Transcript (∆P/P) - (∆P/P) - University of Ottawa

Is the Private Economy SelfCorrecting or Should We Fear
Deflation? Examining the Experience
of the Great Depression and the
Current Great Recession
By
Mario Seccareccia
Professor of Economics, University of
Ottawa, Ontario, Canada
“Fear of Deflation”



From the early postwar years until the mid-1990s, fear of
deflation, which had been a principal concern of numerous
economists and policy makers of the interwar era of the
twentieth century (i.e. from J.M. Keynes to Irving Fisher),
had practically disappeared.
The rise of the postwar Keynesian welfare state with its
built-in stabilizers and the growth of “big” government
associated with a rising public sector, together with the
emergence of a strong trade union movement, all
suggested that such a fear was unfounded and that the real
danger was a new inflationary bias that was supposedly to
characterize the workings of Western capitalist economies
after W.W. II.
However, by the end of the 1990s, the potential for
deflation came once again to haunt economists and policy
makers alike.
Reasons for the Fear of
Deflation




Some have argued that this was the result of the success of monetary
policy of inflation targeting, coupled with “New Economy” which had
brought the inflation rate in most Western countries hovering ever
closer to zero (see, among others, the writings of De Long 1999,
DeLong and Sims 1999, Burdekin and Siklos, 2004, Raines and
Leathers 2008, Siklos 2009).
Others, instead, saw what had been happening to the Japanese
economy since the early 1990s as it lost vitality after a lengthy period
of unprecedented export-led growth together with the bursting of the
real estate bubble, and was suddenly faced with long-term stagnation
and falling prices (Bernanke 2002, Koo 2008).
Others, especially from the anti-globalization movement, feared the
consequences of unrestricted free trade, which was opening up and
integrating vast new regions of the world, such as China, as potential
new sources of cheap labor and very low priced tradable goods (Farrell
2004).
However, it is the major recession of 2008-2010 that accompanied the
worst financial crisis in postwar U.S. economic history, which solidified
the fear of deflation in the psyche of some present-day policymakers
as they observed some ominous signs during the crisis that deflation
can once again become a reality in the world economy (see Bernanke
2010).
Policy Makers’ Fear and Theoretical Models
of the Macro-Economy that Celebrate the
Virtues of Deflation in a Recession


In contrast, most theoretical models, which continue to be
taught by mainstream macroeconomists to both undergraduate
and graduate students using the standard AD/AS analysis,
preach the virtue of deflation as a market-clearing device in
both the product and labour markets. If ever the system
derails and falls into recession, all that is needed to fill the
recessionary gap is wage and price flexibility that will insure
eventually the system’s self correction.
All mainstream economists agree with its simple logic based on
a micro analysis of the macro economy. Largely oblivious to
these fears, the debate among mainstream academic
economists is not really over the self-correcting properties of
the price mechanism, since all agree that the private market
system is self-adjusting, but over the speed at which the
system self-adjusts, with New Keynesians emphasizing the
stickiness of wages and prices in the short run while New
Classicals do not.
Purpose of Present Analysis
 The object of this presentation is to provide an
analysis of how these legitimate fears of deflation
among policy makers can be better understood using
a simple framework of AD/AS that would be widely
understood by mainstream economists of the socalled New Consensus and to show its internal
contradictions.
 These competing analyses of the effects of deflation
are then used to shed further light on what happened
during the Great Depression of the 1930s as well as
during the current Great Recession of 2008-2010.
Simple New Consensus
Framework of Analysis



1. An aggregate demand relation that could explain both the level of output
or, as specified below, its deviation from potential output:
(y – y*) = γ1A + γ2r
[1]
where y is real output or income, y* is potential output (consistent with
some pre-specified NAIRU), A is real “autonomous” spending that is
inelastic to interest rate changes, r is the policy-determined real rate set by
the monetary authorities such as the “overnight” (or interbank) rate, and
the parameters γ1 > 0 and γ2 < 0 in accordance with received theory.
2. There is the usual aggregate supply (or inflation expectations-augmented
Phillips Curve) relation:
∆P/P = α1(y – y*) + α2(∆P/P)e
[2]
where ∆P/P is the rate of inflation, (∆P/P)e is the expected rate, (y – y*) is
as above the output gap, and α1>0 and α2>0.
3. There is also specified a Central Bank reaction function normally of the
Taylor rule variety as described below:
r = ρ0 + ρ1((∆P/P) - (∆P/P)T) + ρ2(y - y*)
[3]
T
where (∆P/P) is the central bank’s target rate with ρ0, ρ1, ρ2>0 and ρ1, ρ2 =
0.5. This model gave all the power to the monetary authorities to tinker
with the real rate of interest r, by keeping the economy at a zero output
gap consistent with a low target inflation rate of, say, 2%, as pursued in
Canada since 1991 (see Seccareccia and Lavoie 2010).
Figure 1: Graphical Integration of the First
Two Relations to Derive an AD Curve in
Inflation/Output Gap Space (Positive
Quadrants Only)
Graphical Analysis of the New
Consensus Approach



Let us now focus on the top right-hand quadrant, by depicting it in
Figure 2 for all possible values (positive and negative for y - y*
and ∆P/P) and adding to it the short-run Phillips Curve relation
described in equation (2) above.
Let us now imagine a situation in which the economy is initially at
price stability as depicted in Figure 2 and, suddenly, there is a
positive demand shock that pushes the inflation rate above the
target level associated with y-y* > 0.
In the New Consensus framework, the usual solution offered is
that, under the Taylor-type assumption about central bank
behavior, the central bank will raise real interest rates much above
the natural rate ρ0 not only because inflation has exceeded its
target level, (∆P/P) - (∆P/P)T> 0, but also because y-y* > 0 in
equation (3). This would therefore have the effect of tightening
monetary policy through higher real interest rates as shown in
Figure 2, consistent with a steady-state inflation with ∆P/P =
(∆P/P)e.
Figure 2: The Effect of a Positive
Aggregate Demand Shock
New Consensus Policy Analysis and
Inconsistency with Traditional
Neoclassical Analysis



In the typical story to be found in the New Consensus
literature, if the policy authorities wish to get the inflation rate
back to zero, one would need a strong dose of either fiscal or
monetary austerity (in the latter case perhaps by some shock
effect that can impact the γ2 coefficient in equation 1) that
would shift the AD down to the original level.
Alternatively, if inflationary expectations had also risen,
thereby shifting upward the Phillips Curve (not shown in Figure
2), the effect would be to reduce the difference between y and
y* automatically as the PC relation would rise upward along
the AD’ curve. But this would still require a similar strong dose
of fiscal/monetary policy to get the higher inflation rate down
to its previous level ∆P/P = (∆P/P)e at (y – y*) = 0.
While this analysis is of interest and points to the need for
discretionary policy, this tale is somewhat inconsistent with the
traditional neoclassical vision, since there is no market
mechanism in place to bring the inflation rate back to zero. All
of this requires a precise government policy mix in place.
Alternative Neoclassical
Laissez-Faire Mechanism
 All those hard-line neoclassical economists
uncomfortable with this New Consensus story could,
however, easily rely on a much simpler mechanism,
going back to A.C. Pigou, to insure that AD falls back
automatically to its original level.
 This is because one of the elements of “autonomous”
expenditures (A) in equation 1 is normally assumed to
be negatively affected by rising prices, owing to negative
wealth effects.
 This could be assumed to be the “self-adjusting” solution
compatible with not only the old textbook view but which
is also consistent with the opinion of certain laissez-faire
economists in recent times that have been critical of the
discretionary actions of the fiscal and monetary
authorities (see, among others, Safner 2010).
Application of the Automatic
Mechanism in a Great Depression



But how would these policy tools or automatic mechanisms
just described work themselves out in an extraordinary
situation like the Great Depression in which a vicious cycle
of deflation had already taken hold and where the “fear of
deflation” had become a reality?
As shown in Figure 3, the self-adjusting mechanism ought
to work itself out smoothly as the positive wealth effect of
the deflation brings the economy back to its natural level
where y - y* = 0.
Indeed, if the expectations of deflation are sufficiently
strong so as to shift down also the short-run Phillips Curve,
this would only speed up the process of self correction. In
fact, the system may even react to such an extent that
there is overshooting, as the AD curve moves outward
towards AD”, but the inflation will then reverse the shift and
bring the economy back to a situation of price stability with
y = y*.
Figure 3: Traditional Pigouvian Mechanism of
Adjustment in a Deflationary Environment
A Fly in the Ointment: Questioning the
Traditional Neoclassical Story
 As much as this is a logically consistent analysis that
questions the validity of those, even among the New
Consensus, who have justified short-term fiscal policy
measures in a situation of deflation, there are a
number of problems.
 Firstly, the shape of AD relation would actually be
different than that shown in Figure 3 above. The
assumed symmetry of the shape (when there is
inflation or deflation) that is depicted in the diagram
above is problematic. This is so for at least three
reasons.
First Problem




The first reason has been sufficiently discussed even in the
mainstream literature since it could reverse the slope of the AD
curve in a situation of deflation. This is because monetary policy
becomes completely ineffective as real interest rates are most
likely to be rising with falling prices as a result of the lower bound
of zero for nominal interest rates.
Moreover, this is undoubtedly an important reason why central
banks fear deflation, since it would take away their only effective
instrument of policy.
As depicted in the Figure 4 below, in such as situation of deflation,
the slope of the AD curve in inflation/output space is reversed. As
prices continue to fall in a downward spiral, real interest rates
would rise and, therefore, reduce further AD.
The perversity of this effect would, of course, depend on the
elasticity of the components of AD to real interest rates, but, even
in the best case scenario in which the interest elasticity of AD is
zero, the shape of the AD curve would become vertical in an
economy faced with a state of deflation.
Figure 4: Effect of Deflation in an Economy
Reaching the Lower Bound of Nominal
Interest Rate
The Resilience of the Neoclassical
Model to the New Consensus Critique




As shown in Figure 4 this does not, however, leave the New
Consensus economists off the hook. Even in this case, there is a
problem with the New Consensus analytics in dealing with the
traditional argument that the positive wealth effects resulting from
the falling prices will eventually move the AD curve back towards a
situation in which y = y*.
Indeed, even in the case in which the short-run Phillips curve does
not spiral downward as with PC shown above, the whole AD
relation will inevitably be shifting rightward toward where ∆P/P =
(∆P/P)e at (y – y*).
If, instead, prices are shocked into an accelerated downward
spiral, as with the downward-shifting PC’, then the adjustment
process merely speeds up and could even lead to overshooting, as
previously discussed. But, once again, the inevitable consequence
is for the private economy to correct itself through an automatic
mechanism of adjustment.
In both cases, the powerful wealth effects that are assumed by
traditional theory make the latter highly resilient to the critical
analysis coming from the New Consensus that has recently opened
the door for the use of discretionary fiscal policy.
The Resilience of the Neoclassical
Model to the New Consensus Critique



This story, which rests on the importance of wealth effects,
has been devastating for those New Keynesian economists
who have pointed to the incapacity of monetary to be
conducted in a deflationary environment, thereby
necessitating discretionary fiscal policy.
Consequently, it has provided since the 1940s a powerful
argument that continues to serve to inoculate mainstream
economists against fears of deflation and it remains the
Achilles’ heal of those New Keynesian economists who have
tried to revive fiscal policy as a legitimate policy tool in the
mainstream literature.
Indeed, literally since the 1940s and 1950s, we have seen,
for instance, the resurrection of consumption functions that
have traditionally emphasized these positive wealth effects
from Patinkin to Friedman and Modigliani (Bodkin, 2010).
Second Problem: Kaleckian
Critique




Starting with Kalecki (1944), post-Keynesian economists
have questioned this analysis on both logical and empirical
grounds.
These critics of the mainstream view argue that, while
consumption may well be affected by positive wealth
effects, it could also be affected by negative Fisher debt
effects (Lavoie 2010).
These debt effects, in a period of deflation, may become
significant in not only offsetting the positive wealth effects
on the consumption function, but even more significantly it
could badly hamper real investment because of problems of
insolvency and negative expectations of return.
This would have the effect of shifting the AD curve further
away from where y = y*. A spiraling deflation with a
downward shifting Phillips Curve would only compound the
effects already resulting from the initial deflation.
Second Problem: Kaleckian
Critique




Hence, in terms of the first New Consensus equation (1), not only
would the first term on the right-hand side have to be revised
because of the significance of real wealth, W/P, as an element in
the determination of A; but one would have to include also the
real debt burden, D/P, with ∂A/∂W/P > 0 and ∂A/∂D/P < 0, as
discussed in Lavoie (2010) and in Baumol, Blinder, Lavoie, and
Seccareccia (2010).
Indeed, from equation (1’), during periods of inflation it may well
be that ∂A/∂W/P > ∂A/∂D/P, even though the effect of inflation
on investment may not be neutral for reasons articulated by
writers such D.H. Robertson in the 1920s.
However, for an economy in a depressive state of deflation, it
would most likely face a situation in which ∂A/∂W/P <∂A/∂D/P
because of the significance of this balance-sheet effect, as
emphasized, among others, by Koo (2008).
(y – y*) = γ1A(W/P, D/P) + γ2r
[1’]
Second Problem: Kaleckian
Critique
 It follows from this that a deflationary episode, as
depicted in Figure 5, would lead to a cumulative
process away from y = y* as the economy plunges
ever further into stagnation.
 The downward expectations of deflation would merely
feed further this negative cumulative process as the
PC relation shifts downward with the private economy
spiraling into the abyss of a Great Depression.
 Without some other counterforce in place, in a world
in which monetary policy is impotent, the only
effective tool to stop this downward spiral is through
strong fiscal policy actions of the New Deal variety.
Figure 5: The “Fear of Deflation” Realized, in
an Economy Characterized by a Kaleckian
Cumulative Negative Process of Deflation
Third Problem: Real Wage as a
Distribution Parameter




Keynes (1936) had argued that, while these instabilities are a
characteristic feature of the capitalist process, such economies
are not violently unstable with certain forces also working to
stabilize the economy.
But what forces other than the negligible positive wealth
effects emphasized by neoclassical economists could serve as
countervailing force against such Kaleckian downward
pressures?
Clearly, for Keynes and Post-Keynesians, consumption
spending is not only affected by real disposable income which
would be falling sharply during a depression, but also by the
distribution of this income over time.
One such key income distribution variables is the evolution of
the real wage itself, ω/P, which normally tends to work in
reverse to what is depicted in the neoclassical conception of
the labor market. This would suggest that the real wage, ω/P,
or some other such distribution variable, ought also to be
considered a critical determinant of A, as depicted in equation
(1”) below.
Third Problem: Real Wage as
Distribution Parameter




The real wage, ω/P ought to be considered a critical determinant
of A, as depicted in equation (1”) below for reasons made clear by
Kalecki, Kaldor and Pasinetti, who had postulated differential
propensities to consume according to income groups, with the
propensity to consume out of labor income being significantly
higher than that from property income.
In contrast to neoclassical theory, which would predict an overall
increase in employment and output resulting from a fall in the real
wage for standard microeconomic reasoning pertaining to the
neoclassical conception of the labor market, Keynesians and PostKeynesians would argue that a fall in the real wage would tend to
exacerbate the situation in the labor market because of the
downward consequences on the product market.
Hence, in terms of equation (1”) below, the relation between the
real wage and private spending would normally be a positive one
unless perhaps the domestic economy is overly dominated by
foreign trade, i.e. ∂A/∂ω/P > 0.
(y – y*) = γ1A(W/P, D/P, ω/P) + γ2r
[1”]
Third Problem: Real Wage as
Distribution Parameter
 From this reasoning, it follows that the reaction of real
wages to a downward aggregate demand spiral is
important and could, therefore, either intensify the
depression or mitigate its effects.
 Indeed, if the traditional neoclassical prescription of
cutting wages actually brings about a decline in the
real wage, a falling real wage would merely aggravate
the situation.
 However, these are all empirical questions to which I
would like to turn by looking at some stylized facts
about the Great Depression of the 1930s and the
recent Great Recession by focusing on the experience
of the United States and Canada for which some data
was readily available.
Are Depressions Self-Correcting? A Brief
Analysis of the Experience of the 1930s and
the Current Great Recession





While there are obvious similarities between the Great Depression
of the 1930s and the Current Great Recession, the current
recession has not degenerated into a serious deflationary episode
characterized by rising real interest rates and falling money wages
and prices.
As argued by Desmedt, Piégay and Sinapi (2010), both crises have
their roots in problems of rising inequality and bubbles in asset
markets.
However, the recent crisis was triggered by problems in the
housing market (starting in 2007) which was then transmitted to
the banking and financial markets (in 2008) and eventually to the
real economy (by 2009).
During the Great Depression, the sequence was somewhat in
reverse, which problems beginning in the stock market, which then
led to the collapse of the banking sector and then had
ramifications in the real economy, including the housing market.
However, by any measure the Great Depression was more severe
than the Great Recession.
120
130
110
120
Index (1926 = 100)
Index (1926 = 100)
Figure 6: Fluctuations in Employment
and Real Gross Domestic Product,
United States and Canada 1926 - 1939
100
90
110
100
80
90
70
80
26 27 28 29 30 31 32 33 34 35 36 37 38 39
US Employment Index
US Real GDP Index
Source: S.B. Carter et al. (eds), Historical Statistics of the United States
(Millenium Edition), New York: Cambridge University Press, 2006. Vol. 2.
Table Ba840, p. 12-112; and Vol. 3, Table Ca9, p. 3-25.
26 27 28 29 30 31 32 33 34 35 36 37 38 39
Canada Employment Index
Canada Real GDP Index
Source: Statistics Canada, CANSIM I, Series Label D14442;
and Statistics Canada, Historical Statistics of Canada (Second Edition),
Ottaw a: Statistics Canada, 1983, Series D528.
180
130
160
120
Index (1926 = 100)
Index (1926 = 100)
Figure 7: Movement in Average Hourly Wage
Rates, the Consumer Price Index, and Average
Real Wages, United States (Mfg.) and Canada
(Industrial Composite) 1926 - 1939
140
120
100
110
100
90
80
80
60
70
26 27 28 29 30 31 32 33 34 35 36 37 38 39
US Average Hourly Earnings
US Consumer Price Index
US Real Average Hourly Earnings
Source: S.B. Carter et al. (eds), Historical Statistics of the United States
(Millenium Edition), New York: Cambridge University Press, 2006, Vol. 2,
Table Ba4396, p. 2-281; and Vol. 3, Table Cc1, p. 3-158.
26 27 28 29 30 31 32 33 34 35 36 37 38 39
Canada Average Wage Rate (General)
Canada Consumer Price Index
Canada Real Wage Rate Index
Source: Statistics Canada, Historical Statistics of Canada (Second Edition),
Ottawa: Statistics Canada, 1983, Series E198 and K8.
Figure 8: Evolution of Saving Rate,
United States and Canada 1926 - 1939
7
12
6
8
5
4
Per Cent
Per Cent
4
3
2
1
0
-4
0
-8
-1
-2
-12
26 27 28 29 30 31 32 33 34 35 36 37 38 39
US Saving Rate
Source: S.B. Carter et al. (eds), Historical Statistics of the United States
(Millenium Edition), New York: Cambridge University Press, 2006,
Vol. 3, Table Ce122, p. 3-312.
26 27 28 29 30 31 32 33 34 35 36 37 38 39
Canada Saving Rate
Source: Statistics Canada, Historical Statistics of Canada (Second Edition),
Ottawa: Statistics Canada, 1983, Series F83 and F90.
Broad Analysis of Charts on the
Great Depression (1929-1939)




1. Collapse in AD, employment and output.
2. Although prices were already gently falling the late 1920s,
the collapse in AD triggers a sharp fall in prices and wages.
Since wages fall less than prices, real wages rise significantly,
despite the growing unemployment.
3. The initial collapse in the saving rate can be explained by
two dominant effects that have little to do with the importance
of wealth effects: (1) there is the traditional Duesenberry
effect where household losing jobs are trying to maintain
consumption levels by drawing down their savings, (2) there is
the distribution effect of a rising real wage, ω/P, which is
initially sustaining consumption demand, despite the fall in
employment resulting from the decline in business spending.
4. The steep rise in the saving rate after 1933 is primarily a
balance sheet phenomenon. Indeed, once employment
stabilized and began to turn around and then reinforced by
New Deal policies, household began to reduce their debt load
in face of still continued uncertainty about income growth.
108
110
106
108
Index (2004 = 100)
Index (2004 = 100)
Figure 9: Fluctuations in Employment
and Real Gross Domestic Product,
United States and Canada 2004 - 2010
104
102
100
98
2004
106
104
102
100
2005
2006
2007
2008
2009
2010
US Employment (Nonfarm) Index
US Real GDP Index
Source: U.S. Bureau of Labor Statistics, Establishment Data, Historical Employ ment B-1;
U.S. Department of Commerce, Bureau of Economic Analysis, Selected NIPA Tables, Table C1.
98
2004
2005
2006
2007
2008
2009
2010
Canada Employment Index (Total, All Industries)
Canada Real GDP (Chained 2002 Dollars)
Source: Statistics Canada, CANSIM Series V2522952 and CANSIM Series V1992067.
Figure 10: Movement in Hourly Earnings, the
Consumer Price Index, and Real Hourly
Earnings, United States and Canada 2004-2010
124
116
120
116
Index (2004 = 100)
Index (2004 = 100)
112
112
108
104
104
100
100
96
2004
108
2005
2006
2007
2008
2009
2010
US Real Hourly Earnings (Total Private Sector)
US Hourly Earnings
US Consumer Price Index
Source: U.S. Department of Labor, Bureau of Labor Statistics, Establishment Data,
Historical Hours and Earnings, B-2, and Consumer Price Index, All Items.
96
2004
2005
2006
2007
2008
2009
2010
Canada Real Average Hourly Earnings Index
Canada Average Hourly Earnings Index
Canada Consumer Price Index
Source: Statistics Canada, CANSIM Series V1806037 and V41693271.
Figure 11: Evolution of Saving Rate,
United States and Canada 2004-2010
6
5.0
5
4.5
Per Cent
Per Cent
4.0
4
3
3.5
3.0
2
1
2004
2.5
2005
2006
2007
2008
2009
2010
US Saving Rate
Source: U.S. Department of Commerce, Bureau of Economic Analysis,
Tables from Personal Income and Outlays.
2.0
2004
2005
2006
2007
2008
2009
2010
Canada Saving Rate
Source: Statistics Canada, CANSIM Series V647038.
Broad Analysis of Charts on the
Great Recession (2007-2010)




1. Collapse in AD, employment and output triggered initially by the
subprime crisis in 2007 and then followed by the widespread financial crisis
in 2008 .
2. Although prices had been gently rising during the early-to-mid 2000s,
the collapse in AD triggers a disinflation in prices, but without making any
significant dent on the upward trend in nominal wages. Hence, despite the
significant fall in employment and rise in unemployment, real wages
continued to grow, much as they had done during the Great Depression,
but, in the recent case, as a result of disinflation (not a deflation) in prices
but with only a small observable disinflation in nominal wages.
3. Unlike the initial collapse in the saving rate during the Great Depression,
there was no fall in the saving rate. Given the weaker fall in employment
and output during the Great Recession, the traditional Duesenberry effect
did not take hold, as a result of the strong distributional effect of a rising
real wage, ω/P.
4. The sharp rise in the saving rate that started immediately with the
collapse of the housing market and fears of household indebtedness out of
control reflects primarily a Kaleckian balance sheet phenomenon. Indeed,
while this phenomenon of rising personal saving rate began just before the
subprime crisis, households have been consistently trying to reduce their
debt load in face of continued uncertainty about income prospects.
Conclusion: Is the Private Economy
Self-Correcting?
 As stated in the previous analysis, the only
self-correcting property of a deflationary
episode is its positive distributional effect
on the real wage, which could sustain some
increased consumption demand despite the
fall in employment.
 The self-destructive property of a
deflationary episode was highlighted long
ago by Fisher and Kalecki and these effects
dominated not only the experience of the
Great Depression after 1933, but also the
current Great Recession.
SEE SEPARATE CHARTS BELOW
 Important: See Figure 14 for
separate series on total labor
compensation (composite measure)
in the US during the Great
Depression.
Figure 12: Fluctuations in Employment
and Real Gross Domestic Product,
United States 1926 - 1939
120
Index (1926 = 100)
110
100
90
80
70
1926
1928
1930
1932
1934
Employment (Nonagricultural)
1936
1938
Real GDP
Source: S.B. Carter et al. (eds), Historical Statistics of the United States
(Millenium Edition), New York: Cambridge University Press, 2006. Vol. 2.
Table Ba840, p. 12-112; and Vol. 3, Table Ca9, p. 3-25.
Figure 13: Fluctuations in Employment
and Real Gross Domestic Product,
Canada 1926 - 1939
130
Index (1926 = 100)
120
110
100
90
80
1926
1928
1930
1932
1934
1936
1938
Employment Index (Industrial Composite)
Real Gross Domestic Product
Source: Statistics Canada, CANSIM I, Series Label D14442;
and Statistics Canada, Historical Statistics of Canada (Second edition),
Ottawa: Statistics Canada, 1983, Series D528.
Figure 14: Movement in Hourly Earnings, Total
Compensation, the Consumer Price Index, and
Real Hourly Earnings and Real Total
Compensation, United States 1926 - 1939
180
Index (1926 = 100)
160
140
120
100
80
60
1926
1928
1930
1932
1934
1936
1938
Hourly Earnings in Mfg
Total Compensation (All Industries)
Consumer Price Index
Real Hourly Earnings in Mfg
Real Total Compensation
Source: S.B. Carter et al. (eds), Historical Statistics of the United States
(Millenium Edition), New York: Cambridge University Press, 2006. Vol. 2.
Table Ba4396, p. 2-281 and Table Ba4419, p. 2-283; and Vol. 3, Table Cc1, p. 3-158.
Figure 15: Movement in Average Wage Rates
(Industrial Composite), the Consumer Price
Index, and Average Real Wages, Canada
1926 - 1939
130
Index (1926 = 100)
120
110
100
90
80
70
1926
1928
1930
1932
1934
1936
1938
Real Wages
Money Wages (Total)
Consumer Price Index
Source: Statistics Canada, Historical Statistics of Canada (Second Edition),
Ottawa: Statistics Canada, 1983, Series E198 and K8.
Figure 16: Evolution of Saving
Rate, United States 1926 - 1939
7
6
5
Per Cent
4
3
2
1
0
-1
-2
1926
1928
1930
1932
1934
1936
1938
SAVINGRATE
Source: S.B. Carter et al. (eds), Historical Statistics of the United States
(Millenium Edition), New York: Cambridge University Press, 2006, Vol. 3, Table Ce122, p. 3-312.
Figure 17: Evolution of Saving
Rate, Canada 1926 - 1939
12
8
Percent
4
0
-4
-8
-12
1926
1928
1930
1932
1934
1936
1938
Saving Rate
Source: Statistics Canada, Historical Statistics of Canada (Second Edition),
Ottawa: Statistics Canada, 1983, Series F83 and F90.
Figure 18: Fluctuations in Employment
and Real Gross Domestic Product,
United States 2004 - 2010
108
Index (2004 = 100)
106
104
102
100
98
2004
2005
2006
2007
Employment (Nonfarm)
2008
2009
2010
Real GDP
Source: U.S. Bureau of Labor Statistics, Establishment Data, Historical Employment B-1;
U.S. Department of Commerce, Burea of Economic Analysis, Selected NIPA Tables, Table C1.
Figure 19: Fluctuations in Employment
and Real Gross Domestic Product,
Canada 2004 - 2010
110
Index (2004 = 100)
108
106
104
102
100
98
2004
2005
2006
2007
2008
2009
2010
Employment Index (Total, All Industries)
Real GDP (Chained 2002 Dollars)
Source: Statistics Canada, CANSIM Series V2522952 and CANSIM Series V1992067.
Figure 20: Movement in Hourly Earnings, the
Consumer Price Index, and Real Hourly
Earnings, United States 2004-2010
124
Index (2004=100)
120
116
112
108
104
100
96
2004
2005
2006
2007
2008
2009
2010
Hourly Earnings (Total Private Sector)
CPI
Real Hourly Earnings
Source: U.S. Department of Labor, Bureau of Labor Statistics, Establishment Data,
Historical Hours and Earnings, B-2, and Consumer Price Index, All Items.
Figure 21: Movement in Hourly Earnings, the
Consumer Price Index, and Real Hourly
Earnings, Canada 2004-2010
116
Index (2004 = 100)
112
108
104
100
96
2004
2005
2006
2007
2008
2009
2010
Real Average Hourly Earnings Index
Average Hourly Earnings Index
Consumer Price Index
Source: Statistics Canada, CANSIM Series V1806037 and V41693271.
Figure 22: Evolution of Saving
Rate, United States 2004-2010
6
Per Cent
5
4
3
2
1
2004
2005
2006
2007
2008
2009
2010
Saving Rate
Source: U.S. Department of Commerce, Bureau of Economic Analysis,
Tables from Personal Income and Outlays.
Figure 23: Evolution of Saving
Rate, Canada 2004-2010
5.0
4.5
Per Cent
4.0
3.5
3.0
2.5
2.0
2004
2005
2006
2007
2008
2009
2010
Saving Rate
Source: Statistics Canada, CANSIM Series V647038.