Download In this presentation

Inequality, Debt and Credit Stagnation
Steve Keen
Kingston University London
IDEAeconomics
Minsky Open Source System Dynamics
www.debtdeflation.com/blogs
Secular Stagnation Mark I
• Hansen 1939
– “Not until the problem of full employment … from the long-run,
secular standpoint was upon us, were we compelled to give serious
consideration to those factors … which tend to make business
recoveries weak .. and which tend to prolong and deepen the
course of depressions.
– This is the essence of secular stagnation—sick recoveries which die
in their infancy and depressions which feed on themselves and leave
a hard and seemingly immovable core of unemployment.” (Hansen,
1939, p. 4)
• Hansen blamed “external factors”
– Technological change & population growth:
• “Fundamental to an understanding of this problem are the
changes in the "external" forces, if I may so describe them,
which underlie economic progress—changes in the character of
technological innovations, in the availability of new territory, and
in the growth of population.” (Hansen, 1939, p. 4)
Secular Stagnation Mark I
• His timing was a bit off…
US Unemployment Rate
28
Hansen1934 Hansen1939
26
24
Percent of workforce
22
21
20
Hansen’s expectation
18
17
16
14
12
10
8
6
4
2
0
1920
1925
1930
1935
1940
1945
1950
1955
1960
Secular Stagnation Mark II
• 80 years later, along comes Larry:
– “a decline in the full-employment real interest rate (FERIR) coupled
with low inflation could indefinitely prevent the attainment of full
employment…”
• This is after the financial crisis, which, of course, is no longer an issue:
– “If a financial crisis represents a kind of power failure, one would
expect growth to accelerate after its resolution as those who could
not express demand because of a lack of credit were enabled to do
so…
– How might one understand why growth would remain anaemic in
the absence of major financial concerns? Suppose that a substantial
shock took place … and that this tended to raise private saving
propensities and reduce investment propensities…
– one would expect interest rates to fall … until the saving and
investment rate were equated at the full-employment level of
output…
– But this presupposes full flexibility of interest rates…”
Secular Stagnation Mark II
• Finance clearly irrelevant when Hansen & Summers extemporised…
USA private debt to GDP
180
Hansen1934
Summers
170
160
150
140
130
Percent of GDP
120
110
100
90
80
70
60
50
40
30
20
10
0
1820
1840
1860
1880
1900
1920
1940
1960
BIS and US Census Normalized Data
1980
2000
2020
Secular Stagnation Mark II
• Back to Larry’s FERIR brainwave:
• “A variety of structural changes … suggest that FERIR levels may have
declined substantially. These include:
– Slower population and possibly technological growth means a
reduction in the demand for new capital goods to equip new or
more productive workers…”
• And the cure for a FERIR is?
– “some major exogenous event will occur that raises spending or
lowers saving…”; or
– “In the long run, as the economy’s supply potential declines, the
FERIR rises, restoring equilibrium – albeit not a very good one.”
• So a FERIR is really, really important!
• Um, but what is it though?…
Secular Stagnation Mark II
• The FERIR was recently discovered by MIT’s fabled CERN* laboratory
– CERN stands for “Crazy Economic Rationalisations of aNomalies”
• A “FERIR” (“Full-Employment Real Interest Rate”) is an antiparticle to a
“NAIRU” (“Non-Accelerating Inflation Rate of Unemployment”)
• FERIR created when a NAIRU interacts with a GFC (“Global Financial
Crisis”) in CERN’s economic particle equilibrator (the “DSGEin”)
– DSGEin crashed—since it assumed that GFC’s don’t exist
– GFC came from orthogonal universe called TRW (“The Real World”)
• Equilibration with a GFC caused the NAIRU to decay…
– And to emit a ZLB (“Zero Lower Bound”) and a FERIR
• The long-lived ZLB particle inverts all other standard particles, so that
– HMDs (“Helicopter Money Drops”) which were mad, are now sane
– Growth, which was high, is now low
– CBs (“Central Banks”) which prevent inflation, now try to cause it; &
– Inflation, which was bad & everywhere, is now good & nowhere
• Footnote *: CERN has been reported to COCOA (the “Campaign to
Outlaw Contrived and Outrageous Acronyms”)
Secular Stagnation Mark II
• Of course, there was no empirical data that might have alerted CERN to
the possible importance of the CREDIT particle…
Credit and Employment (Correlation 0.93)
15
0
Summers
14
13
1
12
11
2
9
3
Percent of GDP
8
7
4
6
5
5
4
3
6
2
1
7
0
0
1
8
2
3
4
5
1990
9
Credit
Unemployment
1992
1994
1996
1998
2000
2002
2004
2006
BIS & BLS Data
2008
2010
2012
2014
2016
2018
Percent of workforce (Inverted)
10
Secular Stagnation Mark I
• Just like there was nothing to alert Hansen in 1934
Credit and Employment (Correlation 0.8)
12
Great Depression
0
Hansen1934
2.5
8
5
6
7.5
4
10
2
12.5
0
0 15
2
17.5
4
20
6
22.5
8
25
 10
 12
1920
Credit
Unemployment
1922
1924
27.5
1926
1928
1930
1932
BIS & Census Data
1934
1936
1938
1940
Percent of workforce (Inverted)
Percent of GDP
10
Secular Stagnation Mark II
• This correlation is zero in the TSM, & therefore can be ignored:
Change in Credit and Employment Change (Correlation 0.88)
4
GFC
 40
Summers
3
 30
2
 20
1
 10
00
1
10
2
20
3
30
4
40
5
50
6
60
7
70
8
80
9
90
 10
100
 11
110
 12
120
 13
 14
 15
1990
130
Credit Change
Unemployment Change
1992
1994
1996
1998
140
2000
2002
2004
2006
BIS & BLS Data
2008
2010
2012
2014
2016
150
2018
Percent change per year (Inverted)
Percent of GDP per year
0
Let’s leave MIT, CERN and TSM for TRW…
• Credit doesn’t exist in Neoclassical macro because of “Loanable Funds”
– “Think of it this way: when debt is rising, it’s not the economy as a
whole borrowing more money.
– It is, rather, a case of less patient people—people who for whatever
reason want to spend sooner rather than later—borrowing from
more patient people.” (Krugman 2012, pp. 146-47)
• ole for credit still not accepted in Post Keynesian macro either…
– “In this primer we will examine the macroeconomic theory that is
the basis for analysing the economy as it actually exists. We begin
with simple macro accounting, starting from the recognition that at
the aggregate level spending equals income.” (Wray 2011)
– “Unless Keen (2014a) can explain how a purchase of a good or
service does not provide income for the seller, then he should
rethink his claim that debt extensions can force an inequality
between expenditure and income at the aggregate level…
– a sector can spend more than its current income, but the sum of
sectors cannot.” (Fiebiger 2014, p. 296)
Integrating Credit into Income  Expenditure
• An expenditure table view:
– Divide economy into 3 non-bank sectors plus banking sector
– Aggregate Expenditure negative sum of diagonal
– Aggregate Income positive sum of off-diagonal elements
– All flows (in $/Year at a point in time) shown in lowercase
– All stocks (in $ at a point in time) shown in UPPERCASE
– Greek r used for interest rate
– First case: lending/borrowing does not occur:
Assets
Loans
Level ($)
S1
S2
S3
BE
Liabilities
S1
-(a+b)
c
e
AE   a  b    c  d    e  f 
AY  a  b  c  d  e  f  AE
S2
S3
Flows ($/Year)
a
b
-(c+d)
d
f
-(e+f)
Equity
BE
Credit and Income  Expenditure
• Loanable Funds and (almost) no role for credit
– Sector 1 borrows l ($/Year) from Sector 2
– Pays interest of r.L ($/Year) to Sector 2
Assets
Loans
Level ($)
S1
S2
S3
BE
Liabilities
S1
-(a+b+l+r.L)
c
e
AE   a  b  r  L    c  d    e  f 
AY  a  b  r  L  c  d  e  f  AE
S2
Flows ($/Year)
a+r.L
-(c+(d-l))
f
S3
b+l
d-l
-(e+f)
Equity
BE
Credit and Income  Expenditure
• Endogenous Money and an essential role for credit
– Sector 1 borrows l ($/Year) from banking sector
– Pays interest of r.L ($/Year) to banking sector…
S1
S2
S3
BE
Assets
Loans
Level ($)
L
l
Liabilities
S1
-(a+b+l+r.L)
c
e
g
S2
S3
Flows ($/Year)
a
b+l
-(c+d)
d
f
-(e+f)
h
i
AE   a  b  l  r  L    c  d    e  f    g  h  i 
AY  a  b  l  r  L  c  d  e  f  g  h  i  AE
• Change in debt (credit) plays an essential role in aggregate
expenditure & aggregate income with endogenous money
• Expenditure is fundamentally monetary
• 2 sources of expenditure: turnover of existing money
• New expenditure financed 1:1 by new debt
Equity
BE
r.L
-(g+h+i)
Credit and Income  Expenditure
• How to measure?
– GDP a (poor) approximate measure of flow of expenditure financed
by existing money in $/Year
– Change in debt a (better) measure of flow of credit created by new
debt in $/Year
– Dimensionally accurate & empirically OK to add together to measure
aggregate expenditure at a point in time
– Analogy
• Flow in river
• with a pump
injecting or removing
water:
Credit ($/Year)
The “Smoking Gun of Credit” & Walking Dead of Debt
• Add GDP to change in debt (credit) to measure aggregate expenditure
• Peak GDP+Credit identifies every economic crisis since Japan…
Japan GDP and Credit
600000
55
50
Average credit 5 years
before Crisis 18% GDP
45
40
35
400000
30
Average after:
minus 1.8% GDP
25
20
15
10
200000
5
0
0
5
 10
 15
0
1970
1975
1980
1985
1990
1995
BIS Data
2000
2005
2010
2015
 20
2020
Change in Private Debt per year as percent of GDP
Billion Currency Units per year
60
Crisis Japan
GDP
GDP + Credit
Credit
The “Smoking Gun of Credit” & Walking Dead of Debt
• USA
USA GDP and Credit
20000
GDP
GDP + Credit
Credit
55
45
40
35
30
Average after:
25
3.7% GDP 20
10000
15
10
5
5000 0
0
5
 10
 15
0
1970
1975
1980
1985
1990
1995
BIS Data
2000
2005
2010
2015
 20
2020
Change in Private Debt per year as percent of GDP
50
Average credit 5 years
before Crisis 11% GDP
15000
Billion Currency Units per year
60
CrisisUSA
The “Walking Dead” of Private Debt
• Countries now in post-debt crisis with low credit-based demand
Private Debt to GDP Ratios in Walking Dead Economies
340
GFC
USA
UK
Japan
Spain
Ireland
Netherlands
Denmark
320
300
280
260
240
Percent of GDP
220
Country
Crisis Date Peak debt
Now
180
USA
2006.4
169
150
160
UK
2008.3
197
160
Japan
1990.25
221
167
Spain
2006
217
172
Ireland
2006.5
330
257
40
Netherlands
2015.25
247
238
20
Denmark
2006
268
230
200
140
120
100
80
60
0
1960
1965
1970
1975
1980
1985
1990
BIS data
1995
2000
2005
2010
2015
2020
The “Walking Dead” of Private Debt
• Flat to negative credit-based demand
Credit in Walking Dead Economies
60
55
50
45
40
35
Percent of GDP per year
Crisis Jap an
USA
UK
Japan
Spain
Ireland
Netherlands
Denmark
GFC
30
25
20
15
10
5
0
5
 10
 15
Country
10 years Pre
Post % GDP
USA
10
4.4
UK
13.75
2.5
Japan
15.8
0.2
Spain
17
1.8
Ireland
21.6
13
0
 20
Netherlands
6.9
12
“Schrodinger’s
Zombie”
 25
Denmark
 30
1960
1965
12.7
1970
1975
8.7
1980
1985
1990
BIS data
1995
2000
2005
2010
2015
2020
The “Smoking Gun of Credit” & Future Debt Zombies
• Future Debt-Zombies:
– Countries with >150% GDP private debt to GDP
– Where debt is growing quickly (above 10% of GDP per year)
– Can’t predict timing
• Can be delayed by government enticement into private debt
– Australia 2008 “First Home Vendors Boost”
– UK “Help to Sell”
– But inevitable since at high levels even stabilisation of debt/GDP
ratio causes fall in aggregate demand & income
The “Smoking Gun of Credit” & Future Debt Zombies
• Low debt ratio
GDP Growth Rate
Debt Growth Rate
Final Debt Growth
Initial Debt Ratio
Years
GDP
Debt
Debt to GDP Ratio
Credit
Total Demand
Demand Growth Rate
10%
20%
10%
50%
0
1000
$500
50%
1
$1,100
$600
55%
$100
$1,200
2
$1,210
$720
60%
$120
$1,330
10.8%
3
4
$1,331 $1,464
$864 $1,037
65%
71%
$144
$173
$1,475 $1,637
10.9%
11.0%
5
6
$1,611 $1,772
$1,244 $1,369
77%
77%
$207
$124
$1,818 $1,896
11.1%
4.3%
The “Smoking Gun of Credit” & Future Debt Zombies
• Medium debt ratio
GDP Growth Rate
Debt Growth Rate
Final Debt Growth
Initial Debt Ratio
Years
GDP
Debt
Debt to GDP Ratio
Credit
Total Demand
Demand Growth Rate
10%
20%
10%
100%
0
$1,000
$1,000
100%
1
$1,100
$1,200
109%
$200
$1,300
2
$1,210
$1,440
119%
$240
$1,450
11.5%
3
$1,331
$1,728
130%
$288
$1,619
11.7%
4
$1,464
$2,074
142%
$346
$1,810
11.8%
5
$1,611
$2,488
155%
$415
$2,025
11.9%
6
$1,772
$2,737
155%
$249
$2,020
-0.2%
The “Smoking Gun of Credit” & Future Debt Zombies
• High Debt Ratio
GDP Growth Rate
10%
Debt Growth Rate
20%
Final Debt Growth
10%
Initial Debt Ratio
125%
Years
0
1
2
3
4
GDP
$1,000 $1,100 $1,210 $1,331 $1,464
Debt
$1,250 $1,500 $1,800 $2,160 $2,592
Debt to GDP Ratio
125%
136%
149%
162%
177%
Credit
$250
$300
$360
$432
Total Demand
$1,350 $1,510 $1,691 $1,896
Demand Growth Rate
11.9% 12.0%
12.1%
• Future Debt Zombies include…
5
6
$1,611 $1,772
$3,110 $3,421
193%
193%
$518
$311
$2,129 $2,083
12.3%
-2.2%
Some Future Debt Zombies
• A sample of 18 vulnerable countries…
Private Debt to GDP Ratios in Future Zombie Economies
240
GFC
China
Canada
Korea
Australia
Norway
Sweden
France
220
200
180
Percent of GDP
160
Country
Debt Ratio %
Credit % GDP
China
210
29
Sweden
237
15
Canada
210
13
Korea
194
13
Australia
207
11
40
Norway
234
6
20
France
181
5
140
120
100
80
60
0
1960
1965
1970
1975
1980
1985
1990
BS data
1995
2000
2005
2010
2015
2020
The “Smoking Gun of Credit” & Future Debt Zombies
• Canada: Debt crisis almost certain during life of Trudeau Government
Canada GDP & Credit
2500
60
Great Recession
GDP
GDP+Credit
Credit
55
50
45
2000
40
30
25
1500
20
15
10
5
1000 0
0
5
 10
 15
500
1990
1992
1994
1996
1998
2000
2002
2004
2006
BIS & BLS Data
2008
2010
2012
2014
2016
 20
2018
Percent of GDP
Billion Loonies per year
35
The “Smoking Gun of Credit” & Future Debt Zombies
• Australia : Debt crisis almost certain during life of ??? Government
Australia GDP & Credit
2000
60
Great Recession
GDP
GDP+Credit
Credit
55
50
45
1500
40
30
25
1000
20
15
10
5
500 0
0
5
 10
 15
0
1990
1992
1994
1996
1998
2000
2002
2004
2006
BIS & BLS Data
2008
2010
2012
2014
2016
 20
2018
Percent of GDP
Billion Dollars per year
35
The “Smoking Gun of Credit” & Future Debt Zombies
• China: debt crisis inevitable, but form it will take???
China GDP & Credit
100000
60
Great Recession
GDP
GDP+Credit
Credit
55
50
45
80000
35
30
60000
25
20
15
40000
10
5
0
0
20000
5
 10
 15
0
1990
1992
1994
1996
1998
2000
2002
2004
2006
BIS & BLS Data
2008
2010
2012
2014
2016
 20
2018
Percent of GDP
Billion Renminbi per year
40
Modeling credit in capitalism
• A “Lucas-critique-immune” modelling paradigm
– Derive macro models from strictly true identities
• My Minsky model can be derived by putting these 3 definitions in
dynamic form:
– Employment rate L/N=l;
– Wages share of GDP W/Y=w;
– Private debt to GDP ratio d=D/Y
• Differentiate with respect to time and you get:
• “Employment will rise if economic growth
1 d
exceeds the sum of population & labor
l  lˆ  YˆR  aˆ  Nˆ
l dt
productivity growth”

• “Wages share of output will rise if wage
rise exceeds growth in labor productivity”
• “Debt ratio will rise if rate of growth of
debt exceeds rate of growth of GDP”
1

d
ˆ R  aˆ
 w  wˆ  w
w dt
1 d
 d  dˆ  Dˆ  YˆR
d dt
Modeling credit in capitalism
• Operationalise with simplest possible linear expressions
Y
dK
dw
L
 IFN  r   Y   KR  K
 w  wFN  l 
a
dt
dt
K
IFn  r    S   r   Z 
wFN  l   lS   l  lZ 
Y
v
• We get this system with intrinsic nonlinearities:
da
 a
dt
dN
  N
dt


 1w  r d





N
 S 

v


l  l 
      KR  
v






w  w   lS   l  lN    


 1w  r d





N
 S 



v
 1w  r d



d   S 
  N   1  w  r  d   d  
  KR 
v
v










Two feasible outcomes
• (1) Convergence to “good” equilibrium
Stable system
(Linear
functions)
Wages
Employment
Private
Share
Debt
of Rate
Ratio
Output
Percent
of GDP
peremployed
year
Percent
of GDP
Percent
of population
100
66
80
64
90
60
62
80
40
60
70
20
58
60
0
56
54
50
20
00
0
50
50
50
100
100
100
www.debtdeflation.com/blogs
www.debtdeflation.com/blogs
Stable
150
150
150
200
200
200
Two feasible outcomes
• (2) Convergence to “bad” equilibrium after apparent “moderation”
Unstable system (Linear functions)
Wages
Employment
Private
Share
Debt
of Rate
Ratio
Output
Percent
of GDP
Percent
of GDP
peremployed
year
Percent
of population
120
400
70
300
100
65
200
60
80
100
55
60
0
 100
50
40
000
Unstable
20
20
20
40
40
40
60
60
60
www.debtdeflation.com/blogs
www.debtdeflation.com/blogs
www.debtdeflation.com/blogs
80
80
80
100
100
100
We’ve seen this before—in complex systems
• Property of Lorenz “chaotic” model of fluid flow
• This
behaviour
cannot be
generated by
standard
equilibriumoriented
“linear”
model
• Inherent
nonlinearity
& nonequilibrium
dynamics are
essential
800
• Decreasing followed by
increasing turbulence
600
400
200
• Convergence to
laminar flow…
0
 200
0
5
10
15
20
Modeling credit in capitalism
• Solving for Eq, wEq and dEq yields:
lE 
q

 lZ
lS


 v      K

R
 Eq  v  
 Z 


S


 v      K

R
v       KR  v  
 Z 


S


dEq 
 




• Wages share is a residual:
wE  1   E  r  dE
e
•
•
•
•
e
e
Inequality stabilises if debt ratio converges to stable equilibrium
Inequality rises if debt ratio continues to rise
Rising inequality as a sign of systemic breakdown
Paradoxes as well: higher investment propensitylower growth
Modeling credit in capitalism
• Higher propensity to invest—higher debt level—lower growth
• Strong sensitivity of
Parameters
Parameters
debt to slope of
 S  6
 Z  3%
 S  5
 Z  3%
investment function
l S  10
l Z  60%
l S  10
l Z  60%
  2%
  1%
KR  6%
v  3
r  4%
• Bankers benefit at
expense of
capitalists, workers
Equilibria given parameters
l Eq  


lS
 l Z   60.2 %

 v       KR
 sEq  v  
S


  Z  25.2 %

 v       KR

v       KR  v  
  Z
S

  60 %
d Eq 

b Eq  r d Eq
So bankers share is
b Eq  2.4 %
wEq  1   sEq  b Eq  72.4 %
Equilibrium growth rate

v      KR
g Eq 
S
v

• Higher desire to
invest, lower
growth rate
  2%
  1%
v  3
Equilibria given parameters
l Eq  


lS
 l Z   60.2 %

 v       KR
 sEq  v  
S



  Z  22.5 %


 v       KR
v      KR  v  
d Eq 
S


b Eq  r d Eq
So bankers share is
b Eq  6 %
wEq  1   sEq  b Eq  71.5 %
Equilibrium growth rate

v      KR
 Z
 2.8 %
KR  6%
g Eq 
S
v

 Z
 2.5 %

  Z
  150 %
r  4%
Simple complex systems model…
• With price dynamics & variable interest rate
 s  1  w  t   r  t   d  t  ; r 
s
v
Inflation-adjusted
r  t   if  inflagnominal
inflag  t  , rrate
 t   0, rb interest
b
1st
1 
1

inf
t


1


w
t




order time lag
determines
 inflation
 P  1  s

  Ifn  r 


1 d
 l  
  Kr       


l dt
v



1 d
 w  w fnaffects
 inf  t  share
Inflation
 l    wages
w dt
 Ifn  r  

 s
 debt

v affects
 Ifn  rgrowth


1 d Inflation


 d
 
  Kr   inf  t  


d dt
d

 v

inf  t  
1
d
1 
inflag  t  rate

1
Lagged interest
reaction
toinflation
inflag  t  dt
 inf  inflag  t  
Simple complex systems model…
• The same model in Open Source system dynamics program Minsky:
Conclusion
• We’re suffering from credit stagnation, not “secular stagnation”
• Summers is as wrong now as Hansen was then…
USA private debt to GDP
180
Hansen1934
Summers
170
160
150
140
130
Percent of GDP
120
110
100
90
80
70
60
50
40
30
20
10
0
1820
1840
1860
1880
1900
1920
1940
1960
BIS and US Census Normalized Data
1980
2000
2020
Debt to GDP Ratios in France
• France
Debt to GDP Ratios France
200
190
GFC
Private
Government
180
170
160
Percent of GDP
150
140
130
120
110
100
90
80
70
60
50
1980
1985
1990
1995
2000
BIS Data
2005
2010
2015
2020
Credit & Unemployment in France
• Credit and Unemployment France…
Credit & Unemployment France (Correlation -0.6)
14
GFC
17.5
13
15
12
12.5
11
10
10
7.5
9
5
8
2.5
7
0
06
 2.5
5
1980
5
Credit
Unemployment
1985
1990
1995
2000
BIS And ILO Data
2005
2010
2015
4
2020
Percent of workforce
Percent of GDP per year
20