How Rising Gas Prices Hurt American Households

No. 2162
July 14, 2008
How Rising Gas Prices Hurt
American Households
Karen A. Campbell, Ph.D.
The upward march of retail gasoline prices has
affected U.S. households regardless of whether their
members drive, take public transportation, or walk. In a
Talking Points
modern economy, the interdependency created by sup-
• Higher gas prices lower employment,
plying specialized labor and trading for all other goods
income, and spending.
and services produced by other people leaves virtually
• Households initially tap their savings to pay
no one unaffected by the price of gas at the pump.
the higher prices.
Analysts at The Heritage Foundation recently
• Consumption growth slows as households
examined how going from $3 and $4 retail to $5 and
adjust to higher gas prices.
$6 retail per gallon of gasoline would affect the U.S.
• Slower growth in consumption and higher
economy. If prices continue to rise at an accelerated
fuel prices for businesses result in lower
pace over the course of a year:1
employment.
• Total employment would decrease by 586,000 jobs,
• Employment is also decreased by workers
• Disposable personal income would decrease by
trying to change their work patterns to avoid
$532 billion,
burdensome commutes.
• Personal consumption expenditure would decrease
• Lower employment slows the growth in
by $400 billion, and
income, which means slower economic
growth over time.
• Personal savings would be spent to help pay the cost.
• Markets are signaling a need for more supply.
What the Numbers Mean
Businesses are trying to respond by making
investments in refining capacity and finding new
Table 1 shows what these numbers mean for
energy sources. This should be encouraged.
three representative households’ income, consump-
tion, and saving patterns. The first column is the
• Deficit reduction and inflation vigilance are
actual data from the 2006 Bureau of Labor Statistics
needed to strengthen the dollar and ease
pressure on prices and speculation in com-
Consumer Expenditure Survey.2 The simulated
modity markets.
impact is in the second column for each type of
household, and the third column shows the dollar
This paper, in its entirety, can be found at:
www.heritage.org/Research/Economy/bg2162.cfm
loss for households.
Produced by the Center for Data Analysis
The estimate is a best case in that mortgage and
Published by The Heritage Foundation
interest payments remain constant. More likely,
214 Massachusetts Avenue, NE
Washington, DC 20002–4999
(202) 546-4400 • heritage.org
Nothing written here is to be construed as necessarily reflect-
ing the views of The Heritage Foundation or as an attempt
to aid or hinder the passage of any bill before Congress.

---

No. 2162
July 14, 2008
How Rising Gas Prices Will Affect Households
Projections are based on the price of gas increasing by $2 per gallon. Income and expenditure ? gures are median values from 2006,
the most recent data available.
Husband and Wife
Married, 2 Children,
Single
Oldest Child Age 6–17
Gas Price
Change in
Gas Price
Change in
Gas Price
Change in
Actual
Actual
Actual
Effect
2008
Effect
2008
Effect
2008
Disposable
$69,350.00 $68,483.13
–$866.88
$86,807.00 $85,721.91 $1,085.09
$37,795.00 $37,322.56
–$472.44
personal income
Income after
paying mortgage
$65,468.78 $64,601.91
–$866.88
$79,934.71 $78,849.62 –$1,085.09
$35,648.05 $35,175.61
–$472.44
and interest
Average annual
purchases of
$55,631.38 $55,130.70
–$500.68
$69,157.47 $68,438.23
–$719.24
$33,996.55 $33,690.58
–$305.97
goods and services
Personal savings
$9,837.40
$9,082.89
–$754.51
$10,777.24
$9,984.52
–$792.72
$1,651.50
$1,424.15
–$227.35
Source: Bureau of Labor Statistics Consumer Expenditure Survey 2006, at http://www.bls.gov/cex/2006/stnderror/cucomp.pdf. Effects are calculations made by
analysts at The Heritage Foundation.
Ta
T ble 1 • B 2162
ble 1 • B 2162
heritage
heritage.org
.org
increased borrowing and less saving will result in
Forecasted Price of Retail Gas
higher interest payments, constraining spending
and decreasing the savings of households yet
$5.50
$6.00
more. It also does not show the increased likeli-
hood that a member of the household will be
5.00
Projected
unemployed.12
4.00
Chart 1 illustrates the baseline gas price fore-
cast and the higher gas price simulation. The
3.00
effect of gas prices operates directly and indirectly.
Baseline
Chart 2 shows the effect on employment. Both the
2.00
$3.15
$3.10
demand for labor and the supply of labor are neg-
atively affected, and this lowers overall employ-
1.00
ment. On the demand side, businesses rely
heavily on transportation to get their goods and
0
1Q
2Q
3Q
4Q
1Q
2Q
3Q
4Q
1Q
services to the consumer. Many suppliers have
2008
2008
2008
2008
2009
2009
2009
2009
2010
their own fleets; others, who outsource their
Source: Data for the employment series are from the Bureau of Labor
transportation service, must pay higher costs for
Statistics Employment series. Personal Disposable Income and
this service. Higher costs along with decreased
Expenditure series are from the Bureau of Economic Analysis National
Personal Income and Outlays Accounts. Retail gas price data are from
consumer purchases will cause businesses to cut
the Energy Information Administration. Heritage Foundation forecast
back on jobs. The decrease in employment is not
simulation uses RATS econometric software.
entirely attributable to labor demand, though;
Chart 1 • B 2162
heritage.org
labor supply may also decrease. Individuals with
1. Dollar prices are in nominal terms.
2. U.S. Department of Labor, Bureau of Labor Statistics, Consumer Expenditure Survey, 2006, at http://www.bls.gov/cex/2006/
stnderror/cucomp.pdf (July 3, 2008).
page 2

---

No. 2162
July 14, 2008
results in more losses to savings. As prices con-
High Gas Prices Will Lead to Job Losses
tinue to rise, consumers do adjust their spending.
This can be seen by the slower growth in spending
If the cost of a gallon of gasoline increases by $2, total
as compared to the baseline. The overall effect
employment in the U.S. will decrease by 2.1 million in 2010.
after just two years can be seen by the gap
Millions of Jobs
between the baseline and the simulation with
139.6
140
higher gas prices.
The rise in energy prices at a time when food
prices and other commodity prices are rising may
139
Baseline
solicit a monetary policy response to fight inflation-
137.9
ary pressures. Although this effect was not included
138
Price
in the analysis, this would further increase interest
Increase
rates and constrain the pocketbooks of businesses
Effect
and households. However, if this policy sends a sig-
137
nal that the Fed is once again targeting inflation, this
137.5
may go a long way to ease pressure on commodity
prices and strengthen the U.S. dollar. Both of these
136
1Q
2Q
3Q
4Q
1Q
2Q
3Q
4Q
1Q
two effects would serve to ease pressure on prices.
2008
2008
2008
2008
2009
2009
2009
2009
2010
Source: Data for the employment series are from the Bureau of Labor
Conclusion
Statistics Employment series. Personal Disposable Income and
Americans are now facing the prospect of even
Expenditure series are from the Bureau of Economic Analysis National
Personal Income and Outlays Accounts. Retail gas price data are from
higher prices at the pump. While there are many
the Energy Information Administration. Heritage Foundation forecast
simulation uses RATS econometric software.
other economic influences on household expendi-
ture, personal savings, personal disposable income,
Chart 2 • B 2162
heritage.org
and total employment, the Heritage analysis simu-
lated the dynamic movement of these variables in
response to movements in the retail price of gasoline.
long commutes may decide to look for other jobs
The results of the analysis show that house-
closer to home or give up working altogether.
holds react by using personal savings in the short
Chart 3 shows the effect on other household
term. This reduction in assets slows other spend-
variables. Households tap into personal savings
ing, leading to slower growth in purchasing.4 On
immediately to pay for higher fuel costs because
the supply side, businesses experience higher pro-
personal consumption expenditures (buying hab-
duction costs while demand for their goods is
its) are not adjusted downward as fast as real3 dis-
lower, causing them to adjust their employment
posable income is decreased. Disposable income
downward. Individuals, too, may begin to adjust
decreases almost immediately. Growth in income
their work choices as longer commutes make
rebounds but then adjusts to a slower rate as the
working outside the home less beneficial. These
feedback from job losses begins to drag it down.
two effects reduce overall employment.
Consumption expenditure is then further
There is a feedback effect between employment
crimped by decreased disposable income, and
and personal disposable income. After a sharp
higher interest payments from increased borrow-
decrease, disposable income starts to grow, but
ing start to crowd out other purchases. This
this growth is quickly slowed by the loss of jobs.
3. Real variables are adjusted for inflation and therefore are a measure of what people can “really” purchase with their income
(purchasing power).
4. At these prices, it is arguable that the elasticity of demand for gasoline is higher, which should reduce expenditures. The
income effect of the higher prices also has a ripple effect throughout the economy.
page 3

---

No. 2162
July 14, 2008
Higher prices signal the need for
Higher Gas Prices Will Affect the Entire Economy
more oil. Businesses are attempting to
respond to that need by finding new
If the cost of a gallon of gasoline increases by $2, personal expenditures and
disposable income will be affected. The charts below show projections, by
reserves to drill and increasing invest-
quarter, through the first quarter of 2010:
ment in refining capacity. The high
price also signals entrepreneurs to
Personal Consumption
Disposable Personal
look for new, more efficient ways to
Expenditures
Income
supply the energy needs of consum-
In Billions of Dollars
In Billions of Dollars
ers.
10,500
11,000
Baseline
Baseline
To the extent that the high price
10,400
10,900
is not a clear signal because of ex-
10,800
10,300
cessive taxes and regulations that
Price
10,700
artificially make oil scarcer, the gov-
10,200
Increase
Price
Effect
10,600
ernment can implement policies that
Increase
10,100
10,500
Effect
would allow for more oil production.
The government can also encourage
10,000
10,400
1
2
3
4
1
2
3
4
1
1
2
3
4
1
2
3
4
1
innovations in energy supply by
2008
2009
’10
2008
2009
’10
keeping regulatory burdens to a
Source: Data for the employment series are from the Bureau of Labor Statistics
minimum.
Employment series. Personal Disposable Income and Expenditure series are from the Bureau
of Economic Analysis National Personal Income and Outlays Accounts. Retail gas price data
Congress should focus on rein-
are from the Energy Information Administration. Heritage Foundation forecast simulation
ing in spending and reducing the defi-
uses RATS econometric software.
cit. This would serve to lower long-
Chart 3 • B 2162
heritage.org
term interest rates and strengthen
the U.S. dollar, which would help to
The overall result after just two years is lower
ease the pressure on prices. Mone-
employment, lower disposable income, and lower
tary policy that is expected to be inflation-fighting
overall consumption of goods and services.
could also aid in stabilizing prices and curb the
current flight to commodities.
This is purposefully a short-term forecast. The
U.S. economy is known for its innovative responses
—Karen A. Campbell, Ph.D., is Policy Analyst in
to economic scarcity. Markets that are unconstrained
Macroeconomics in the Center for Data Analysis at
by excessive regulation can give a clear price signal.
The Heritage Foundation.
page 4

---

No. 2162
July 14, 2008
APPENDIX
DATA, METHOD, AND RESULTS
Analysts used Vector Autoregression (VAR) anal-
shows that after its own lags, retail gas accounts for
ysis to decompose the historical data and conduct a
much of the variance in personal consumption
dynamic forecast simulating a two-dollar increase in
expenditure over the longer term (22 percent). The
the retail price of gasoline. Vector Autoregression is
analysis also shows that over time, rather than
a tool now widely used by economists to gain
diminishing, gas prices have an increasing effect on
insights into the dynamic interactions of historical
all three variables.
data, measure the impact on other variables to a
A forecast from the second quarter of 2008 to the
shock in one of the variables, and make economic
first quarter of 2010 was generated to establish a
forecasts.5
baseline. The baseline predicted an average retail
This analysis uses the quarterly series for retail
gas price of $2.61 in the second quarter of 2008,
gasoline prices (GAS) from the International Energy
which is slightly higher than the average price in the
Agency. Real disposable personal income (DPI) and
first quarter of 2008 of $2.58 (in real terms). A
real personal consumption expenditure (PCE) are
shock in gas prices was implemented via a price
from the Bureau of Economic Analysis, and total
path that increases the difference in gas prices by 40
employment (TE) is from the Bureau of Labor Sta-
percent in the first two quarters, 5 percent in the fol-
tistics. The data run from the first quarter of 1980 to
lowing quarter, and 10 percent in the quarter after
the first quarter of 2008. The log of the data was first
that. The remaining quarters were not “shocked.”
differenced to achieve stationarity,6 resulting in 107
This resulted in a forecast of average gas prices in
usable observations. The series were in real terms.
the second quarter of 2008 of $3.90 and climbing to
The following model was estimated.7 Difl is the
$4.58 (in real terms).
logged difference of the variable. J = {1,2,3,4}.8
In order to estimate the nominal prices, the real
Diflgas(t) = constant + diflgas(t ) + diflgas(t-j)
gasoline prices were adjusted by a deflator of 1.2.
This was the deflator used by the Bureau of Eco-
Diflpce(t) = constant + diflgas(t ) + diflgas(t-j) +
nomic Analysis to adjust the consumption and
diflpce(t-j) + difldpi(t+1-j)
income series. Arguably, this is a best-case scenario
Difldpi(t) = constant + diflgas(t ) + diflgas(t-j) +
of the differences in the baseline versus forecast
diflte(t-j)+ difldpi(t-j)
because the same rate of inflation is assumed for
Diflte(t) = constant + diflgas(t ) + diflgas(t-j)+
both. To the extent that higher gas prices help to
diflte(t-j) + diflpce(t-j)
fuel inflation, the gap between the baseline and
forecasted amounts will be wider.
Factor variance decomposition of the model
5. An analysis was run using the Global Insight model of the U.S. economy. These results also showed significant decreases in
the variables, with savings being hit hardest first and the employment impact being felt a few quarters later. The Global
Insight simulation was for a $1 increase in the price, and the effect on employment was 544,000 jobs lost.
6. The Kwiatowski, Phillips, Schmidt, and Shin test for stationarity was used. See D. Kwiatowski, P. C. B. Phillips, P. Schmidt,
and Y. Shin, “Testing the Null Hypothesis of Stationarity Against the Alternative of a Unit Root: How Sure Are We That
Economic Time Series Have a Unit Root?” Journal of Econometrics, Vol. 54 (1992), pp. 159–178.
7. The model was run on RATS v. 6.2 on a PC using the Windows XP OS. Different lag structures were tested and resulted in
qualitatively similar results.
8. Durbin–Watson statistics were between 1.89 and 2.04 for the four equations estimated. The standard errors of the
estimates for each of the four equations were 0.07, 0.005, 0.007, and 0.003, respectively.
page 5

---