Download SALES AND USE TAX

SALES AND USE TAX
Taxes on Consumption--Introduction
The most important taxes on consumption are sales and excise taxes. Sales
taxes usually get paid on such things as cars, clothing and movie tickets. Sales
taxes are an important source of revenue for most states and some large cities
and counties. The tax rate varies from state to state, and the list of taxable goods
or services also varies from one state to the next.
Excise taxes, sometimes called “luxury taxes,” are used by both state and
Federal Governments. Examples of items subject to Federal excise taxes are
heavy tires, fishing equipment, airplane tickets, gasoline, beer and liquor,
firearms, and cigarettes.
The objective of excise taxation is to place the burden of paying the tax on the
consumer. A good example of this use of excise taxes is the gasoline excise tax.
Governments use the revenue from this tax to build and maintain highways,
bridges, and mass transit systems. Only people who purchase gasoline—who
use the highways—pay the tax.
Some items get taxed to discourage their use. This applies to excise taxes on
alcohol and tobacco. Excise taxes are also used during a war or national
emergency. By raising the cost of scarce items, the government can reduce the
demand for these items.
Consumption-Theoretical Framework
Permanent-Income Hypothesis
The central idea of the permanent-income hypothesis, proposed by Milton
Friedman in 1957, is simple: people base consumption on what they consider
their "normal" income. In doing this, they attempt to maintain a fairly constant
standard of living even though their incomes may vary considerably from month
to month or from year to year. As a result, increases and decreases in income
that people see as temporary have little effect on their consumption spending.
The idea behind the permanent-income hypothesis is that consumption depends
on what people expect to earn over a considerable period of time. As in the lifecycle hypothesis, people smooth out fluctuations in income so that they save
during periods of unusually high income and dis-save during periods of unusually
low income. Thus a pre-med student should have a higher level of consumption
than a graduate student in history, if both have the same current income. The
pre-med student looks ahead to a much higher future income, and consumes
accordingly.
In order to test the theory, Friedman assumed that on the average people would
base their idea of normal or permanent income on what had happened over the
past several years. Thus if they computed permanent income as the average of
the past four years, and income had been $13,000, $10,000, $15,000, and
$8,000, they would consider their permanent income as $11,500. Though our
expectations of future income do not depend solely on what has happened in the
past, these additional factors are almost impossible to include into attempts to
test the theory with data.
Both the permanent-income and life-cycle hypotheses loosen the relationship
between consumption and income. This is more clearly seen in the permanentincome hypothesis, which suggests that people will try to decide whether or not a
change of income is temporary. If they decide that it is, it has a small effect on
their spending. Only when they become convinced that it is permanent will
consumption change by a sizable amount. As is the case with all economic
theory, this theory does not describe any particular household, but only what
happens on the average.
The life-cycle hypothesis introduced assets into the consumption function, and
thereby gave a role to the stock market. A rise in stock prices increases wealth
and thus should increase consumption while a fall should reduce consumption.
Hence, financial markets matter for consumption as well as for investment. The
permanent-income hypothesis introduces lags into the consumption function. An
increase in income should not immediately increase consumption spending by
very much, but with time it should have a greater and greater effect. Behavior
that introduces a lag into the relationship between income and consumption will
generate the sort of momentum that business-cycle theories saw. A change in
spending changes income, but people only slowly adjust to it. As they do, their
extra spending changes income further. An initial increase in spending tends to
have effects that take a long time to completely unfold.
The existence of lags also makes government attempts to control the economy
more difficult. Policies taken do not have their full effect immediately, but only
gradually. By the time they have their full effect, the problems that they were
designed to attack may have disappeared.
Finally, though the life-cycle and permanent-income hypotheses have greatly
increased our understanding of consumption behavior, data from the economy
does not always fit theory as well as it should, which means they do not provide a
complete explanation for consumption behavior.
The Life-Cycle Hypothesis
In examining why people spend the amount they do, a logical starting point is to
ask what goals they have. Two goals seem reasonable for many people. First,
they prefer a higher standard of living to a lower standard of living. In other
words, people want the highest level of consumption spending they can get.
Second, most people prefer to have a roughly constant standard of living through
time. They do not like to live on a roller coaster, with one year of feast followed
by a year of famine.1 Put together, these two goals suggest that we assume that
people try to maintain the highest, smooth consumption path that they can
get. Presented in this way, discussion of consumption behavior becomes a
problem, that is, the tools of microeconomics are designed to examine (with
budget lines and indifference curves).
Consider, for example, a person who will earn $10,000 this year and $5,000 next
year. Ignoring all future years, how should he spend? The idea that he wants a
smooth consumption path suggests that he should save part of this year's
income and spend it next year. Alternatively, if the person earned the $5000 this
year and the $10000 next year, the goal of a high smooth consumption path
suggests that he borrow in the first year and pay back in the second. An
important function of financial markets from the point of view of consumers is that
these markets help one maintain a constant standard of living despite
fluctuations in income.
The idea that people have fluctuations in income which they want to smooth is
the basis of the life-cycle hypothesis of consumption which was produced by
Franco Modigliani, Richard Brumberg, and Albert Ando in a series of articles in
the 1950s and 1960s. The life-cycle hypothesis asks why people save. It
answers that people generally live longer than they earn income--that is, people
usually retire. If they are to keep spending after they no longer earn income, they
must have accumulated assets while they were earning so that they can dissave. (Few are willing to lend to those who have no prospect of future income.)
Suppose a 20-year-old person expects to live 50 years more, but only to work for
40 of those years. He expects to earn $20,000 each year. Ignoring interest, this
person will have earnings of $800,000 to spread over 50 years. If he spends
$16,000 a year, he will die with zero assets left. To get this spending pattern, he
saves $4,000 each year while he works, and at retirement will have assets of
$160,000.
Now suppose that the person in the above example begins with assets of
$200,000. He will then have a lifetime amount of $1,000,000 that can be spent.
He will be able to spend $20,000 each year and die with zero assets. Thus the
life-cycle hypothesis introduces wealth as a factor into the consumption function.
Consumption can be financed either through income or through the sale of
assets, and an increase in either should increase consumption.
By looking at what a "typical" individual should do, the life-cycle hypothesis builds
microeconomic footings for the consumption function. Behavior is goal-directed in
the life-cycle hypothesis, while it is not in the original Keynesian consumption
function. This latter consumption function is mechanical without a reason. Since
economists prefer behavior that can be explained in terms of people pursuing
goals, it is no surprise that the life-cycle hypothesis has become popular.
The life-cycle hypothesis can be expanded to take into account uncertainty of
when death will occur, the existence of social security, the interest rate, savings
for bequests for heirs, and various patterns of lifetime earnings. It does not deal
well with what should happen if incomes fluctuate erratically over time, but for
this situation another theory, the permanent-income hypothesis, provides an
answer. The permanent-income and life-cycle hypotheses are not contradictory
theories, but theories which nicely complement each other.
Forecast Sales and User Tax
Tax rate and taxable components
-------------Sales Tax Rate by State----------------Prescription
Drugs
State
ALABAMA
Tax Rates
none
ARIZONA
5.6
ARKANSAS
CALIFORNIA (3)
COLORADO
CONNECTICUT
DELAWARE
Food
4
ALASKA
*
*
6
*
*
7.25 (2)
*
*
2.9
*
*
6
*
*
*
*
none
FLORIDA
6
*
*
GEORGIA
4
* (4)
*
HAWAII
4
IDAHO
5
ILLINOIS (2)
INDIANA
IOWA
KANSAS
*
*
6.25
1%
1%
6
*
*
5
*
*
5.3
6
*
*
LOUISIANA
4
* (4)
*
MAINE
5
*
*
MARYLAND
5
*
*
MASSACHUSETTS
5
*
*
6
*
*
MINNESOTA
6.5
*
*
MISSISSIPPI
7
MISSOURI
4.225
MONTANA
none
NEBRASKA
NEVADA
NEW HAMPSHIRE
1.225
*
5.5
*
*
6.5
*
*
6
*
*
5
*
*
NORTH DAKOTA
4
*
*
4.5
* (4)
*
5
*
*
OHIO
5.5
*
*
OKLAHOMA
4.5
OREGON
PENNSYLVANIA
*
none
NEW MEXICO
NORTH CAROLINA
*
*
NEW JERSEY
NEW YORK
1%
*
KENTUCKY
MICHIGAN
Non-prescription
Drugs
*
*
*
none
6
*
*
*
RHODE ISLAND
7
SOUTH CAROLINA
5
*
SOUTH DAKOTA
4
*
TENNESSEE
*
*
7
6%
*
TEXAS
6.25
*
*
UTAH
4.75
VERMONT
*
*
*
6
*
*
*
5 (2)
2.5% (2)
*
*
6.5
*
*
WEST VIRGINIA
6
5%
*
WISCONSIN
5
*
*
WYOMING (3)
4
VIRGINIA
WASHINGTON
DIST. OF COLUMBIA
5.75
*
*
*
*
NY Sales Tax Forecasting Models
Tax Base and Rate
New York State has imposed a general sales and use tax since 1965. It is
currently the State’s second largest tax revenue source generating over $12
billion annually. The tax rate has been 4 percent since 1971 although a
temporary surcharge to 4.25 percent was imposed from June 1, 2003, to May 31,
2005. Counties and cities within the State are authorized to impose an additional
3 percent sales and use tax, although most have temporary authorizations to
impose the tax at a higher rate. New York City and 37 counties currently have a
State and local combined rate of 8 percent, including the 0.375 percent
Metropolitan Commuter Transportation District tax in the MCTD area. The
highest maximum combined State and local rate is 9.5 percent in Oneida County.
The tax applies to sales and uses within the State of tangible personal
property (unless specifically exempt), certain utility service billings, restaurant
meals, hotel and motel occupancy, and specified services and admission
charges. Certain exemptions such as food, prescription drugs, residential
energy, and college textbooks have been enacted to lessen the regressiveness
of the tax. Other items, including machinery and equipment used in production
and property purchased for resale, are excluded from tax to avoid tax
pyramiding.
Administration
Persons selling taxable property or services are required to register with
the Department of Taxation and Finance as sales tax vendors. Vendors
generally are required to remit the tax that they have collected quarterly.
However, vendors who record more than $300,000 of taxable sales in any of the
immediately preceding four quarters must remit the tax monthly, by the twentieth
of the month following the month of collection. Vendors collecting less than
$3,000 yearly may elect to file annually, in March. Finally, monthly filers
collecting more than $500,000 in tax annually are required to remit the tax by
electronic funds transfer (EFT). The collections for the first 22 days of the month
must be remitted electronically within three business days after the 22 nd day.
DATA SOURCES
The primary sources of data used in the estimation and forecasting
methodology for the sales tax are as follows:
● AS043, Department of Taxation and Finance Monthly Report of Receipts.
This report contains gross and net receipts data.
● Various reports, Department of Taxation and Finance. Other reports
supplementing the RS-43 provide information on data such as audit
collections, prior period adjustments and daily receipts.
● Various U.S. and New York government agencies, including the U.S.
Bureau of Economic Analysis of the Commerce Department. These
agencies provide economic data used in the econometric equations.
●
STATUTORY CHANGES
The Division of the Budget has developed a series of State fiscal year
sales and use tax receipts that has been adjusted for Tax Law, and
administrative and other changes to allow for year-to-year comparisons of the
taxable sales base.
Major legislative and administrative events causing divergent growth in
actual sales tax receipts from the constant law series include:
● large taxable base expansion in 1991-92;
● one-time spin-up due to the implementation of EFT in 1992;
● exceptional audit collections in 1994-95;
● implementation of vendor credit program in 1995-96;
● week-long exemptions for clothing and footwear biannually from 1997-98
to 1999-2000;
● exemption for promotional materials in 1997-98;
● exemption for college textbooks in 1998-99;
● expansion of the vendors’ credit in 1999-2000;
● permanent exemption for clothing and footwear priced under $110
beginning March 1, 2000;
● lower tax rate on charges for separately purchased transmission and
distribution of electricity and gas in 2000-01;
● rate surcharge from 4 percent to 4.25 percent effective June 1, 2003 to
May 31, 2005; and
● suspension of the permanent clothing exemption between June 1, 2003,
and May 31, 2007; replaced by two exemption weeks annually at a
threshold of $110 per item.
Collection Components
(millions of dollars)
1,400.0
Sales and Use Tax
1,200.0
Collections
1,000.0
800.0
600.0
400.0
200.0
0.0
Jan-70
1,000.0
900.0
800.0
700.0
600.0
500.0
400.0
300.0
200.0
100.0
0.0
Jan-70
Jan-77
40.0
30.0
20.0
10.0
0.0
(10.0)
(20.0)
(30.0)
(40.0)
(50.0)
(60.0)
Jan-70
Jan-91
Jan-98
Jan-05
Jan-91
Jan-98
Jan-05
Jan-98
Jan-05
Jan-98
Jan-05
Trend
Jan-77
600.0
500.0
400.0
300.0
200.0
100.0
0.0
(100.0)
(200.0)
(300.0)
(400.0)
(500.0)
Jan-70
Jan-84
Jan-84
Seasonal
Jan-77
Jan-84
Jan-91
Irregular
Jan-77
Jan-84
Jan-91
FORECAST MODELS
MODEL 1: TAXABLE CONSUMPTION MODEL
Consumption of Taxable Goods and Services in New York
Detailed components of nominal U.S. consumption of durable and
non-durable goods are weighted based on what percentage is estimated
to be taxable in New York. These weighted components are then
summed and multiplied by the ratio of New York to U.S. employment to
estimate State taxable consumption of durable and non-durable goods.
To more closely capture the lag between economic activity and tax
collections, one third of the prior quarter's State taxable consumption is
added to two thirds of the current quarter value.
As for goods, detailed components of nominal U.S. consumption of
services are weighted based on what percentage is estimated to be
taxable in New York. The same steps taken for goods to estimate State
consumption and adjust for the collections lag are repeated for services.
Other Economy Related Explanatory Variables
The fourth difference of the log of U.S. investment in equipment and
software, lagged one period, is used to capture the sales taxes paid by
businesses.
The log of the current period value of the S&P 500 index minus the log of
the value for the same quarter of the prior year captures the importance of
the financial sector to the New York economy.
The model specification appears below:
TAXABLE CONSUMPTION MODEL
 4 ln SALESAdjt   0.018  0.616  4 ln CDNTX t  0.560  4 ln CSTX t  0.076  4 ln IPDENRt
(0.005) (0.094)
(0.116)
(0.039)
 0.034  4 ln SP 500t  0.060 DUMCLOTHt  0.083 DUM1986t  0.038 DUM 2004t
(0.014)
(0.012)
(0.021)
(0.015)
 0.025 DUM1990t
(0.011)
Adjusted R 2  0.77
SALESAdj
Adjusted quarterly sales tax receipts
CDNTX
Taxable durable and nondurable consumption goods
CSTX
Taxable services
IPDENR
Investment in equipment and software
SP500
S&P 500 index
DUMCLOTH
Clothing dummy
DUM1986
Dummy variable (=1 for 1986 Q1; 0 elsewhere)
DUM2004
Dummy variable (=1 for 2004 Q1 and Q2; 0 elsewhere)
DUM1990
Dummy variable for cable exclusion
PERCENT CHANGE IN EXOGENOUS VARIABLES — STATE FISCAL YEARS 1999-2000 TO 2009-10
Consumption of goods in NY
Consumption of services in NY
S&P Index
99-00
10.1
6.6
19.4
00-01 01-02 02-03
7.4
1.9
3.3
7.4
1.4
2.4
1.8 (16.9) (19.7)
03-04
4.8
5.1
11.2
04-05
6.1
5.1
11.0
05-06
6.1
5.0
7.4
06-07
3.7
5.1
9.5
07-08 08-09
5.1
-1.2
5.2
3.4
8.3 (25.5)
09-10
Estimated
-4.6
1.6
(7.8)
MODEL 2: ERROR CORRECTION MODEL WITH STATE INCOME AND EMPLOYMENT
This model exploits the long-run equilibrium relationship between sales tax
receipts and New York State disposable income and total nonfarm employment.
That relationship is estimated and the lagged deviations appear on the righthand-side of the sales tax model within an error correction model framework that
allows for a gradual dynamic adjustment back toward equilibrium. Consistent
with that framework, the model also includes the year-ago differences in State
disposable income and employment. Also appearing on the right-hand-side are
a lagged value of the dependent variable, the S&P 500, and two dummy
variables, one to account for changes in the clothing exemption and the other for
an outlier.
Economy Related Variables
The log of current-quarter total nonfarm New York State employment
minus the log of the value for the same quarter of the prior year.
The log of current-quarter New York disposable income minus the log of
the value for the same quarter of the prior year.
The log of the current-period value of the S&P 500 index minus the log of
the value for the same quarter of the prior year captures the importance of
the financial sector to the New York economy.
The model specification appears below:
ERROR CORRECTION MODEL INCLUDING INCOME AND EMPLOYMENT
Long-term Eqm. Model :
ln SALESAdjt  1.086 ln NYSEMPt  0.744 lnYDNYt  0.050 SEASONQ 3t
(0.005)
(0.007)
(0.006)
Adjusted R 2  0.99
Error Correction Model:
 4 ln SALESAdjt  0.233  4 ln SALESAdj t 1  0.833 ECt  4  0.070  4 ln NYSEMPt  0.465  4 lnYDNYt
(0.095)
(0.192)
(0.016)
(0.085)
 0.071  4 ln SP 500t  0.034 DUMCLOTHt  0.055 DUM1986t
(0.016)
(0.014)
(0.021)
Adjusted R 2  0.66
SALESAdj
Adjusted quarterly sales tax receipts
NYSEMP
NY employment
YDNY
NY disposable income
SEASONQ3
Seasonal dummy for third quarter
EC
Error correction term, i.e., the deviation from the long term equilibrium relationship
SP500
S&P 500 index
DUMCLOTH
Clothing dummy
DUM1986
Dummy variable (=1 for 1986 Q1; 0 elsewhere)
PERCENT CHANGE IN EXOGENOUS VARIABLES STATE FISCAL YEARS 1999-200 TO 2009-10
NY Disposable Income
NY Employment
S&P Index
99-00
3.6
2.3
19.4
00-01 01-02 02-03
6.2
1.3
3.2
1.9
(1.6)
(1.2)
1.8 (16.9) (19.7)
03-04
4.9
(0.5)
11.2
04-05
6.1
0.9
11.0
05-06
5.7
0.9
7.4
06-07
6.3
1.1
9.5
07-08
4.4
1.4
8.3
08-09
3.8
-0.2
(25.5)
09-10
Estimated
-1.5
-2.1
(7.8)
MODEL 3: AUTO SALES AND RETAIL TRADE EMPLOYMENT
This model exploits two alternative indicators of the growth in taxable
sales. To capture the large portion of taxable sales that are attributable to the
auto market, this model includes growth in the number of State vehicle
registrations. Retail trade employment represents yet another indicator of the
strength of taxable sales and is also included in the model. Also appearing on
the right-hand-side are the S&P 500 and three dummy variables, one to account
for changes in the clothing exemption and two for outliers. A forecasting model
for vehicle registrations is also specified below.
Nominal Value of Auto and Light Truck Registrations
The logarithm of New York new auto and light truck registrations multiplied
by the national average price of new light vehicles minus the logarithm of
the same concept for the prior year. These data are not seasonally
adjusted.
An additional model forecasts vehicle registrations in New York State.
Vehicle registrations are explained by national light vehicle sales
multiplied by the ratio of State to U.S employment to determine the share
attributable to New York. Both contemporaneous and lagged auto sales
are statistically significant in the model. In addition, the lagged value of
the dependent variable and the year-ago change in 5-year Treasury yield
are included, with the latter capturing borrowing costs. A dummy variable
is included in the model to account for the inclusion of light trucks in the
data series as of the first quarter of 1993.
Retail Trade Employment
It is expected that as retail sales grow, outlets will increase their demand
for workers. Employment is an indicator of real economic activity, while
sales tax receipts reflect changes in both real activity and prices.
Therefore, retail employment is multiplied by a measure of the price level
constructed to capture inflation trends unique to New York.
All variables except the price deflator are not seasonally adjusted. The
model specification appears below:
VEHICLE SALES AND RETAIL EMPLOYMENT MODEL
 4 ln SALESAdj t  0.008  1.082  4 ln(EMP 46t * CPICOMPt )  0.070  4 ln NOMCARSt  0.058  4 ln SP 500 t
(0.004) (0.082)
(0.016)
(0.013)
 0.027 DUMCLOTHt  0.054 DUM 2004t  0.055 DUM1986t
(0.011)
(0.015)
(0.021)
Adjusted R 2  0.78
EMPNYt
EMPNYt 1
 4 lnVEHREGNYt  0.642  4 ln(SQLVt *
)  0.149  4 ln(SQLVt 1 *
)  0.108  ln VEHREGNYt 1
EMPUS
EMPUSt 1 (0.056) 4
(0.074)
(0.079)
t
 0.010  4RMGF 5NSt 1  0.281 DUM1993t
(0.004)
(0.033)
Adjusted R 2  0.76
NOMCARSt  VEHREGNYt * JPLVt
SALESAdj
Adjusted quarterly sales tax receipts
EMP46
NYS retail sector employment
CPICOMP
NYS CPI
NOMCARS
Nominal value of vehicles sold in NY
SP500
S&P 500 index
DUMCLOTH
Clothing dummy
DUM2004
Dummy variable (=1 for 2004 Q1 and Q2; 0 elsewhere)
DUM1986
Dummy variable (=1 for 1986 Q1; 0 elsewhere)
VEHREGNY
Vehicle registrations in NY
SQLV
U.S. light vehicle sales
EMPNY
NYS Employment
EMPUS
US Employment
RMGF5NS
5-year U.S. Treasury yield
DUM1993
Dummy variable (=1 for 1993 Q1; 0 elsewhere)
JPLV
Average sales price of light vehicles
PERCENT CHANGE IN EXOGENOUS VARIABLES — STATE FISCAL YEARS 1999-2000 TO 2009-10
Nominal Value of Registered Autos
and Light Trucks
S&P Index
CPI NY
Retail Trade Employment
99-00
00-01
13.5
19.4
2.4
2.9
(4.9)
1.8
3.2
1.9
02-03
03-04
04-05
05-06
06-07
07-08
08-09
09-10
Estimated
8.5
3.1
(16.9) (19.7)
2.4
2.5
(2.2)
(0.6)
2.7
11.2
2.6
(0.1)
(1.8)
11.0
3.6
1.8
0.3
7.4
3.6
0.9
(2.6)
9.5
3.4
0.7
7.9
8.3
3.1
1.6
(18.1)
(25.5)
3.1
(1.0)
(6.9)
(7.8)
0.6
(3.0)
01-02
Adjustments
Budget Division forecasts for the relevant economic variables are used to
produce an estimate of underlying growth in base receipts. This growth rate is
arrived at by taking a weighted average of the forecasts from the three models
described above and applying it to a prior year sales tax receipt base that has
also been adjusted for Tax Law and other changes. However, the final receipts
forecast must include the impact of these factors. Consequently, in a final step,
the base forecast is converted back into a cash forecast by accounting for Tax
Law and administrative changes, audits, court decisions, tax cuts being phased
in, and prior period adjustments.
Cash Receipts
PERCENTAGE DISTRIBUTION OF CASH RECEIPTS
1996-97
1997-98
1998-99
1999-2000
2000-01
2001-02
2002-03
2003-04
2004-05
2005-06
2006-07
2007-08
2008-09
2009-10 (est.)
1st Quarter
24.4
24.5
24.8
24.3
24.4
24.7
23.9
22.7
25.6
25.5
24.8
25.2
25.7
24.4
2nd Quarter
25.3
25.8
25.6
24.7
25.7
23.5
26.6
26.3
25.3
25.5
25.5
25.3
26.7
25.4
3rd Quarter
25.5
25.3
25.0
26.1
25.4
26.7
24.8
26.4
25.2
24.5
25.9
25.2
24.3
25.4
4th Quarter
24.8
24.4
24.6
25.0
24.5
25.1
24.7
24.5
23.9
24.5
23.8
24.3
23.3
24.8
Risks to the Forecast
Errors in the forecasts of the exogenous variables provide a degree of risk
to the sales and use tax forecast. Forecast error in prior years can largely be
attributed to the forecasts of the exogenous variables. Variation in the estimate
may also occur as a result of administrative changes or unanticipated legislative
action.
Cash collections are reduced by credits and increased by collections from
audits and other administrative processes, which, due to payment schedules, are
unrelated to economic liability in the month remitted. To adjust the sales tax
series to more closely correspond to the economic activity that generated the
receipts, collections from the first ten days of the quarter are placed in the
previous quarter, non-voluntary collections (audit collections, tax compliance) are
removed from the series, the March prepayment (now repealed — applied to
March 1976 through March 1990 only) is placed in April, and an adjustment is
made for allocation errors made in prior periods.
Issues
What to be taxed?
Who to pay?
Policy Issue and What-ifs
Need to look at some micro data: Consumer Expenditure Survey
See Consumer Expenditure and Income.Doc