China’s High Saving Rate: Analytics and Prospects...government transfer to poor households •...
Transcript of China’s High Saving Rate: Analytics and Prospects...government transfer to poor households •...
China’s High Saving Rate: Analytics and Prospects
May 18th, 20171
China has one of the world’s highest saving rate…
2
… which is the crux of external/internal imbalances
3
0
4
8
12
30
35
40
45
50
2006
2007
2008
2009
2010
2011
2012
2013
2014
2015
Sources: CEIC Data Company Ltd.; and IMF staff calculations.
Rotation from External to Internal Imbalance(In percent of GDP)
CA balance/GDP (RHS)
Investment/GDP
Excessive savings depress consumption and lower welfare
4
Questions
• What are the drivers of high national savings?
• What policies could reduce excessive savings?
• What is the likely trajectory of national savings in the medium-term?
5
Key Takeaways• High national savings today mostly emanate from the household
sector, though corporate sector was also a main driver in early 2000, partly reflecting under-valued exchange rate.
• Rising household savings mostly reflect the one-child policy induced demographic changes, and the dismantle of social safety net during economic transition. Housing reform and rising income inequality also contributed to the rise in savings.
• The high fiscal saving was driven by low expenditure, reflecting the stage of pension system and insufficient provision of public service, especially health and social security.
• Saving rate to fall substantially in the medium-term with demographic shift and policies to strengthening social safety net and reducing income inequality
6
Composition of national savings
7
-10
0
10
20
30
40
50
60
1992
1994
1996
1998
2000
2002
2004
2006
2008
2010
2012
2014
2016
government financialnon-financial firms householdsTotal(FF)
National Saving Rate (in percent of GDP)
Sources: Flow of funds, IMF staff estimates.
Global shift from household to corporate saving
8
-4
-2
0
2
4
6
8
10
12
14
16
HouseholdCorporateGovernment
Shifting Pattern of Global Savings( in percent of GDP)
Sources: Chen et al (2017), "The Global Rise of Corporate Saving." .
The imbalance today mostly reflects the household sector
9
0
5
10
15
20
25
30
1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013
China Global average
Household Saving(in percent of GDP)
0
5
10
15
20
25
30
1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013
Corporate Saving(in percent of GDP)
-5-3-113579
111315
1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013
Government Saving(in percent of GDP)
Sources: CEIC and Chen et al (2017)
Roadmap
• Household saving• Corporate saving• Government saving• The role of State and policy recommendations• Prospect of Saving
10
HOUSEHOLD SAVING
11
Household saving and reform overview
12
-0.05
0
0.05
0.1
0.15
0.2
0.25
0.3
0.35
1955
1958
1961
1964
1967
1970
1973
1976
1979
1982
1985
1988
1991
1994
1997
2000
2003
2006
2009
2012
2015
Household Saving Rate(in percent of disposable income)
Sources: Household Survey and Modigliani and Cao (2004)
One-child PolicyReform and opening-up
PoliticalInstability, growth declined
South Tourof Deng
SOE reformAsian Financial Crisis
WTO EntryGreat Famine
Drivers of Household Savings
• Demographics: Modigliani and Cao (2004), Choukhmane and others (2013), Curtis, Lugauer, and Mark (2015)
• Social safety net: Blanchard (2005), Prasad (2010), and He et al. (2017)
• Financial repression: Nabar(2011) • Housing: Li et al(2013)• Income inequality
13
One-child policy has led to rapid decline in youth-dependence ratio
14
10
30
50
70
90
110
0 10000 20000 30000 40000
Yout
h-ag
ed de
pend
ency
ratio
in 20
15
GDP per Capita in 2014, PPP based
Youth Dependence Ratio and GDP Per Capita
Sources: United Nations; and Penn World Table.
China
Household saving and demographics
• OLG Model (Curtis et al 2015)• Main features• 95 generations, 20-95 decide consumption, retirement
at 64• Children’s welfare in parents’ utility function as in Barro
and Becker (1989) – fewer children makes you effectively more patient
• Income transfers from working agents to retirees –fewer children means less old-age support
• Demographic composition effect – more people in prime savings age
15
Demographic changes account for half of increase in household saving
16
Household saving and social safety net
• Transition in the 80s and 90s from a central planned system to market economy has led to dismantle of social safety net:
Break of the “ iron rice bowl”, no more job security with large SOE layoffs (He et al. 2017).
Individual bears increasing share of healthcare and education expenditure
Housing reform has eliminated subsidized housing for employees provided by SOEs.
17
Historical health care cost allocation
18
0.0
10.0
20.0
30.0
40.0
50.0
60.0
1978
1980
1982
1984
1986
1988
1990
1992
1994
1996
1998
2000
2002
2004
2006
2008
2010
2012
GovernmentCorporateIndividual
Share in total health expenditure( in percent)
Sources: China Health Statistics Yearbook.
Empirical estimation
• Regression at prefecture level(287 prefectural-level municipalities, data available
since 2013 for rural households)
• Regression at provincial level( 31 provinces, data available from 1995, with structural break in 2013)
19
Pension spending most effective for urban households
20
Table 1. Urban Households: Consumption and Government Spending, Prefecture Level Cross-Sectional Analysis
2013 2014(1) (2) (1) (2)
Disposable Income 0.94 0.93 0.99 1.00
(Standard error) (0.04) (0.04) (0.05) (0.04)
[P-val] [0.00] [0.00] [0.00] [0.00]
Gov. Social Security Spending 0.07 0.06 0.08 0.08
(Standard error) (0.02) (0.02) (0.02) (0.02)
[P-val] [0.00] [0.00] [0.00] [0.00]
Gov. Health Spending 0.01 ... -0.07 ...
(Standard error) (0.04) ... (0.06) ...
[P-val] [0.85] ... [0.27] ...
Gov. Education Spending -0.02 ... 0.03 ...
(Standard error) (0.04) ... (0.05) ...
[P-val] [0.52] ... [0.50] ...
Constant 0.21 0.34 0.0 -0.38
R-squared 0.76 0.77 0.87 0.86
Jarque-Bera test of normality 0.44 0.26 2.65 2.72
[P-val] [0.80] [0.88] [0.26] [0.26]# Obs. 217 217 103 121
Sources: CEIC databaseNote: All variables are in logs. The dependent variable is household expenditure per capita, and government
spending variables are expressed per capita. The 2013 sample excludes 6 outliers, while the 2014 sample excludes 2 outliers.
Healthcare Spending most effective for rural households
21
Table 2. Rural Households: Consumption and Government Spending, Prefecture Level
2013 2014
(1) (2) (1) (2)
Disposable Income 0.86 0.87 0.75 0.76
(Standard error) (0.05) (0.03) (0.04) (0.04)
[P-val] [0.00] [0.00] [0.00] [0.00]
Gov. Social Security Spending -0.01 ... 0.08 0.08
(Standard error) (0.04) ... (0.03) (0.03)
[P-val] [0.87] ... [0.01] [0.01]
Gov. Health Spending 0.20 0.11 0.18 0.19
(Standard error) (0.08) (0.04) (0.06) (0.05)
[P-val] [0.01] [0.00] [0.00] [0.00]
Gov. Education Spending 0.00 ... 0.02 ...
(Standard error) (0.06) ... (0.03) ...
[P-val] [0.95] ... [0.60] ...
Constant 1.13 0.97 2.15 2.09
R-squared 0.80 0.83 0.86 0.86
Jarque-Bera test of normality 1.26 0.17 0.99 0.94
[P-val] 0.53 0.91 0.61 0.62
# Obs. 118 167 104 104
Sources: CEIC database
Note: All variables are in logs. The dependent variable is household expenditure per capita, and government
spending variables are expressed per capita. 2013 sample excludes 3 outliers; 2014 sample excludes 5 outliers
Financial repression less relevant with interest rate liberalization.
22
0
4
8
12
16
20
0
1
2
3
4
5
6
2000Q1 2002Q1 2004Q1 2006Q1 2008Q1 2010Q1 2012Q1 2014Q1 2016Q1
Benchmark deposit rate WMP Wenzhou index (rhs)
Interest Rates (In percent, period average; 1 year)
2016Q4
Rising income inequality in China
23Sources: Li (2013)
income inequality associated with saving inequality
24
0
5
10
15
20
25
30
35
40
45
50
1 2 3 4 5 6 7 8 9 10
200220072013
Urban Household Saving Rate by Income Percentile( saving in percent of disposable income)
Sources: CHIP.
Higher Gini associated with steeper saving curve
25
Slovenia 2010
France 2010
Taiwan 2013Australia 2010
Serbia 2013
Israel 2012
Peru 2013
Mexico 2012 South Africa 2012
Taiwan 1981
Taiwan 1986Taiwan 1991
Taiwan 1995Taiwan 1997
Taiwan 2000
Taiwan 2005Taiwan 2007
Taiwan 2010
0
10
20
30
40
50
60
70
0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7
Diffe
renc
e in
Avg
Hou
seho
ld S
avin
g Ra
te b
tw Q
90 a
nd Q
10
Gini Coefficient
Gini Coefficient VS Average Household Saving Rate Difference btw P90 and P10
Sources: Luxembourg Income Studies (LIS)
Chinese save more at any income decile, but larger gap at the bottom
26
-50
-40
-30
-20
-10
0
10
20
30
40
50
60
10 20 30 40 50 60 70 80 90 100
Aver
age
Hous
ehol
d Sa
ving
Rat
e
Household Income Decile
Average Household Saving Rate by Income Decile(in Percent)
France 2010, Gini 0.289
Taiwan 2013, Gini 0.308
Australia 2010, Gini 0.33
Mexico 2012, Gini 0.459
South Africa 2012, Gini 0.572
China 2013, Gini 0.47
Sources: Luxembourg Income Studies (LIS) for other countries. China 2013 from CHIP.
Household saving and public transfer
• Negative saving rate at bottom income percentiles reflects government transfer to poor households
• Poor in China also save high, reflecting the possible lack of transfer and social safety net for poor
• The introduction of health care (New Rural Cooperative Medical Scheme, NRCMS, 2003) and pension system (New Rural Social Pension Scheme, NRSPS, 2009) in rural area could be the driving force behind the declining HH saving rate in rural China, in a sharp contrast to urban China
27
Household saving rate: rural vs. urban
2810
15
20
25
30
35
40
1990 1992 1994 1996 1998 2000 2002 2004 2006 2008 2010 2012 2014
UrbanRural
Sources: CEIC; and staff calculations.
Figure 3 China: Urban and Rural Saving Rates, 1991-2015(In percent of Income)
Household saving and homeownership
29
0
10
20
30
40
50
60
70
80
90
100
1988 1995 2002 2007 2013
Perc
ent
homeownership rate saving rate
Sources: 1988, 1995 UHS; 2002, 07, 13 CHIP
Housing reform
Household saving and homeownership
30
0
5
10
15
20
25
30
35
40
1988 1995 2002 2007 2013
Perc
ent
tenant's saving rate homeowner's saving rate
Sources: 1988, 1995 UHS; 2002, 07, 13 CHIP
Household saving, homeownership, and housing reform
31
• DiD regression to estimate impact of housing reform (late 1990s-early 2000s) on HH saving rate
• Explanatory variables: homeownership dummy and other controls (HH head’s age, sex, edu, occupation, income, HH size,…)
• Data: Urban household survey (UHS): 1988, 1995; CHIP: 2002, 2007,2013.
Regression results
32
Year Homeownership Obs P_value
1988 -0.0010 12852 0.886
1995 0.0060 5906 0.511
2002 0.0283 6835 0.009
2007 0.0319 9998 0.019
2013 -0.0299 6255 0.104
Household saving, home ownership, and housing reform
33
• Home ownership has a significant impact on saving rate after the housing reform.
• The transmission channels include: down payment effect, mortgage effect and wealth effect
• Further controlling mortgage dummy (WIP) and interaction term shows some evidence that paying mortgage induced HH to save more
• Overall, given homeownership in China has plateaued, housing factor not likely a significant driver of household saving going forward.
CORPORATE SAVING
34
Corporate savings
• Main conclusions:– At the firm level, Chinese firms do not have
significantly higher savings than the global average
– SOEs tend to save less in gross terms but more in net terms, reflecting their lower profits and lower capex.
– Exchange rate has a significant impact on corporate saving for both SOE and private firms
35
Firm Level Summary Statistics
36
-0.01
0
0.01
0.02
0.03
0.04
0.05
0.06
0.07
China World ex. China Asia ex. China andJapan
Gross Saving
Net Saving
Corporate Saving Rate( in percent of assets)
Sources: Worldscope.
Saving ratios on downward trend after the GFC
-0.08
-0.04
0
0.04
0.08
0.12
0.16
2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015
Savings over Assets: SOEs vs. Private Companies(sample mean)
gs1_SOE gs1_Private ns1_SOE ns1_Private
37
Regression Results
• Following Bayoumi et al 2010, firm level regression controlling for size, q ratio, sector, year, exchange rate and lending rate
• SOEs tend to save less in gross terms (profits minus dividend)– SOEs make less profits and pay less dividend
• But SOEs tend to save more in net terms (gross saving minus capex)– SOEs spend less on capex
38
Regression ResultsGross savingratio
Net savingratio
Profit ratio Dividend ratio Capex ratio
Exchange Ratedepreciation
0.0965* 0.522*** 1.65 -6.47** -47.6***
Interest rate 0.0086* 0.039*** 0.27 -0.47* -3.48***
SOE dummy -0.0043*** 0.0075*** -0.634*** -0.21** -1.07***
Size 0.0046*** 0.0012 0.55*** 0.067 0.47***
Q ratio 0.0079*** 0.0058*** 1.05*** 0.255*** 0.31***
Constant -0.097*** -0.266*** -6.77* 2.32 17.0***
Observations 12,246 12,236 12,252 23,509 23,492
R-squared 0.14 0.076 0.16 0.068 0.095
39• The interaction terms of exchange rate and tradable sectors, and interest rate and SOEs are
not significant.
Policy Issues: SOE dividend
• SOE dividend payment should increase.
40
SOEs Private Companies
Dividend/Profit
0.22(mean)
0.31(mean)
Observa-tions
7,362 6,802
0
5
10
15
20
2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015
SOE profit transfer to fiscal budget (Ministry of Finace; net)
Central SOE profit transfer to fiscal budget (SASAC)
Transfer of SOE Dividend to Fiscal Budget(Percent of aggregate SOEs' profit)
Policy Issues: SOE entrusted lending
• Why do SOEs invest less than private firms—the role of entrusted loans?
41
0
1
2
3
4
5
6
7
8
9
2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015
SOE GFCF and Entrusted Loans (in yuan, trillion)
Entrusted
GFCF by SOEs (implied)
GOVERNMENT SAVING
42
Government savings volatile over time…
43
… driven mostly by revenue
44
high fiscal savings driven by low consumption
45
Relatively high savings in China… …due to low consumption, not high income
reflecting policy focus on investment and insufficient social spending
46
THE ROLE OF STATE
47
High Saving and the role of State• While increasing saving is often observed in economies where
growth take off, policy has played an important role in shaping China’s extreme high savings.
• One-child policy implemented in late 70s has drastically reduced fertility rate and raised saving rate.
• The transition in the 90s from a planned economy to market economy has dismantled social safety net, and led to higher precautionary savings of households.
• WTO entry has led to export-oriented growth surge, with significant rise in corporate saving as ER undervalued, and higher government saving from corporate income tax. Household savings also increased with rapid income growth and widening income inequality.
48
Policy Recommendation• Fiscal Policy Making income tax more progressive and more family friendly Increasing expenditure on social safety net, including health care, pension,
and targeted transfer to poor households. Increasing spending on general public service, e.g. staffing shortage in NBS Increasing public spending on childcare to boost fertility rate [Limit potential burden on households from bank recapitalization]
• Hukou reform to give migrants equal access to education and social safety net
• SOE reform Increasing dividend payment to budget Hardening budget constraint• Improve access to formal financing of private firms• Ensure exchange rate in line with fundamental
49
Make income tax more progressive
50
Increasing SOE dividend payment to government
51
Increasing expenditure on social safety net
52
0.0
0.5
1.0
1.5
2.0
2.5
3.0
3.5
4.0
4.5
200020022003200420052006200720082009201020112012201320142015
EducationSocial securityHealth
Government Spending(in percent of GDP)
Sources: CEIC and IMF staff estimates.
PROSPECT OF SAVING RATE
53
Ageing to accelerate
54
0
10
20
30
40
50
60
70
80
90
2015
2020
2025
2030
2035
2040
2045
2050
2055
2060
UN medium fertility
UN high fertility
UN low fertility
Old age dependency ratio across UN scenarios(65+ over 20-64 population, percent)
Sources: UN, staff estimates.
Saving rate to fall with rapid-ageing
55
Saving to fall further with lower wage growth and interest rate
56
But results sensitive to assumptions
57
Higher social spending and lower income inequality also help reduce household saving
58
-6
-5
-4
-3
-2
-1
0
16
17
18
19
20
21
22
23
2016 2017 2018 2019 2020 2021 2022
social safety net (RHS)
Ageing effect (RHS)
income equality (RHS)
Household Saving Rate to Fall -- Proactive(In percent of GDP)
Source: IMF staff estimates and projections.
Household saving rate
Government and corporate saving to fall with changing policy and declining capital return
59
-10
0
10
20
30
40
50
60
1992
1994
1996
1998
2000
2002
2004
2006
2008
2010
2012
2014
2016
2018
2020
2022
government financial
non-financial firms households
National Saving Rate to Fall(in percent of GDP)
Sources: Flow of funds, IMF staff estimates and projections.
International experience: Household Savings
60
yearincome level (1990 US$)
Taiwan POC 1993 11950Japan 1974 11145Korea 1991 10610China 2010 9558Sources: Haver and Madison Database.
Table: Peak of household saving rate
International experience: Household Savings
61
-5
0
5
10
15
20
25
30
35
t=0 2 4 6 8 10 12 14 16 18 20 22 24 26 28 30 32 34 36 38 40
Taiwan
Japan
Korea
Household Saving Rate( in percent of disposable income)
Sources: t=0 is defined as the year when household saving rate peaked, 1974for Japan, 1991 for Korea and 1993 for Taiwan.
International experience: Taiwan
62Sources: Luxembourg Income Studies (LIS)
-50-40-30-20-10
0102030405060
10 20 30 40 50 60 70 80 90 100
Aver
age
Hous
ehol
d Sa
ving
Rat
e
Household Income Decile
Taiwan 1981, Gini 0.267
Taiwan 1991, Gini 0.271
Taiwan 2005, Gini 0.305
Taiwan 2010, Gini 0.317
Average Household Saving Rate by Income Decile, Taiwan(in Percent)
THANK YOU!
63
Interaction of saving across sectors
• Globally, the increase of corporate saving was largely offset by falling household saving.
• In China, corporate saving and household saving tend to move together, likely reflecting common income effect and limited household ownership of corporates.
• Government saving increases with higher corporate saving reflecting tax revenue, while insufficient government expenditure on social safety net also increase household saving.
64
Similar results from recent provincial data
65
Table 3. Households: Consumption and Government Spending, Provincial Level, 2013-15
Urban Rural Total
Disposable Income 0.92 0.76 0;92(Standard error) (0.05) (0.05) (0.03)[P-val] [0.00] [0.00] [0.00]
Gov. Social Security Spending 0.07 ... 0.10(Standard error) (0.03) ... (0.03)[P-val] [0.04] ... [0.00]
Gov. Health Spending … 0.22 …(Standard error) … (0.03) …[P-val] … [0.01] …
Gov. Education Spending 0.11 ... …(Standard error) (0.07) ... …[P-val] [0.03] ... …
Constant -0.23 0.45 -0.13
R-squared 0.95 0.92 0.83# Obs. 30 30 30
Sources: CEIC database
Note: All variables are in logs, and are an average for the period 2013-15. The dependent variable is household expenditure per capita, and government spending per capita.
Provincial regression before data revision
66
Table 4. Households: Consumption and Government Spending, Provincial Level, 1995-2012
Urban Urban Rural Rural
2007-12 1995-2012 2007-12 1995-2012
Disposable Income (real 1/) 0.96 0.76 0.55 0.31
(Standard error) (0.12) (0.05) (0.17) (0.14)
[P-val] [0.00] [0.00] [0.00] [0.03]
Gov. Social Security Spending (real 1/) -0.03 … 2/ -0.07 … 2/
(Standard error) (0.02) ... (0.03) ...
[P-val] [0.12] ... [0.04] ...
Gov. Health Spending (real 1/) 0.05 0.02 0.12 0.03
(Standard error) (0.01) (0.01) (0.02) (0.02)
[P-val] [0.00] [0.04] [0.00] [0.21]
Gov. Education Spending (real 1/) -0.05 ... -0.05 ...
(Standard error) (0.01) ... (0.01) ...
[P-val] [0.00] ... [0.00] ...
Population growth 1.72 ... 1.72 ...
(Standard error) (0.45) ... (0.45) ...
[P-val] [0.00] ... [0.00] ...
Lagged dependent -0.03 -0.11 -0.03 -0.11
(Standard error) (0.06) (0.03) (0.06) (0.08)
[P-val] [0.56] [0.00] [0.56] [0.15]
Constant -0.02 0.01 0.01 0.05
R-squared 0.64 0.59 0.58 0.53
# Obs. 151 458 151 458
Sources: CEIC databaseNote: All variables are in logs change terms. The dependent variable is household expenditure per capita, and government
spending variables are expressed per capita.
1/ In real terms, deflated by the provincial consumers’ price index
2/ Social security spending data is not available prior to 2007 with a coverage consistent with post 2007 data.