Transcript Document

The economics of consumption
1
Midterm
Grading will probably be ready for sections next week.
Midterm makeup:
- Bring your Dean’s excuse to the exam.
- We will schedule the exam for this week in the evening.
Meet after class today to schedule it.
2
Importance of consumption in macro
1. Consumption is two-thirds of GDP –
understanding its determinants is major part
of the ball game.
2. Consumption is the entire point of the
economy:
3. Consumption plays two roles in
microeconomics:
a. AD: It is a major part of AD in the short
run: recall IS curve in which
Y = C(Yd) + I + G + NX
b. AS: What is not consumed is saved and
influences national investment and economic
growth
3
Growth in C and GDP
Rate of growth per year
.08
.06
.04
.02
.00
Consumption
GDP
-.02
-.04
1970
1975
1980
1985
1990
1995
2000
2005
2010
4
The importance of fiscal policy today
When the economy is in a liquidity trap and recession, major available
policy tool is fiscal policy (remember IS-LM)
But, fiscal policy is controversial inside and outside economics:
Purchases:
- Controversial because increases size of government
- Long lags (recognition, decision, implementation)
- Infrastructure and other programs have long gestation periods.
Tax Cuts:
- One view: people will smooth consumption, and even anticipate a
future tax increase, and there will be little or no response.
- Other view: people are short-sighted and/or liquidity constrained,
and they will spend a substantial fraction of increased incomes
Deficit hawkism: Today, many economists and others worry about
impact of stimulus on government debt
Here is where we need to study carefully the economics of
consumption.
5
Alternative Theories of Consumption
The basic Keynesian insight is that consumption depends
fundamentally on personal income (“consumption function”)
This enters into the Keynesian models as C = α + βYd
On a closer look, a major puzzle: the short-run and cross-sectional
consumption functions looked very different from the long-term
consumption function.
6
Short-run v. Long-run Consumption Function
Mankiw, p. 499.
7
Alternative Theories of Consumption
The basic Keynesian insight is that consumption depends
fundamentally on personal income (“consumption function”)
This enters into the Keynesian models as C = α + βYd
On a closer look, a major puzzle: the short-run and cross-sectional
consumption functions looked very different from the long-term
consumption function.
There are four major approaches in macroeconomics:
*1. Fisher's approach: sometimes called the neoclassical model
2. Keynes original approach of the consumption function
*3. Life-cycle or permanent income approaches (Modigliani,
Friedman)
4.Rational expectations (Euler equation) approaches (Hall, Barro,...)
*We will sketch the life cycle model in class; Fisher in Mankiw and
section.
8
Consumption and Disposable Income
Real personal consumption expenditures
10,000
8,000
6,000
4,000
2,000
0
0
2,000
4,000
6,000
8,000
10,000
Real personal disposable income
9
Basic Assumptions of Life Cycle Model
Basic idea:
People have expectations of lifetime income; they determine their
consumption stream optimally; this leads consumers to “smooth”
consumption over their lifetime.
Assumptions:
“Life cycle” for planning from age 0 to D.
Earn Y per year for ages 0 to R.
Retire from R to D.
Maximize utility function:
D
max V(C1 , C 2 , ..., C D )   ( 1   ) z U( C z ), for ages z  0 to D.
z0
Budget constraint:
D
D
-z
(1

r
)
C

(1

r
)
Yz


z
-z
z0
z0
Discount rate on utility (δ) = real interest rate (r) = 0 (for simplicity)
10
Techniques for Finding Solution
1. Two periods:
max U(C1 )  U(C2 )  U(C1 )  U(Y1  Y2 - C1 )
{Cz }
Maximizing this leads to U’(C1)=U’(C2). This implies that C1 = C2 ,
which is consumption smoothing. The Cs are independent of
the Ys.
2. Lagrangean maximization (advanced math econ):
D
 D
-z
max L C1 ,...,CD  =  (1+ δ) U(C z ) + λ   (1+ r ) C z -  (1+ r )-z Yz

{C z }
z=0
z=0
 z=0
D
-z




Maximizing implies that U’(C1)=U’(C2)=-λ. This implies that Ct  C
which again is consumption smoothing independent of Y.
11
Initial Solution
C, Y, S
Diagram of Life
Cycle Model
Showing
Consumption
Smoothing
Income, Y
Consumption, C
Saving, S
|
0
R
age
|
D
12
Anticipated change in timing of income
C, Y, S
Income “splash” (Y’) with no W increase
Income, Y
Anticipated income
change of ΔY.
Because it is
anticipated, no
change in lifetime
income, so no
change in
(smoothed)
consumption. MPC
= 0; MPS = 1.
Consumption, C’=C
Saving, S’
|
R
0
age
|
D
13
Unanticipated change in permanent income
C, Y, S
Y’ =unanticipated increase; W increases.
Unanticipated windfall
of ΔY.
Leads to smoothing
the windfall over
remaining
lifetime.
(a) one time splash:
MPC = ΔY/(D-z).
For life
expectancy of 40
years, would be
MPC = .025.
(b) Permanent
income increase:
MPC = ΔY(Rz)/(D-z) = .6 to .8
Y
C’
C
|
R
0
age
|
D
14
Taxes and Consumption
1. Theory of temporary tax cuts:
– What is the impact if taxes are anticipated and paid back during
lifetime? No impact! MPC from taxes = 0.
– Barro (Ricardian) model extends this to future generations
2. Empirical estimates
– Actual evidence definitely shows substantial MPC (0.3 to 0.7)
– Evidence from random assignment of 2008 tax cut; MPC perhaps
0.5 in the first two quarters
3. Why discrepancy?
– Liquidity constraints on low-income
– “Behavioral economics”
15
Example of consumption smoothing: the 2008 tax rebate
Changes in C, DY, and S
800
600
C
DY
S
Estimated MPC= 0.46 (0.19)
400
200
0
-200
-400
06M01
06M07
07M01
07M07
08M01
08M07
16
17
Behavioral economics
Basic idea: That people are not optimizers (make mistakes)
Real-world examples for all of us:
- procrastination
- dealing with addictive substances
Why is it “behavioral”? Because lead to inconsistent
decisions that are regretted later
- bad grades, hangovers, addictions, drug wars
Examples from macroeconomics:
- MPC too high; low savings for retirement; subprime
mortgages; sticky housing prices; too high discount rate
18
Example of the Life Cycle Model at Work:
• How would the consumption and saving of people with volatile or
stable income streams look?
• See figure for Entrepreneur Ghates and Professor Nerd.
19
Major result of LCM: consumption smoothing
Y: Entrepreneur
Y: professor
C of both!
age
R
D
20
Taxes and Consumption
1. Theory of temporary tax cuts:
– What is the impact if taxes are anticipated and paid back during
lifetime? No impact! MPC from taxes = 0.
– Barro (Ricardian) model extends this to future generations
2. Empirical estimates
– Actual evidence definitely shows substantial MPC (0.3 to 0.7)
– Evidence from random assignment of 2008 tax cut; MPC perhaps
0.5 in the first two quarters
3. Why discrepancy?
– Liquidity constraints on low-income
– “Behavioral economics”
21
Further Extensions
Liquidity constraints
– Case of Yale students where income growing rapidly
– Here consumption is limited by borrowing constraint.
– Is this reason for MPC higher than life cycle prediction?
(Partially, but cannot explain response of non-constrained
consumers)
22
Further Extensions
3.Wealth effects:
Examples: How would you spend an unanticipated inheritance of
$1m? What is MPC of “trust-fund babies”? What would be the
effect of stock-market decline or housing bubble and burst?
- Life cycle model predicts that initial wealth (or surprise
inheritances) would be spread over life cycle.
• Intuition: an inheritance is just like an income splash.
- So the augmented life cycle model is
p
Ct = β0 + β1 Y t + β2 Wt
p
where Y t is permanent or expected labor income and Wt is
wealth.
23
What is the Effect of Stock Market
Booms and Busts on Consumption?
Price-earnings ratios, US
40
Roaring 20s
50
30
10
Roading 90s
20
0
1890 1900 1910 1920 1930 1940 1950 1960 1970 1980 1990 2000
24
The stock market, the housing market, and consumption
• Economists think that the bursting of the stock market bubble in
2000 or the housing market today contributed to recessions.
• Reasons? Decline in consumption (today) and investment (later)
• Rationale: the “wealth effect” on consumptoin
• Analysis in the life-cycle model:
p
– In augmented life-cycle model Ct = β0 + β1 Y t + β2 Wt
standard estimates are that β2 = .03 - .06 (example in a minute)
– Effect in the “Roaring 90s” and the housing crash today.
25
Regression
Dependent Variable: Real consumption expenditures
Method: Least Squares
Sample: 1960.1 2010.2
Variable
Coefficient
Std. Error
P
Real Disposable income
0.78
0.009
.0000
Real wealth
0.029
0.0014
.0000
R-squared
0.9993
26
Wealth and Consumption through Two Bubbles
Billions of 2005 $
16,000
12,000
500
Financial
bank
crisis
Tech
bubble
400
8,000
300
4,000
200
0
100
-4,000
0
-8,000
-100
-12,000
-200
Change in net worth (left scale)
Change in consumption (right scale)
-16,000
00
01
02
03
04
05
06
07
-300
08
09
10
27
Loss of wealth and savings rate increase
6.4
6
Net worth/personal income (<--)
6.0
Savings rate
5
(-->)
5.6
4
5.2
3
4.8
2
4.4
1
2005
2006
2007
2008
2009
2010
28
Key ideas
1.
2.
3.
4.
5.
6.
Consumption derived from consumer maximization
Pure model leads to consumption smoothing
All kinds of fun predictions
But impediments to pure model
Remember the wealth effect
Big open issue: how big is the short-run MPC?
29