income
(save+pension)/income
Saving as a proportion of income
14. Looking ahead, how do you think you will be
financially a year from now, will you be:
Worse off
About the same
Better off (i.e. financially optimistic)
Used in previous literature on household
finances: Brown et al. (2005, 2008): more
financially optimistic take on more debt.
15. Low pay is defined at the individual level
(below 2/3 median gross weekly pay from
ASHE).
The Annual Survey of Hours and Earnings
(ASHE) is the most comprehensive source of
earnings information on the structure and
distribution of earnings in the UK.
ASHE is based on a 1% sample of employee
jobs taken from HM Revenue & Customs
(HMRC) PAYE records.
19. , binary variable which indicates whether
the individual saves, denotes the individual
and denotes time.
Dynamic model (Wooldridge, 2005):
20. Control variables in :
Indicator of being in low pay at t-1;
Indicator of not being financially optimistic;
Gender; Marital Status; Age; Education
(degree, A levels, GCSE, other qualification);
Number of children in household; Number of
adults in household; Housing tenure (own
home outright, own home with mortgage,
rent); Self-reported health status; Region;
Month of interview; Years.
21. Random Effects Probit Model: Probability of saving on a
monthly basis
Key
Variables
M.E. T-stat M.E. T-stat
Not
optimistic
0.0277 6.78 0.0263 5.79
Low paidt-1 -0.0440 6.51 -0.0303 4.21
Savedt-1 - - 0.2176 50.75
23. Random Effects Probit Model: Probability of saving on a
monthly basis
Key
Variables
M.E. T-stat M.E. T-stat
Not
optimistic
0.0268 6.44 0.0253 5.59
Y < 25th -0.1350 13.25 -0.1276 11.49
Y 25th to 50th -0.0809 10.37 -0.0757 8.89
Y 50th to 75th -0.0424 7.12 -0.0387 5.96
Savedt-1 - - 0.2155 50.53
24. Random Effects Probit Model: Probability of making a
monthly private pension contribution
Key
Variables
M.E. T-stat M.E. T-stat
Not
optimistic
-0.0079 2.88 -0.0075 2.77
Low paidt-1 -0.0125 2.70 -0.0124 2.37
Pensiont-1 - - 0.1161 39.82
32. From 1996, 8 types of financial hardship over
time:
◦ paying for their accommodation;
◦ loan repayment;
◦ keeping their home adequately warm;
◦ been able to pay for a week’s annual holiday;
◦ replace worn-out furniture;
◦ been able to buy new, rather than second-hand,
clothes;
◦ been able to eat meat, chicken, fish every second
day; and
◦ been able to have friends or family for a drink or
meal at least once a month.
33. Person in household responsible for
financial decisions.
Typically head of household, .
Aged from 25 to 59;
Employed heads of household;
NT = 34,496; N = 8,285 (households);
T = 1996 to 2013/14.
34.
35.
36.
37. is number of financial problems.
=1 if saved in t-1, 0=otherwise.
=1 if low paid in t, 0=otherwise.
Additional control variables in :
Gender; Marital Status; Age; Education (degree, A
levels, GCSE, other qualification); labour income;
non-labour income; Number of children in
household; Number of adults in household;
Housing tenure (own home outright, own home
with mortgage, rent); Self-reported health status;
Region; Month of interview; Years.
38. Count Model: Number of Financial Problems
INCLUDE INCOME – WHETHER SAVE
Fixed Effects Random Effects
Key Variables COEF T-stat COEF T-stat
Savingt-1 -0.1210 3.67 -0.3030 10.67
Low paid 0.0128 0.26 0.0368 0.84
39. Count Model: Number of Financial Problems
INCLUDE INCOME – WHETHER PENSION
Fixed Effects Random Effects
Key Variables COEF T-stat COEF T-stat
Pensiont-1 0.0286 0.53 -0.0762 1.68
Low paid 0.0143 0.28 0.0378 0.86
40. Fixed Effects Count Model: Number of Financial Problems
EXCLUDE INCOME – WHETHER SAVE
Fixed Effects Random Effects
Key Variables COEF T-stat COEF T-stat
Savingt-1 -0.1228 3.72 -0.3497 12.28
Low paid 0.0739 2.57 0.2203 5.58
41. Fixed Effects Count Model: Number of Financial Problems
EXCLUDE INCOME – WHETHER PENSION
Fixed Effects Random Effects
Key Variables COEF T-stat COEF T-stat
Pensiont-1 0.0280 0.52 -0.1012 2.22
Low paid 0.0760 2.61 0.2309 5.81
42. Fixed Effects Count Model: Number of Financial Problems
INCLUDE INCOME – LOG AMOUNT SAVED
Fixed Effects Random Effects
Key Variables COEF T-stat COEF T-stat
Log Savingt-1 -0.0347 4.72 -0.0817 13.36
Low paid 0.0133 0.26 0.0409 0.94
43. Fixed Effects Count Model: Number of Financial Problems
INCLUDE INCOME – LOG AMOUNT
PENSION
Fixed Effects Random Effects
Key Variables COEF T-stat COEF T-stat
Log Pensiont-1 0.0056 0.44 -0.0215 2.05
Low paid 0.0142 0.28 0.0381 0.87
51. Effect of saving on financial hardship varies
by type of problem faced;
Same pattern of results as above found for
affording holidays and new furniture;
No effects from saving or pensions found for
the probability of reporting problems with:
keeping home warm; new clothes;
Very small protective effect for saving is
found for ability to afford meat; entertain
friends/family.
52. Our findings suggest that financial pessimism
is positively associated with active saving and
negatively associated with private pension
contributions;
Being in low pay is inversely associated with
both types of saving;
Evidence of persistence in saving behaviour;
Evidence supports a protective role for saving
with lagged saving being inversely associated
with current financial problems.
53. Five main areas to develop:
Future research will explore the effects of
subjective measures of job (in)security;
Accuracy of financial expectations;
Eligibility and membership of occupational
pension schemes;
Dynamic zero-inflated Bayesian models of
financial problems;
Interaction of saving behaviour between
household members;
54. Employed couples.
Unit of observation head of household i.e. person
responsible for making financial decisions.
NT = 21,876; N = 5,681
Controls as previous, , but now condition on:
◦ Whether head in low pay;
◦ Whether spouse in low pay;
◦ Whether both head and spouse in low pay;
◦ Saving of head and spouse – amounts and whether save
55. Count Model – Poisson Number of Problems
Whether saved by
head and/or
spouse t-1
Amount saved by
head and/or
spouse t-1
Key Variables M.E. T-stat M.E. T-stat
Savingt-1 -0.3095 7.99 -0.0879 11.19
Low paid head 0.1990 1.82 0.1925 1.77
Low paid spouse 0.1750 3.70 0.1548 3.29
Both low paid 0.0786 0.62 0.0911 0.72
56. Count Model – Poisson Number of Problems
Whether saved by
head and/or
spouse t-1
Amount saved by
head and/or
spouse t-1
Key Variables M.E. T-stat M.E. T-stat
Savingt-1 head -0.2167 5.17 -0.0020 9.75
Savingt-1 spouse -0.2233 4.93 -0.0024 2.27
Low paid head 0.1686 1.49 0.1842 1.69
Low paid spouse 0.1482 3.00 0.1815 3.86
Both low paid 0.1180 0.91 0.0873 0.69