Transcript First estimates of UK National Transfer Accounts

First estimates of UK National Transfer
Accounts
David McCarthy & James Sefton
13th April 2011
ICBS Research Fest 2011
1
© Imperial College Business School
Outline
•
•
•
•
•
The ageing society & the economic lifecycle
National Transfer Accounts (NTA’s)
NTA estimates for the UK
Cross-country comparisons
Applications
• Are we life-cycle savers?
• How much do children cost?
• The response to public transfers to the elderly: lower savings, or higher
bequests?
• Consumption shocks and assets prices (incl residential housing)
The ageing society
• Societies all around the world are ageing
• The fundamental cause is a demographic transition from high fertility –
high mortality societies to low fertility – low mortality societies
• (Primary cause of population ageing is low fertility)
Demographic transition
• In the future, the increase in the proportion of the population which is
elderly will dominate population change
• 2008: 13.6% > 60
• 2033: 20.6% > 60
Economic lifecycle
• Consumption exceeds labour income for the very young and the very old
• Changes in population structure affect everyone
30,000
Per capita consumption and labour income, UK (2007)
25,000
£ p.p. p.a.
20,000
15,000
10,000
5,000
0
0
10
20
30
Consumption
40
50
Age
Labour income
60
70
80
Population ageing presents major challenges
• How sustainable are our transfer systems?
• Public sector: Number of taxpayers will decline relative to number of
beneficiaries
• Private sector: Number of working-age adults will decline relative to
dependent children and elderly
• How resilient are our asset markets (housing / financial assets)?
• If the elderly depend on selling assets to fund their retirement, how will
asset prices respond to population ageing?
• How sustainable is our political consensus?
• What is the implication of population ageing for generational equity?
Economic lifecycle: UK
• Consumption exceeds labour income below age 23 and above age 57
30,000
Per capita consumption and labour income, UK (2007)
25,000
£ p.p. p.a.
20,000
15,000
Age 57
10,000
Age 23
5,000
0
0
10
20
30
Consumption
40
50
Age
Labour income
60
70
80
Methodology and data sources
• Basic principle is to use a nationally representative survey to estimate
age profiles for the variable of interest
• Smooth the profile using a Gaussian smoother
• Use the national population to set the aggregate value implied by the
smoothed profile equal to the appropriate control value from the
National Accounts
• We estimated 27 separate profiles (phew!) and used three surveys for
our data:
• Food and Expenditure Survey, 2006
• Family Resources Survey, 2006
• Wealth and Asset Survey, 2006/8
• We used National Accounts data for 2007 from the 2008 Blue Book
• (Data problems: students, care homes, age top-coding)
Numbers are consistent with National Accounts
30
Aggregate consumption and labour income, UK (2007)
25
£bn p.a.
20
15
10
5
0
10
20
30
40
50
60
70
80
Age
£193bn , around
14% of GDP
Consumption
Labour income
£217bn, around
15% of GDP
Components of consumption
30,000
Components of per capita consumption by age, UK (2007)
25,000
£ p.p. p.a.
20,000
15,000
10,000
5,000
0
0
10
20
30
40
50
60
70
80
90
Age
Public education
Other public
Actual Rentals
Public health
Private education
Imputed rentals
Public long-term care
Private health
Other private
• Consumption includes both consumption of services provided by
government (education, health care, defence etc), and private
consumption
Per-capita age reallocations, UK (2007)
40,000
Components of per capita age reallocations, UK (2007)
30,000
£ p.p. p.a.
20,000
10,000
0
0
10
20
30
40
50
60
70
80
90
-10,000
Age
-20,000
Private transfers
largely downward,
few upward transfers
Asset-based Reallocations
Bequests
Public Transfers
Public transfers upward
and downward
Private Transfers
ABR’s largely positive,
and significant
Asset-based reallocations and saving
20,000
Components of per capita asset-based reallocations, UK (2007)
15,000
10,000
£ p.p. p.a.
5,000
0
0
10
20
30
40
50
60
70
80
90
-5,000
-10,000
-15,000
Age
Private asset income
Most private
saving occurs
between the ages
of 43 and 65
Private saving (negative)
Public ABR's
Does appear to be real
dissaving at older ages
(although most may be
private pensions)
Asset income largely received after
the age of 40, declines with dissaving
Components of consumption, UK (2007)
Percentage of total consumption, synthetic cohorts, UK (2007)
mortality
Components of consumption, synthetic cohorts, UK (2007)
150%
125%
100%
75%
50%
25%
0%
Age 0-19
Age 20-64
Age 65+
-25%
-50%
Asset-based reallocations
Private transfers
Labour income
Public transfers
Bequests
Applications
•
•
•
•
International comparisons of the economic lifecycle
How much does it cost to raise a child?
Are we life-cycle savers?
Consumption and asset prices
• Residential housing
• Other assets
© Imperial College Business School
Comparison with other NTA countries
• NTA results are available on the NTA website for project members
(www.ntaccounts.org)
• This allows detailed comparisons of NTA quantities between NTA
countries
• Chose a group of ‘peer’ countries to the UK
COUNTRY
Austria
Germany
Hungary
Japan
Spain
Sweden
UK
US
YEAR
2000
2003
2005
2004
2000
2003
2007
2003
© Imperial College Business School
Age profiles of consumption
Propn of labour income (30-49)
1.6
All profiles scaled
by average labour
income between
30-49 in each
country
Consumption
1.4
1.2
1
0.8
0.6
0.4
0.2
0
0
10
20
30
60
40 Age 50
80
70
Austria
Germany
Hungary
Japan
Spain
Sweden
UK
US
90
Proportion of consumption at each age which is public
2
Public transfers
Propn of labour income (30-49)
1.5
Austria
1
Germany
Hungary
Japan
0.5
UK has average
consumption at
young ages but a
low proportion of
this is publicly
financed
Spain
Sweden
0
0
10
20
30
40
50
70
80
90
UK
US
-0.5
-1
60
Age
A much higher
proportion of the
consumption of
the elderly is
publicly financed
Components of 65+ consumption (synthetic cohorts)
Heavy reliance on
assets and labour
income to finance
old-age
consumption
Heavy reliance on
transfers
Transfers
Austria
Hungary
1/3
Sweden
2/3
Germany
Japan
2/3
Spain
1/3
UK
US
Assets
0
0
2/3
1/3
Labour
income
Components of 0-19 consumption (synthetic cohorts)
Much less
variation between
countries than
with old-age
consumption
Private Transfers
1/3
2/3
US
Japan
UK
Spain
Germany
Sweden
Austria
Hungary
2/3
UK relies more
heavily on private
transfers than its
peer countries
appear to do,
including the US
1/3
Public
transfers
0
0
2/3
1/3
Labour
income
© Imperial College Business School
Applications: Are we life-cycle savers?
• Previous attempts to test the life-cycle model of consumption and saving
have suffered from the problem that they have not had a comprehensive
measure of wealth, including (private) transfer wealth
• NTA’s provide this measure of wealth (provided you are willing to make
some strong assumptions)
Total financial and transfer wealth, UK (2007)
700000
600000
500000
Wealth (pounds)
400000
300000
200000
100000
0
-100000
0
10
20
30
40
50
Age
-200000
-300000
Public transfers
Private transfers
Labour income
Financial wealth
Bequest
60
70
80
90
Some theory
• Consider a life-cycle model of consumption and savings, with
no bequest motives, no uncertainty, labour and transfer
income, and mortality:
w
max
{c0 ,...,cw }
å
r i p i u(ci ) s.t. wi+ 1 = (wi - ci + t i + yi )(1+ r )
i= 0
• Model solution given by:
w
å
wi +
ci =
w
y j (1 + r )
j= i
w
å
j= i
((r (1 + r ))
i- j
+
å
t j (1 + r )i-
j= i
j- i p j 1/ g
pi
) (1 + r )i-
j
j
Average propensity to consume out of wealth
• Average propensity to consume out of total wealth at age i is defined as:
ci
APCi =
w
w
i- j
wi + å y j (1 + r ) + å t j (1 + r )i- j
j= i
j= i
• Which using the solution to the model can be rewritten as:
APCi =
1
w
å
((r (1 + r )) j- i
p j 1/ g
pi
) (1 + r )i-
j
j= i
» r-
1
g
( r - (1- r ) - (1-
p i+ 1
pi
)) ,
• (Which in turn allows us to estimate rho by age, provided we know r,
gamma and the mortality probabilities)
© Imperial College Business School
NTA-derived APC, by age (UK, 2007)
• Using NTA-derived estimates of total wealth at each age, and NTA
consumption profiles, we can obtain the APC at each age
Average propensity to consume out of total wealth,
by age, UK (2007)
Average propensity to consume out of total wealth
0.15
0.1
APC
0.05
0
0
10
20
30
40
50
60
70
Age
• Estimates compare well with Sefton & Kirsanova (2006)
80
Estimated subjective discount factors, by age, UK (2007)
Estimated subjective discount factors, by age, UK (2007)
Subjective discount factor
0.2
0.15
0.1
0.05
0
0
10
20
-0.05
30
40
50
60
70
80
Age
Gamma = 1
Gamma = 2
Gamma = 3
(Slope depends heavily on assumed level of risk aversion)
Applications: How much does it cost to raise a child?
• There is a huge literature on the determinants of fertility, parental
investment in children, and the role of bequests, going back to Becker,
1974.
• NTA’s allow at least reasonable estimates of the cost – public and private
– of raising a child
• In the UK, PV of private transfers to a child between 0 and 19 are
£140,000, public transfers are £57,000 (around 8.6 years of labour
income on average)
0-19
Austria
Germany
Hungary
Japan
Spain
Sweden
UK
US
Labour Private
Public
income transfers transfers
1.1
3.4
3.3
0.3
4.3
2.9
0.1
3.0
4.6
0.2
5.0
4.4
0.4
5.1
3.0
0.5
4.4
3.8
0.8
5.5
2.3
0.3
4.7
3.5
ABR's
0.2
0.1
0.1
0.1
-0.1
0.1
-0.1
0.1
TOTAL
8.1
7.7
7.8
9.7
8.3
8.7
8.6
8.6
Applications: Response to higher old-age transfers
• Two schools of thought in economics on the response of individuals to
higher transfers to the elderly
• Feldstein (1974): individuals save less in anticipation of receiving larger
transfers
• Barro (1974): individuals may bequeath more in response to larger
transfers having been received (undoing the effect of the transfers)
Components of 65+ transfer mechanisms
Since private
transfers are
small, there is
almost a one-forone trade-off
between ABR’s
and public
transfers
Sweden
Public Transfers
Hungary
Austria
1/3
Germany
2/3
Spain
Japan
UK
2/3
US
1/3
Assets
0
0
2/3
1/3
Private
transfers
But do Swedes/Hungarians/Austrians save less or bequeath more?
Saving appears high in high-transfer countries.........
0.5
Private saving
0.4
Propn of labour income (30-49)
Very high savings
as a proportion of
average income
30-49 in the UK
between ages 50
and 60. Is this
result confirmed
in any other
work?
0.3
Austria
Germany
0.2
Hungary
Japan
0.1
Spain
0
Sweden
0
10
20
30
40
50
60
-0.1
70
80
90
UK
US
-0.2
-0.3
Age
… suggesting that bequests must be higher there too?
Total savings (65+) vs. Public transfers / LCD
• Absolute level of savings appears to be higher in countries where a
higher proportion of the LCD is financed by public transfers
Total savings, propn YL30-49, synth
cohort, 65+
5
Savings and public transfers, NTA countries, various years
4
3
2
1
0
0
20
40
60
80
100
120
-1
-2
Propn of LCD financed by public transfers
• What about bequests?
© Imperial College Business School
Application: Consumption and asset prices
• Serious work will only be possible once a full set of NTA estimates has
been completed
• Finance theory suggests a strong link between consumption and assets
prices, but empirical relationships are weak
• What role do transfers (and the sensitivity of transfers to economic
shocks) play in explaining the weak empirical findings in the asset-pricing
literature?
© Imperial College Business School
Application: Consumption and house prices
• Estimated NTA’s separately for owner-occupiers and renters in 2007
• Very different structure of transfers across the lifecycle
40,000
40,000
Components of per capita age reallocations, renters, UK (2007)
20,000
20,000
£ p.p. p.a.
30,000
£ p.p. p.a.
30,000
Components of per capita age reallocations, owner-occupiers
UK (2007)
10,000
10,000
0
0
0
10
20
30
40
50
60
70
80
90
0
-10,000
10
20
30
50
60
70
80
90
-10,000
Age
-20,000
Asset-based Reallocations
20,000
Age
Bequests
Public Transfers
-20,000
Private Transfers
5,000
5,000
0
10
20
30
40
50
60
70
80
90
£ p.p. p.a.
10,000
-5,000
-10,000
-10,000
-15,000
-15,000
Private asset income
Private saving (negative)
0
10
20
30
-20,000
Age
Public ABR's
Public Transfers
Private Transfers
0
-5,000
-20,000
Bequests
Components of per capita asset-based reallocations, owneroccupiers, UK (2007)
15,000
10,000
0
Asset-based Reallocations
20,000
Components of per capita asset-based reallocations, renters, UK
(2007)
15,000
£ p.p. p.a.
40
40
50
60
70
Age
Private asset income
Private saving (negative)
Public ABR's
80
90
Conclusion & future work
• NTA’s for the UK appear to indicate that we transfer less to young
people than our peer countries (more to older people, especially the
very old) even controlling on the level of consumption
• UK public age-based reallocations appear quite low
• Our children are quite reliant on private transfers, with possible
implications for inter-generational transmission of inequality
• Our older people are quite reliant on asset-based reallocations relative
to other countries
• We illustrated some potential applications of NTA’s
• Currently assessing whether NTA methodology can shed light on
differential savings patterns within sub-groups of the UK population
© Imperial College Business School