distribution, asset prices, debt and aggregate demand in ... · • carvalhoand rezai(2014): var(2)...
TRANSCRIPT
Distribution, Asset Prices, Debt and
Aggregate Demand in OECD Countries
Engelbert Stockhammer & Rafael Wildauer
Kingston University, London
Part of INET project Rising Inequality as a Cause of the Crisis
Motivation
• macro trends over last 30 years
• strong house (and equity price) growth prior to the recession
• strong shifts in functional and/or personal income distribution
• Debates
• Effects of distribution on demand
• Wage-led vs profit-led demand regimes
• Consumpution cascades hypothesis: inequality increases demand
• Finance and growth
• Keynesian theory: debt important in theory (Minsky cycles), but little empirical
• Level vs flow effects of debt
• Wealth effects in mainstream
This paper (contribution)
• extend Bhaduri Marglin framework in order to test
competing hypothesis
• personal income distribution:
Consumption cascades vs Kalecki
• wealth and debt
rational vs no aggregate vs collateral/Minsky vs stock/flow
• Growth drivers (1998-2008): debt-driven growth?
• panel of 18 OECD countries, 1980-2013
Basic Bhaduri Marglin model
� � � �,��, �� �,��, �� � �,��, ���
•��
���� 0;
��
���� 0;
���
���� 0;
• WS ... wage share (W/Y)
• no disagreement over signs of��
���and
��
���
��∗
���� �1/�1 � �2�
• � ���
���
��
���
���
���
• �! ���
��
��
��
���
��
• If � � 0 profit-led demand
• If � � 0 wage-led demand
Effects of inequality and wealth: extending
the Bhaduri Marglin model
• Veblen effect: ‘keeping up with the (richer) Joneses’
• inequality ↑ → consumption emulation → C↑: "�/"# � 0
• that’s ad odds with Kaleckian intuition: "�/"# � 0
• Including wealth and debt
• � � ���,��, #, $�, %�,&$�
• rational wealth effects: �� � $� %� � &$, thus "�/"�$� %�� � �"�/"&$
• no aggregate wealth effects: Buiter (2010): "�/"$� � 0
• Muellbauer: housing is collateral "�/"$� � "�/"%�
• � � ���,��, #, &$, ∆&$�, with "�/"&$ � 0, "�/"∆&$ � 0
• positive flow and negative stock effect
• Debt-led vs debt-burdened demand regimes (Dutt 2006; Nishi 2012)
• Minsky for households: "�/"$� � 0 (and "&$/"$� �� 0)
• Asset-driven growth: if combined effect of
• ()�∆$� (*�∆%� (+)∆&$ �� 0
• for C as well as I
Hypothesis Theoretical prediction Implications
the rich save moreRicher households have a higher
propensity to save.
,-
,.� /
expenditure cascadesHouseholds make consumption
decisions with respect to richer peers.
,-
,.� /
housing wealth is no wealth
Rising house prices lead to wealth effect
for home owners and higher savings of
willing-to-be-homeowners.
,-
,01� /
net wealth effectNet wealth (NW) is the relevant wealth
measures for consumption decisions.
,-
, 01203� �
,-
,41
credit constraints
Due to credit constraints, changes in
housing wealth effect consumption
even if shocks are anticipated.
,-
,01� / and
,-
,01�
,-
,03
Minskyian householdsRising asset prices leads to increasingly
optimistic lending and spending.
,-
,01� /
stock and flow effects of debt
The stock of debt implies interest rate
payments which affect consumption
negatively whereas the flow of debt
affects consumption positively.
,-
,41� / and
,-
,∆41� /
the literature
Empirical literature on Bhaduri-Marglin
models
single equation systems approach Panel
Basic model
(few controls)
Bowles & Boyer 1995Stockhammer & Onaran
2004
Hartwig
2014
Naastepad & Storm 2006/2007 Barbosa-Filho & Taylor 2006
Hein & Vogel Carvalho & Rezai 2014
Stockhammer & Stehrer 2011 Rada & Kiefer 2014
Onaran & Galanis 2012
Financialisation
variables
Hein & Schoder 2011 (USA, D): int, div
Nishi 2012a (J): DOnaran et al. 2011 (USA): int, div, FW,
HW, D
Detailed
treatment of
globalisation
Stockhammer et al. 2009
Stockhammer et al. 2011
Extensions of Bhaduri Marglin model
• Onaran et al 2011: extend BM model to include FW, WH,
rentier payments and incomes for USA
• Nishi 2012: VAR(3) BM model with debt for Japan
• Carvalho and Rezai (2014): VAR(2) BM with regimes
according to high or low personal inequality; USA
Wealth and debt effects in PK literature
• Growing theoretical literature on the effects of finance and debt on growth, but little on consumption• Minsky models (Keen 1995, Fazzari 2006, Charles 2008, Ryoo 2013): business
debt
• SFC models: stock flow norms
• Flow vs stock effects of debt: Dutt (2006), Hein (2012) debt without asset P
• Bezemer et al 2014: panel of 40 countries; growth as function of disaggregated debt; find neg. effect of asset-related debt on growth
• Kim et al (2014): PK theory of consumption based on relative income and rules of thumb; estimate C as function of Y, W, DH, ∆ DH for USA
• Zezza (2009): PE = C+I = f(Y, FAF,G, SP, HP, ∆DH, ∆DB) for USA
• None of them controls for distribution
Mainstream wealth effects literature
• Rich (if recent) empirical literature, but theoretically
ambiguous
• Rational behaviour assumption (HH never worry about bubbles)
• No role for debt in baseline NK models
• Girourard et al 2006, Ludwig and Slok 2004, Slacalek 2009
• Muellbauer 2009: credit constraints
• Use either asset P or wealth
• Find wealth MPC ≈ 0.05
• USA, UK MPC(HW)>>MPC(FW), but not for Eu
Veblen effects? Empirical
• (somewhat) related literature: relative income hypothesis, micro
data, happiness, labour supply
• Effects of income relative to peer group on self-reported happiness (McBride
2001)
• Neumark and Postlewaite 1998: labour supply decisions as depending on
income of husband of sister
• Bowles and Park (2005): labour supply as function of relative income
• Alvarez-Cuadrado et al 2012: micro data consumption as positive function of
average consumption of geographical reference group
• (somewhat) related: current account as negative function of
inequality [motivated by consumption cascades, but not
documented] (Behringer and van Treeck 2013)
• Brown (2004): macro consumption function for USA, finds negative
effect of inequality
Data, Estimation and Results
Data
• 18 OECD countries (AT, AU, BE, CA, CH, DE, DK, ES, FI, FR,
IE, GB, IT, JP, NL, NO, SE, US)
• sample: 1980-2013
• Sources:
• AMECO (national account data)
• OECD (property prices, stock prices, interest rates)
• BIS (household debt)
• World Top Income Database (top income shares)
• IMF (stock price indices)
A Panel Approach
• BM: estimate determinants of each GDP component
• Country fixed effects (FD estimator) and HAC standard
errors
• consumption: � � 5��,��, #, 66, �6, &$�
• investment: � 5��,��, 7, #, 66, �6, &$,&8�
• foreign sector: � 5��5,��, 9:, 66�
; � 5��,��, , 9:, 66�
Comments on panel estimtors
• Panel = impose identical coeffs
• Report MG (mean group) estimator to check robustness of pooling
• We have a macro panel small N, medium T (N=18, T=33)
• Compared to much of the panel literature (assumption of large N) we
have to worry more about TS aspects
• Can’t reject unit roots for our variables
• Estimate FD (first difference) estimator
• Check with dynamic estimators A&H, restricted difference GMM
• Experiment with cointegration estimators (PMG, MG, pooled)
• FD preferred estimator
Results Consumption
Results: Consumption
Interpretation results consumption
• no stable cointegration relationship
• Kalecki effect:��
���� 0supported
• Veblen effect: ��
�=� 0 not supported
• Rational wealth effects: ���
� )�2*���
��
�+)not supported
• Buiter (2010):��
�)�� 0 not supported
• Muellbauer (and Minsky for HH):��
�)��
��
�*�supported
• Debt-led vs debt-burdened: ��
�+)� 0,
��
�∆+)� 0not supported
Results Investment
Results Foreign Sector
Demand regimes (marginal effects)
and
Growth contributions
Private excess demand
• Domestic demand is wage led
• Very open economies can have profit-led demand
Debt-driven growth?
Conclusion
• Functional distribution:
• WS positive on C but modest
• Personal income inequality:
• no stat sign effect (or negative on I)
• robust to using Gini
• Wealth and debt:
• strong and robust asset effect (HW, FW, DH)
• no evidence for -DH , +∆DH in C;
• In decade prior to the crisis: asset price effects are
substantial; effects of functional distribution modest
(asset-driven growth)
Appendix III: Cointegration
Consumption
Investment
Appendix IV: Crisis dummies and
interactions
Appendix V: Investment (construction)
Appendix VI: Investment (non-
construction)
Zezza (2009): USA
Kim et al 2014