Full text of Working Papers (Federal Reserve Bank of Chicago) : The Effect of Sales Tax Holidays on Household Consumption Patterns, Working Paper 2010-06

The full text on this page is automatically extracted from the file linked above and may contain errors and inconsistencies.

Federal Reserve Bank of Chicago

The Effect of Sales Tax Holidays on
Household Consumption Patterns
Nathan Marwell and Leslie McGranahan

WP 2010-06

The Effect of Sales Tax Holidays on Household
Consumption Patterns
Nathan Marwell and Leslie McGranahan
Federal Reserve Bank of Chicago

July 27, 2010
PRELIMINARY AND INCOMPLETE – PLEASE DO NOT CITE

Abstract
Sales tax holidays (STHs) are the temporary suspension of state (and some local) sales taxes on
selected retail items for a brief period of time. The policy has gained popularity in recent years,
beginning in one state in 1997and growing to twenty by 2008. Despite the increased frequency
with which states use STHs, little research has been conducted to study how households respond
to this temporary tax manipulation. Our paper offers the first household-level, microeconometric evaluation on the effect of STHs on household consumption patterns. We find that
on STHs, households increase the number of clothing and shoes bought by over 49 percent and
45 percent, respectively, relative to what they buy on average. Further, we find that this increase
in consumption is limited to children’s apparel and that the wealthiest households and
households consisting of married parents and young children have the largest, statistically
significant response to STHs; for example, households with incomes over $70,000 increase the
number of children’s clothing items purchased by 136 percent, while households that consist of
married parents and young children increase the amount spent on children’s clothing and shoes
by 117 and percent and 295 percent, respectively.
The opinions expressed in this paper are those of the authors and do not reflect the views of
the Federal Reserve Bank of Chicago or the Federal Reserve System. We would like to
thank Kevin Feeney for excellent research assistance with this project. We welcome
comments. Please send them to Leslie.McGranahan@chi.frb.org.

I. Introduction
It is common for policymakers to temporarily manipulate tax rates to incentivize
consumer behavior. Sales tax holidays (STHs)—the suspension of sales taxes levied on targeted
retail goods for a set period of time—have grown in popularity over the past two decades. The
first sales tax holiday was instituted in 1997 in the state of New York to help clothing retailers
compete against stores in neighboring New Jersey, where clothes are not subject to the sales tax.
By 2008, STHs had expanded to 20 states to cover goods ranging from fanny packs to
refrigerators. The number of states with sales tax holidays declined to 17 in 2008 and 2009,
reflecting the worsening fiscal health of many states.1 However, this number is set to grow in
2010, with Florida reinstating its STH for the first time since 2007 and Illinois recently passing
its first STH. With nearly 100 million people living in a state with an STH in 2009, STHs can
potentially have a large impact on households, retailers, and state governments.
While the stated purposes for state tax holidays are varied, they are generally justified to
accomplish three goals. First, by eliminating the state-level sales tax, the STH lowers the final
sales price of retail goods to consumers. Further, the dates for STHs are often chosen to coincide
with periods of high seasonal demand, such as back-to-school shopping periods, with the
intention of providing relief to lower-income, liquidity-constrained households or those
otherwise deemed worthy of tax relief by policymakers. In justifying Illinois’s 2010 holiday, the
lead sponsor said, “[w]orking families with kids going back to school, we want to give them a
break.” (Associate Press, 2010) Second, lowering sales tax rates helps retailers lower prices
without reducing profits. And relatedly, STHs can help retailers compete better against their
counterparts in bordering states with lower sales tax rates. Third, by targeting specific items for
tax exemption, policymakers can encourage the consumption of specific goods that they believe
1

We are included the District of Columbia as one of the “states” with a sales tax holiday.

have inherit social benefits—such as goods needed during the hurricane season in Florida or
computers in Vermont.
Despite the growing popularity of the STH, little formal evaluation has been conducted to
determine whether they actually accomplish policymakers’ stated goals. There are a number of
reasons why instituting an STH may be ineffective, or why doing so may be considered bad
public policy. First, while the sales tax is eliminated for a brief period of time, consumers would
not see any benefit if retailers increased the pre-tax price of items; the degree by which prices
may be lowered depends on the incidence of taxation. Second, even if after-tax prices do drop
consumers may not purchase any more of the targeted item—the degree of change in consumer
response depends on the elasticity of demand for the eligible targets. Third, the temporal nature
of the STH may cause consumers to shift planned consumption over time. Fourth, because the
STH is available to all consumers, unintended recipients, such as wealthy households or those
without children, may also take advantage of the holiday or may even have the means to take
better advantage of the STH than the intended beneficiaries. In particular, this may be too blunt
a policy instrument for subsidizing the consumption of targeted groups. Fifth, there have been
complaints of onerous compliance costs for retailers who are forced to participate in an STH.2
The purpose of this paper is to test both the economic and policy implications of the
STH, and will focus on how consumers are affected. Through our analysis, we hope to address a
number of concerns. Does the STH actually encourage consumers to increase consumption of
both covered goods and ineligible items? How successful is the STH at targeting different types
of households, as well as different members within households? Is it the lower-income
households that are taking advantage of the STH or wealthier households? And within families,
are parents using the STH to purchase items for their children or themselves? Lastly, do
2

For further criticisms, see Hawkins and Mikesell (2001).

consumers increase overall spending or simply shift intended consumption across time so that it
occurs during an STH? Or do STHs cause consumers to shift consumption from ineligible items
to eligible ones?
We organize the remainder of our paper as follows. In section II, we provide background
information: a brief history of the STH, along with a review of relevant literature. In section III,
we summarize the data sources we use. We provide an overview of our econometric approach in
section IV. In section V, we present our findings. And we give our conclusion in section VI.

II. Background
Brief History of Sales Taxes and the STH
Sales taxes are levied on almost every tangible good and a number of services—the
exceptions being a few necessity goods such as food and prescription drugs. The tax applies at
the retail level, is ad valorem, and is remitted by vendors. Sales taxes are levied by different
levels of government; 45 states have some type of state-level sales tax, with counties, cities, and
even transportation authorities levying additional sales taxes. In 2010, the average state-level
sales tax was 5.6 percent and the median was 6.0 percent, with a range of 2.9 percent in Colorado
to 8.25 percent in California. Just under a quarter of state own-source revenues are derived from
the general sales tax. Sales taxes affect the vast majority of Americans almost every day of the
year.
Cole (2008) presents a historical synopsis of the STH. Briefly, the first STH occurred in
1997 in New York and was enacted to help the city’s retailers compete with their tax-free
competitors in neighboring New Jersey. Lasting seven days, the holiday exempted general use
clothing and footwear priced under $500 from the statewide 4 percent sales tax. The STH then

spread to Florida in 1998 and Texas in 1999; by 2007, 20 states had instituted at least one STH,
offering over a third of the U.S. population a brief respite from the sales tax. STHs are state
level policies although in a number of cases, localities have also suspended their sales tax during
the state holiday. The duration of the STH, the type of exempted items, the ceiling below which
these items could be purchased tax-free, and even the number of STHs offered each year are all
parameters set by the state legislative bodies; in a given year, no two STHs have been identical.

Previous research on the STH
There has been little academic research conducted on the consumer responses to STHs. A
wider body of literature has analyzed how prices respond to different tax rates and changes in tax
rates more generally. Besley and Rosen (1998) look at the prices of a bundle of consumption
goods across municipalities and use the heterogeneity in sales tax rates to estimate the degree to
which the tax burden is shifted to retailers or consumers. Their findings are mixed, with the tax
burden being shifted completely to consumers for some goods (for example, Big Macs, eggs,
Kleenex, and Monopoly games), while the tax is less than fully shifted to consumers for other
goods (for example, bananas, bread, Crisco, milk, shampoo, soda, and boys underwear). Doyle
and Samphantharak (2008) look at the how gasoline prices shift during a gas tax moratorium.
Analyzing prices at the station level throughout the Midwest and aggregating prices to the state
level, they find that 70 percent of the tax suspension and 80 percent to 100 percent of the
reinstatement is shifted to consumers. Fullerton (1987) analyzes the aggregate shifts in
consumption resulting from changes in the sales tax rate in Idaho, and finds that the shift in the
sales tax rate from 4.5 percent to 4.0 percent induces a deviation from consumption’s long-term

trend between $3 to $4 million. In other words, the decline in the sales tax led to an increase in
consumption.
One of the first studies to actually analyze an STH is Harper et. al (2003). They sample
prices on bundles of goods exempt from the sales tax before and during the 2001 Florida STH
and compare price dynamics for these same bundles to those in neighboring Mobile, Alabama
which did not have a STH and consequently experienced no change in the sales tax rate. They
find that, on average, prices of tax exempt items in Florida fall by 80 percent of the sales tax rate
during the STH. Mogab and Pisani (2007) survey shoppers at a shopping mall in Texas during
the 2004 STH to discern their attitudes and perceptions concerning the STH, and find that it
plays an important role in deciding whether or not to shop.
Cole (2009a) presents the first fully specified econometric evaluation of an STH. In Cole
(2009a), the author estimates how an STH affects the prices and quantities of computers and find
weak evidence that the sales tax savings is either fully passed through or mildly overshifted to
consumers during the STH (most of his estimates are not statistically significant). Cole (2009a)
simulates the aggregate quantity of computers that would have been consumed absent an STH,
and compares this counterfactual to the actual numbers to estimate the impact of the STH on the
quantity purchased. His main findings are that an STH induces consumption of between 6 and
17 additional computers per 10,000 people and that the timing of purchases to coincide with an
STH can account for between nearly 40 percent and 90 percent of the increases in purchases in
states with STHs over a 30-week period. Cole (2009b) analyzes the fiscal impact of these tax
holidays, and finds that there is about a 4 percent reduction in states’ sales and use tax collection
in the months with an STH.

The previous literature suggests that retailers will decrease the final sales price of items in
response to an STH and that an STH in at least one state played a significant role in consumers’
decision to go shopping. In addition, the effects of STHs on an aggregate level imply that they
are successful in encouraging consumption of computers (although without a great deal of
statistical certainty) and that in general, sales tax receipts fall during months that have an STH.
It remains to be seen, however, how households directly respond to the lower prices that occur
during a STH.
III. Data
Our empirical approach to exploring STH effects on consumption relies on merging a
data set on daily household consumption with a census of STHs, which contains information on
dates for STHs, the corresponding states that instituted them, and the items that were tax-exempt
during these STHs. The micro data on household consumption are from the Diary portion of the
U.S. Bureau of Labor Statistics’ (BLS) Consumer Expenditure Survey (CEX) from 1997 to 2008;
and the STH data are from Cole (2008).
The Diary Survey tracks the spending patterns of households for two weeks. Households
use diaries to track spending on detailed items. The data from the Diary Survey are used in
conjunction with the data from the better known Interview survey to generate the market baskets
used in the construction of the Consumer Price Index (CPI). The purpose of the Diary survey is
to gather detailed information on purchases made very frequently. The CPI relies on Diary data
for the measurement of all home food consumption, all footwear purchases, and most clothing
purchases. However, the Diary Survey itself covers all purchases and expenses. From the Diary
Survey, we know the date of a given purchase, the type of item purchased, the state where the

household resides, and the after-tax cost.3 The CEX also includes detailed household and person
level socio-demographic data.
Households are instructed to record pre-tax prices for items purchased during their
sample period, after which the BLS calculates and subsequently publishes the after-tax cost. The
CEX does not take into account STHs when adding the sales tax -- sales taxes are added during
data processing for items that are temporarily exempt from the sales tax. Hence, we have had to
remove the sales tax from the all the eligible items purchased during STHs. Because we cannot
identity geographic locations beyond the state of residence, we can only remove the state-level
sales taxes in our analysis. Hence, our analysis of the dollar amount spent on items exempted
from sales taxes during STHs may be slightly biased upward, since we are unable to remove
local sales taxes in those cases when the local sales tax is also suspended.
The STH data contain information on every STH that occurs between 1997 (the year of
the original STH) and 2008. The STH data contain information on the states, the dates, the items
exempt from the sales tax, and the price ceiling for tax-exempt items (under which consumers do
not have to pay the sales tax). Cole (2008) contains data on all the STHs between 1997 and
2007. For the year 2008, we use data from the Federation of Tax Administrators4. The five most
popular items that are exempt from sales tax during an STH are clothing, shoes, clothing
accessories, computers, and school supplies. The modal price ceilings at which these items are
exempt is $100 for clothing, shoes, and clothing accessories; $2,500 for computers; and $75 for
school supplies. As stated earlier, the dates for STHs are chosen to coincide during periods of
high seasonal demand, the most popular being back-to-school season. The STH data reflect this
3

The Diary survey asks for respondents to report the pre-tax price. Sales tax is applied to expenditure items by the
BLS before the data are transmitted to the data users. The state of residence is available in the public use microdata
for most households. It is omitted or recoded for some households. We only use those households whose state of
residence we are certain of.
4
http://www.taxadmin.org/

claim, with 65 percent of all STHs beginning in August. In addition, 48 percent of STHs begin
on Friday and many only last for a weekend.
Using the merged Diary–STH data, we measure spending on a given day by a household.
We separately tabulate spending for a number of different categories of expenditure and focus on
the categories for which states have enacted sales tax holidays. Our unit of analysis is a
“household- date”: on a given date some households live in states where there is a sales tax
holiday, while other households live in states where sales taxes remain unchanged.
Conceptually, our identification strategy is to use consumption patterns among individuals living
in a state without a sales tax holiday on a given date as a control for consumption patterns among
individuals living in states where there is a sales tax holiday. Because sales tax holidays have
been concentrated in clothing, shoes, accessories, computers, and school supplies, we focus on
those categories of expenditure in our analysis. Our resulting merged data set tracks the daily
purchasing decisions of just under 50,000 households, with over 600,000 household-date
observations.
Summary statistics from the merged data are presented in Table 1. The table shows the
daily means of household expenditures for all items as well as for items traditionally tax-exempt
during an STH. We show expenditure data in three ways—total dollar amount spent, the number
of items purchased, and the percentage of days in which any purchase is made in a given
category. On an average day, households spend $136 on all expenditures. Of this total, $11.80
is spent on food purchased for home use, and $115 on other items not tax-exempt during STHs,
leaving $8.35 spent on items that are traditionally tax-exempt during STHs. This amount
accounts for only 6.14 percent of all daily spending. Similarly, only 4.74 percent of all items
purchased on an average day are items that are tax-exempt during STHs. Of the goods that are

typically tax-exempt during an STH, clothing constitutes the bulk of the purchases; over 50
percent of both the amount spent on and the number of items purchased of traditionally taxexempt items are devoted to buying clothing.
Not only do STHs only pertain to a small percentage of daily expenditures, but they only
affect a small percentage of days each year. Table 2 shows the frequency of STHs in the CEX
data. While our data set consists of over 600,000 household-date observations, only 2,241 of
them fall on an STH (0.37 percent). STHs that exempt clothing from tax are the most popular,
applying to 0.33 percent of our observations, followed by shoes (0.27 percent) and school
supplies (0.17 percent). STHs are revealed again to only affect a sliver of household
consumption activity. The combination of the minimal share of daily consumption devoted to
tax-exempt items on STHs and the infrequent occurrence of STHs suggest that STHs have a
relatively small window for influencing household consumption.
Table 2 also shows variable means for the demographic variables we include as controls
in our analysis of consumption patterns.

IV. Model
Our econometric analysis seeks to identify the consumer response to a sales tax holiday.
Briefly, our econometric model is as follows:
_

where
_

,

(1)

is an outcome measuring the purchases by household i in state s on day t.
is an indicator variable equal to 1 if there is a sales tax holiday on date t in the state

s where household i lives.

is a vector of the household-level controls of age, sex, and race of

reference person; marital status; household composition; and household income.
state and time fixed effects, respectively, and

and

are

is a household-level error term. Our time fixed

effects are calendar dates (for example, Jan 5, 2006). As a result of our including these fixed
effects, we are assessing whether individuals living in states with sales tax holidays on a given
date have purchasing patterns that differ from otherwise similar individuals living in states
without STHs evaluated on the exact same date. Because we include state and time fixed effects,
our model estimates the average treatment effect that experiencing an STH has on a variety of
outcome measures.
Including the calendar date controls is crucial for our identification strategy because
STHs are intentionally positioned to occur during periods of high seasonal demand—for
example, weekends during back-to-school shopping season (August). If we did not compare
households shopping on the exact same calendar date, we would not be able to take into account
the increase in demand that occurs on weekends and in August. By including the calendar date
controls, we are able to net out the shared seasonal preference shocks that affect all households,
isolating the effect of the STH.5
For the majority of our empirical work, we use daily data on purchasing behavior, using
outcome variables such as the total amount spent and number of units purchased on day t by
household i in state s.. We investigate spending for a number of different expenditure categories.
When using these outcome variables, we use an ordinary least squares (OLS) regression,
clustering standard errors at the household level. In a few instances, we are also interested in
seeing how an STH affects the probability of purchasing an item. For these circumstances, we
use a linear probability model.
5

One concern with this procedure is that sales tax holidays are placed at periods of high seasonal demand specific to
a given state. We address this concern in the robustness section that follows our results.

V. Results
Baseline Results
Table 3 presents our baselines results of running model (1) for five different items
frequently exempt from sales taxes during STHs: clothing, shoes, clothing accessories,
computers, and school supplies. For each item category, we employ three different outcome
variables: the dollar amount spent on the item on a given day, the total number of items in that
category purchased on a given day, and an indicator variable equal to 1 if an item was purchased
and equal to 0 if no items in that category were purchased. Our baseline results reveal that STHs
have a statistically significant and economically meaningful impact on household consumption
decisions. Specifically, we find sizeable effects on the amount spent, the number of items
purchased, and the likelihood of making a purchase during STHs that exempt clothing, shoes,
and school supplies from the sales tax. On days that have an STH, households increase the
quantity of clothing purchased on average by an additional 0.083 items and the quantity of shoes
by 0.013 items. Similarly, the probability that a household will purchase clothing and shoes
increases by 1.4 percentage points and 1.0 percentage points respectively, on days that have an
STH. Consumption of school supply items also increases by 0.055 items on days with a STH.
The purchasing of clothing accessories and computers do not appear to be affected by an STH.
Most of the point estimates are negative and imprecisely estimated.6
While the point estimates for the increased consumption of clothing and shoes on days
that have an STH are small and may appear economically meaningless, when compared with
average daily consumption they actually represent a significant increase in economic activity.
The bottom row of Table 3 displays the ratio of the coefficients of STH_Day to their
6

We observe a negative and statistically significant coefficient for the number of computers purchased. Given the
relative infrequency with which people purchase computers and the rarity of computer STHs, we think there is too
much noise to take this coefficient seriously.

corresponding daily means. Relative to average daily consumption, days that have an STH result
in a sizeable increase in the number of shoes and clothing purchased. On an STH, the number of
clothing items purchased by households rises almost 50 percent; the probability that they will
purchase any clothing item rises 21 percent; and the amount they spend on clothing goes up 26
percent. Similarly, on STHs, the number of shoes purchased by a household increases 45 percent
and the probability that they will make any shoe purchase rises 44 percent. The amount of
school supplies purchased increases the most on STHs; household purchase over 60 percent
more school supply items on days with an STH than on days without one.
Given that our initial results show that STHs increase the consumption of both clothing
and shoes, we next want to explore how this increased consumption is distributed within
households. Specifically, we are interested in seeing whether STHs cause an increase in the
consumption for adults or children.7 A significant political motivation behind STHs is the need
to provide children with new apparel; children, though, rely on their parents to fulfill their
consumption needs. Examining whether parents respond to an STH by increasing their
children’s consumption or their own shows us whether STHs are effective as a policy mechanism
to increase consumption of targeted goods.8
Table 4 disaggregates the clothing and shoe items into those purchased for men, women
and children.9 The results reveal that the increases found in consumption of clothing and shoes
on days that have an STH are limited solely to children’s items; somewhat surprisingly, men’s
and women’s apparel exhibit no statistically significant increase on STHs in either the amount

7

It is not possible to disaggregate school supplies in a similar manner. Further, in households with children,
purchase of school supplies are likely made on their behalf
8
The difficulty for retailers to determine whether an item is explicitly meant for children is likely the predominate
reason why STH rules do not simply make only children’s apparel eligible for tax exemption.
9
The breakdown between men’s, women’s and children’s purchases is based on for whom the item was purchased
by the household. The children category includes infant apparel.

spent or quantity purchased. On average, households increase the amount spent on children’s
clothing by $0.91, and they increase the quantity purchased by 0.068 clothing items and 0.010
shoe items. These results amount to substantial increases in consumption relative to the daily
means; the amount spent on children’s clothing and the number of children’s clothing items
purchased increased 91 percent and 129 percent, respectively, and the number of children’s
shoes purchased increased 130 percent. On STHs, there is also an increase in the likelihood of
any purchase of children’s apparel being made; on average, households are 63 percent and 125
percent more likely to purchase any children’s clothing and any children’s shoes, respectively,
on dates that are STHs compared with dates that are not. Given these household responses, state
tax holidays appear to support policymakers’ stated objective of helping families with school-age
children, often right before the academic year begins.

Heterogeneous Households
To justify the enactment of state tax holidays, policymakers focus their political rhetoric
not only on the consumption needs of children, but also on the hardships of lower-income,
working class households. While policymakers might want to target the tax relief of STHs to
households of certain income levels, this has proven to be very difficult, if not impossible, given
the current designs of STHs. Since retailers cannot charge different sales tax rates on the same
items to households based on their income levels (or based on whether or not they have
children), there is little policymakers can do to incentivize households of different backgrounds
differently. Further, there are many reasons why lower-income households may be less likely to
reap the benefits of STHs. In fact, households with more financial resources may be better able
to take advantage of STHs. This would be consistent with the findings in Blattberg et al. (1978).

In that paper, the authors present a model on household buying behavior that identifies
characteristics related to cost parameters as playing the most important roles in determining the
consumption response to promotions for household items. Their empirical results indicate that
variables reflecting household resources, such as homeownership and automobile ownership, are
important predictors of households that are prone to take advantage of promotions. Given their
results, we may expect lower-income households to be less likely to take advantages of STHs.
To see how the consumption response to STHs may differ among households with
different familial and economic makeups, we segment our sample by household composition and
income groups, and run model (1) separately for each household type. The eight different
sample segments we investigate are households consisting of one parent with children;
households consisting of a married couple with no children, with young children (any children
under or equal to seventeen), and with old children (all children older than seventeen);
households consisting of single adults; “other” households;10 and households with annual
incomes of under $30,000, between $30,000 and $70,000, and over $70,000.11 When we run
model (1) on segments defined by household’s composition we include the income controls.
When we define the segments by income, we include the family composition controls.
The results of the regressions are displayed in Table 5. Each entry in the table represents
the result from a separate regression. We find a number of notable trends in the output. First,
with the exception of an increase in the number of shoes purchased by single households, there is
no discernable pattern concerning the change in spending among households without children or
among single parent households. For these groups, some of the point estimates are negative,
10

“Other” households are those most likely consisting of non-married partners.
These dollar delineations were kept constant for every CEX survey between 1997 and 2008; for example, the
group consisting of households earning below $30,000 includes both households earning below $30000 1997 dollars
for the 1997 CEX and households earning below $30,000 2007 dollars for the 2008 CEX In other words, these are
nominal rather than real break points..

11

some are positive and almost all of the effects are imprecisely estimated. Second, of all the
different types of household compositions, the only ones that consistently increase their
consumption on STHs consist of married couples with young children. STHs cause these
households to increase the amount spent on the daily consumption of clothing by $6.12 and the
number clothing and shoe items purchased by 0.277 and 0.057, respectively. These increases
account for approximately a doubling of the dollars spent and number of items purchased. For
married households with older children, the point estimates suggest an increase in clothing and
school supply purchases, but these effects are imprecisely estimated.
When we break down households by income group, we find that there is that there is no
statistically significant change in consumption for the lowest-income households. However,
most of the point estimates are positive and large. For the middle income group, we observe
significant increases in the number of clothes and school supplies purchased. Households
earning between $30,000 and $70,000 increase the quantity of clothing purchased during STHs
by 0.091, constituting a 54 percent increase relative to what these households spend on an
average day. These households also increase their number of school supply items purchased by
0.135 units, a quantity over 160 percent larger than what they purchase on average. Similarly,
households earning over $70,000 increase the number of clothing items purchased by 0.119
items, representing a 48 percent increase. The point estimates for the other categories of
expenditure are also positive (with the exception of money spent on school supplies), but are
imprecisely estimated.
We next breakdown spending on clothing and shoes based on the household member for
whom the items were purchased. We only perform these breakdowns for the three main groups
where we observed statistically significant increases in expenditure during STHs – married

couples with young children, families with income $30,000-$70,000, and families with income
>$70,000. The results are displayed in Table 6, and we show that the largest and only statistically
significant and positive increases in consumption are restricted to children’s apparel.
Up to this point, our analysis has focused on the marginal effects of STHs on households’
consumption behavior. We find large marginal effects for consumption of children’s clothes and
among higher income households and households consisting of married parents and young
children. It should be noted, however, that sales tax holidays also subsidize the consumption of
eligible items when households do not make any additional purchases as a result of the STH.
Tabulating the composition of household types that purchase eligible items during STH periods
is important in understanding who benefits from this tax subsidy. Table 7 presents the average
August daily expenditures on clothing, shoes, and school supplies and the sum of these for nine
population groups. We perform these tabulations for households in states without an STH and
on non-STH days for households in states with an STH. We choose to look at August
expenditure because it is the most popular month for STHs. We do not include STH dates to
isolate average spending independent of the influence of a tax holiday. In the final column of the
table, we calculate the average tax saving that would correspond to this expenditure had it been
tax free.12 The purpose of this table is to illustrate the amount of money different households
would save during a STH in absence of any additional spending spurred by the STH. As the table
shows, there exists a large amount of variation in economic activity by household composition.
Perhaps expectedly, the amount spent on STH eligible items increases monotonically in
income, with the households earning under $30,000 spending on average $3.94 while households
earning over $70,000 spend $12.0. If households were not to increase their spending in any way
during a sales tax holidays, the wealthiest households would save over twice as much in sales
12

This calculation is based on the tax rate of the state where the households live.

taxes as the poorest households. The other household type that spends the most on STH items in
August are households with children; households with both young and old children spend almost
the same amount on the tax-exempt items – a little over 12 dollars – although the former
households spend more on apparel while the latter spend more on school supplies. Another
interesting trend we observe is that while households that increase their spending on days with a
STH also spend more on eligible items in general, the reverse is not true: high average spending
does not necessarily translate into increased additional spending. For example, households
consisting of one parent spend nearly as much on household apparel as households with married
parents and young children ($8.83 versus $9.74), yet we do not observe them increasing
consumption in response to STH.
Table 7 illustrates that even in the absence of any stimulative effect, sales tax
holidays are only partially effective at helping subsidize policy makers the intended recipients.
While households with children shop the most during the period when STHs occur, lower
income households are among those that spend the least.

Consumption Shifting: Non-Exempt Items
In addition to seeing whether households respond to STHs by increasing consumption of
items exempt from sales tax, we are also interested in seeing how STHs may affect households’
other consumption decisions. Specifically, we hypothesize two distinct ways in which
households may change their normal shopping behavior to take advantage of an STH. First, they
may shift consumption away from items that are not exempt from sales tax. If households
devote a set amount of time or money for consumption every day, then consumers that spend

more time or money purchasing items that are exempt from sales tax may subsequently have less
time our money to purchasing items that are not. Alternatively, they may increase their
consumption of other items because they are already out shopping. This is one of the
justifications for the Illinois sales tax holiday—that the consumption of “other items pulled off
shelves while Mom’s in the store” will also increase. (Associated Press, 2010) To investigate,
we run model (1) with two small changes. First, we replace our dependent variable to two
different non-exempt items: food for home consumption and all other non-exempt items. Second,
instead of running the regressions separately for the STHs that exempt different items from sales
tax, we group all the STHs together, resulting in our treatment variable equal to 1 if any item is
temporarily exempt from the sales tax in the state where the household lives. The results are
shown in Table 8. On STHs, we do not find any statistically significant response to consumption
of either food for home use or non-exempt items. Further analysis that disaggregates the nonexempt category into separate consumption categories does reveal some statistically significant
decreases in consumption (for example, in purchases of home supplies, non-prescription drugs,
and entertainment).

This limited evidence supports the claims that STH are more likely to

reduce rather than increase consumption of non-exempt items, although further analysis is
needed to determine the magnitude of this effect.

Temporal Shifting of Consumption
Households may also adapt their consumption behavior to take into account an STH by
reducing their consumption of items that are tax-exempt on STHs in the periods preceding or
following an STH. Most STHs are annual events, and there are often many advertisements and

significant press coverage in the days and even weeks before an STH.13 The ephemeral nature of
STHs provides households with the incentive to defer the consumption of items that will be taxexempt during an STH.
In the marketing literature, there exists a large body of work on households delaying
consumption in advance of sales promotions that has similar application in the STH context.
Given that the items that are tax-exempt are storable goods, consumers may also engage in
stockpiling behavior that would depress consumption following an STH. Hendel and Nevo
(2004) provide a literature survey on stockpiling behavior in response to promotions, and
conclude that there exists evidence of such behavior. Further, Hendel and Nevo (2006) apply a
model of consumer inventory holding to scanner and household-level data and find that when
households make purchases during sales they are more likely to postpone their next purchase.
In order to test whether households reduce consumption in periods before and after the
STH, we estimate the following model:

∑

_

_

∑

_

(2)

Model (2) is nearly identical to model (1), but the model (2) includes indicator variables denoting
time periods either preceding (

_

_

) or following (

) an STH. These time periods

consist of consecutive days. The identification strategy is same as that used in model (1); by still
including the calendar date fixed effects, the variables

_

and

_

examine the

consumption patterns of households on the same calendar dates in states that have an STH with

13

An exception to this was the first STH in Massachusetts, when the STH began two days after the state legislature
passed the legislation.

those in states that do not—the only difference is that instead of focusing on STH dates, we now
extend our analysis to other dates.
For our main specification, the time periods we analyze are the dates two weeks and three
weeks before and after an STH.14 We do not look at consumption in the period a week before
and a week after the STH, dropping these observations for households in states with STHs
because we are concerned about recall bias. If some households do shop on a STH but report
their purchases as occurring on a non-STH, we may observe an increase in consumption in the
days occurring around a STH, artificially increasing our estimates of household behavior before
or after a STH. This week long recall window is based on the belief that while households may
not remember the exact day of their purchases, they will remember the exact week. In addition,
by dropping these dates, we drop all the observations for non-STH days from weekly Diaries that
contain at least one day that occurs on a STH . Because we do not employ a difference-indifference model, this non-overlap between the treatment period and non-treatment period does
not affect our estimation strategy.15
Table 9 presents the coefficients for

_

,

_

, and

_

. Each column

of the Table represents the results from one regression. Similar to the other tables, the dependent
variables are daily expenditure categories. Our results provide mixed and unconvincing
evidence of shifting behavior. In the period three weeks before the sales tax holiday, we observe
depressed consumption in clothing, shoes, and school supplies. Further, a number of the
coefficients are statistically significant at the 10 percent and one percent level, specifically in the
consumption of clothing and shoes. These amounts all tend to be smaller than the subsequent
increases seen on STH. By contrast, we observe positive coefficients in the week both before
14

Other period lengths yield similar results.
Recall bias would bias our main results down because some purchasing that occurred during an STH would be
reported as occurring outside of it.

15

and after the recall period surrounding a STH, although nearly all of these are statistically
insignificant. It is challenging to develop a theory as to why consumption would fall three weeks
before an STH, but increase two weeks before and after. This suggests that different data are
probably needed to address whether an STH causes consumption shifting16.

Robustness—Seasonal Demand Peaks
One concern with these results is that we may be finding effects of STHs on consumption
because states time sales tax holidays to match state specific periods of increased seasonal
demand. If this is the case, our regressions are merely capturing the effects of seasonal demand
spikes and not the tax incentives of the STH. We attempt to address this issue by comparing
consumption patterns across states (some with STHs and some without) on the same dates.
However, if seasonal demand peaks are heterogeneous across states, we could accidentally
engage in an apples-to-oranges comparison of households experiencing different levels of
seasonal demand occurring on the same dates. If indeed the timing of seasonal demand differs
by state and policymakers place STHs on dates with peak seasonal demand, our results may be
biased upward.
To investigate what role state-specific seasonal demand may have in interfering with our
results, we run the following experiment. Focusing on the years before the first STH in New
York, we analyze how individuals in a state behave on days that would eventually have an STH
in their state. This experiment is valid if we assume that timing of state-specific demand shocks

16

For example, data with longer panel periods so that one may observe household consumption behavior that occurs
before, during, and after a STH.

is invariant across time. If we observe households increasing consumption on days that in later
years would have a sales tax holiday, our main findings may be biased.
The main challenge of our approach is determining exactly when to project the simulated
STH. STHs are never the on the exact same date but rather vary slightly year to year. We
simulate sales tax holidays in two manners. First, we create a modal sales tax holiday for each
state based on the modal month, week of the month, and week day of the month for the sales tax
holidays enacted since 1997. For instance, for West Virginia, we determine that the modal sales
tax holiday begins on the first Friday in August. (In fact all West Virginia STHs have begun on
this date) 17. We also use the modal length of STH in determining the simulated duration
(although there is almost no variation in the lengths of STHs across time within a state). We
label these “modal” sales tax holidays.

Our second approach is to use the STH dates over the

period 2006–08 and project them backwards based on the month, day of week and week of
month when they occurred18. We label these “recent” STHs. We test the effects of these
simulated STHs for the period 1994–1996. We are restricted to these dates because state
identifiers are first introduced in the public use CEX in1994.
Table 10 summarizes the results of running model (1) on the simulated STHs. Column 1
reproduces the results from Table 3, and columns 2 and 3 show the coefficients of STH_Day for
the two different experiments: one using the modal STH dates and the other using the most
recent STH dates. The findings of this experiment suggest that our main results are indeed
driven by the temporary suspension of sales taxes during STHs and not by the heterogeneity of
seasonal demand across different states. In every instance where the coefficient of STH_Day

17

Using the modal month, week of the month, and weekday of the month matches the actual STH 77 percent of the
time.
18
For example, Connecticut’s 2008 STH occurred on Sunday, August 17th. The third Sunday of August in 1994
occurred on August 21st.

shows a statistically significant increase in the consumption of clothing and shoes during an
STH, the regressions using the simulated STHs show no corresponding positive response. For
example, compared with an increase of 0.068 items of children’s clothing and 0.010 items of
children’s shoes purchased on an STH, in the absence of the STH, the coefficient on STH_Day
is–0.022 (with a standard error of 0.015) for children’s clothing and 0.002 (with a standard error
of 0.005) for children’s shoes when using the most recent STH dates—both statistically
indistinguishable from zero. This difference, as well as the fact that they hold regardless of using
the modal STH dates or the most recent STH dates, supports the evidence that STHs lead to
increased consumption of tax-exempt items.
V. Conclusion
Our analysis of household consumption patterns during STHs reveals that the policy
leads to a sizeable increase in the purchasing of apparel (clothing and shoes) and school supply
items; our baseline findings show that on days that have an STH, households increase the
number of clothing and shoes purchased by over 40 percent relative to what they normally buy.
STHs that exempt either apparel accessories or computers from sales taxes do not lead to an
increase in the amount spent, the number of items purchased, or the probability of making a
purchase. Because policymakers are unable to target STHs to benefit specific types of
households, the policy creates the potential for households of different economic and familial
makeups to respond idiosyncratically. Indeed, we find that our baseline results vary both within
and across different types of households, categorized by income and/or household structure.
Within households, an STH causes increases in the purchases of only children’s apparel:
spending on children’s clothing increases 91 percent, and the number of children’s clothing and
shoes purchased increases around 130 percent. Across different households, we find those

households consisting of married couples with young children and earning over $30,000 per year
are the only ones that statistically significantly increase consumption during STHs. An aspect of
STHs that critics raise is that they may lead to consumers to engage in distortionary behavior.
Specifically, consumers may either shift consumption away from non-exempt items on days with
STH, or may suspend planned consumption to occur on days with a STH. We find some
evidence that both may occur, although they are based on results with little statistical accuracy.
We conclude that more research needs to be conducted to fully address these criticisms of the
policy, but our initial investigation does not find large distorting behavior.
One criticism of our approach may be that since states intentionally place STH during
periods of increased seasonal demand, our estimation strategy my fail to account for
heterogeneity of demand cycles that vary by state. We address this concern by simulating when
STH would occur in the years before states institute them, and seeing whether or not we still
observe an increase in consumption. Using two specifications, we find no such increase on days
that would later have a STH, supporting our findings that it is indeed the STH that cause
consumers to increase consumption and not simply the varied timing of state back-to-school
shopping seasons.
As a policy that aims to accomplish multiple goals for disparate parties, the STH falls
short of satisfying all of the policymakers’ stated intentions. STHs are designed to fulfill three
main goals: to provide monetary relief to lower-income families with children; to increase sales
for retailers; and to promote the consumption of specific goods that have some inherit social
value. Our main findings show that STH do lead households to increase the consumption of
children’s apparel and school supplies. However, our subsequent analysis reveals that it is only
the wealthiest households that statistically significantly take advantage of STHs. The major

theme our analysis reveals is that the STH is too blunt a policy tool for addressing the many
problems it seeks to resolve. As a method for stimulating economic activity for a limited number
of goods, the STH is shown to be quite effective. As tool for providing economic relief to
certain households, STH has fared quite poorly. Policymakers should be aware that the STH is
not a panacea—especially when weighing its effects against the foregone tax revenue.

References
Associated Press, 2010, “Sales Tax Holiday for Back-to-School Shopping,”
myFOXchicago.com, May 31, available at
www.myfoxchicago.com/dpp/news/metro/sales-tax-break-back-to-school-shopping20100531.
Besley, Timothy J. and Harvey S. Rosen (1999) “Sales Taxes and Prices: An Empirical
Analysis.” National Tax Journal. Vol. 52, June
Blattberg, Robert; Buesing, Thomas; Peacock, Peter and Subrata Sen. (1978) “Identifying the
Deal Prone Segment.” Journal of Marketing Research. Vol 15(3) August
Cole, Adam J. (2008) “Sales Tax Holidays, 1997-2007: A History.” State Tax Notes.
Vol. 47, Num 13
Cole, Adam J. (2009a) “Christmas in August: Prices and Quantities During a Sales Tax Holiday”
Dissertation, University of Michigan
Cole, Adam J. (2009b) “The Fiscal Impact of Sales Tax Holidays” Dissertation, University of
Michigan
Doyle, Joseph J. Jr and Krislert Samphantharak (2008) “$2.00 Gas! Studying the effects of a gas
tax moratorium” Journal of Public Economics, Vol. 92(3-4) April
Fullerton, Thomas M. Jr. (1987) “Rational Reactions to Temporary Sales Tax Legislation: An
Idaho Case Study” Public Budgeting & Finance Vol. 7, Num. 2 Summer
Harper, Richard K.; Hawkins, Richard R.; Martin, Gregory S.; Sjolander, Richard. (2003).
“Price Effects Around a Sales Tax Holiday: An Exploratory Study” Public Budgeting &
Finance. Vol. 23(4)
Hawkins, Richard R. and John L. Mikesell. (2001) “Six Reasons to Hate Your Sales Tax
Holiday” State Tax Notes. Vol 20, March
Hendel, Igal and Aviv Nevo (2004) “Intertemporal Substitution and Storable Products” Journal
of European Economic Association. May
Hendel, Igal and Aviv Nevo (2006) “Sales and Consumer Inventory” The RAND Journal of
Economics. Vol 37(3) Autumn
Mogab, John W. and Michael J. Pisani (2007) “Shoppers’ Perceptions of the State Sales Tax
Holiday: A Case Study from Texas” American Journal of Business, Vol. 22(2) Fall

Table 1
Summary Statistics of Daily Expenditures (2008 Dollars)
Items
Total

Category
$ Spent
# Purchased
% Days

Mean/Percent
136.00
6.67
70.4 %

Std. Dev.
740.0
10.70

$ Spent
# Purchased
% Days

8.35
0.32
17.3 %

64.0
1.18

Clothing

$ Spent
# Purchased
% Days

4.31
0.17
7.0 %

30.3
0.84

Shoes

$ Spent
# Purchased
% Days

1.21
0.03
2.3 %

11.8
0.21

Accessories

$ Spent
# Purchased
% Days

1.09
0.03
2.3 %

37.8
0.22

Computers

$ Spent
# Purchased
% Days

0.72
0.01
0.5 %

29.4
0.08

School Supplies

$ Spent
# Purchased
% Days

1.02
0.09
5.3 %

15.7
0.46

$ Spent
# Purchased
% Days

11.80
3.42
33.9 %

31.4
8.18

$ Spent
# Purchased
% Days

115.00
2.93
65.6 %

729.0
4.10

STH Tax-Exempt
Exempt Total

STH Non-Exempt
Food

Other

Table 2
Frequency of STH, Demographics
Frequency of STH
Any STH
Clothing
Shoes
Accessories
Computers
School Supplies
Demographics
Age of Reference Person

Mean/Percent
0.37 %
0.33 %
0.27 %
0.09 %
0.11 %
0.17 %

Single Household
One Parent
Married – No Children
Married – Young Children
Married – Old Children
“Other” Household
Income: < 30,000
Income: 30,000-70,000
Income: > 70,000

48.3
(17.1)
51.3 %
82.8 %
2.6
(1.5)
27.0 %
6.0 %
21.4 %
20.1 %
7.1 %
18.3 %
34.9 %
36.3 %
28.8 %

Number of Observations
STH Observations

606,907
2,241

Reference Person is Male
Reference Person is White
Household Size

Table 3
The Effect of STH on the Amount Spent, Number of Items Purchased, and Percentage of Days Shopped
Clothing
Shoes
Accessories
STH_Day
β/Mean

$ Spent
1.106*
(0.655)
25.7 %

# Purchased
0.083***
(0.029)
49.2 %

Days %
0.014**
(0.007)
20.6 %

$ Spent
0.281
(0.317)
23.2 %

# Purchased
0.013*
(0.007)
45.4 %

Days %
0.010**
(0.005)
44.1 %

$ Spent
0.039
(0.539)
3.60 %

# Purchased
-0.006
(0.008)
-20.7 %

Days %
-0.004
(0.006)
-19.5 %

Computers
School Supplies
$ Spent
# Purchased
Days %
$ Spent
# Purchased Days %
STH_Day -0.269
-0.003*
-0.003
0.328
0.055*
-0.000
(0.305)
(0.002)
(0.002)
(0.304)
(0.032)
(0.007)
-37.5 %
-62.7 %
-62.0 %
32.0 %
63.5 %
-0.6 %
β/Mean
All regressions include controls for age, sex, and race of reference head, household composition and income controls, and state and calendar date fixed effects.
Standard Errors clustered at the household level

Table 4
The Distribution of Intra-Household Consumption of Clothing and Shoes
Clothing-Men
Clothing-Women
STH_Day
β/Mean

$ Spent
0.081
(0.270)
7.43 %

# Purchased
0.005
(0.008)
13.6 %

% Days
0.001
(0.003)
7.36 %

$ Spent
0.141
(0.342)
6.95 %

# Purchased
0.010
(0.012)
13.6 %

% Days
0.002
(0.005)
4.66 %

Clothing-Children
$ Spent
0.910**
(0.394)
91.4 %

# Purchased
0.068***
(0.020)
129 %

% Days
0.016***
(0.005)
62.8 %

Shoes-Men
Shoes-Women
Shoes-Children
$ Spent
# Purchased % Days
$ Spent
# Purchased
% Days
$ Spent
# Purchased
% Days
STH_Day -0.146
-0.001
-0.001
0.107
0.003
0.003
0.319
0.010**
0.008**
(0.123)
(0.002)
(0.002)
(0.174)
(0.004)
(0.003)
(0.207)
(0.005)
(0.003)
β/Mean
-36.7 %
-8.03 %
-12.3 %
19.1 %
25.0 %
28.0 %
126 %
130 %
124.9 %
All regressions include controls for age, sex, and race of reference head, household composition and income controls, and state and calendar date fixed effects.
Standard Errors clustered at the household level

Table 5
Heterogeneous Household’s Consumption of Clothing and Shoes
STH_Day
Clothing
Shoes
Household Composition
One Parent
β/Mean
Married-No Children
β/Mean
Married-Young Children
β/Mean
Married-Old Children
β/Mean
Single Households
β/Mean
“Other” Households
β/Mean
Income Groups
Income < $30,000
β/Mean
Income $30,000-$70,000
β/Mean
Income > $70,000

School Supply

$ Spent
0.155
(2.932)
3.27 %

# Purchased
0.121
(0.119)
62.4 %

$ Spent
0.030
(1.512)
1.86 %

# Purchased
-0.013
(0.023)
-33.8 %

$ Spent
-0.802
(1.086)
-81.7 %

# Purchased
-0.060
(0.082)
-80.4 %

-1.558
(1.039)
-37.0 %

-0.014
(0.038)
-10.3 %

-0.303
(0.555)
-30.0 %

-0.006
(0.009)
-28.0 %

0.038
(0.317)
4.19 %

-0.047*
(0.028)
-51.2 %

6.115**
(2.770)
96.5 %

0.277**
(0.121)
98.8 %

1.763
(1.253)
102.3 %

0.057*
(0.032)
128.8 %

0.503
(0.956)
32.5 %

0.202
(0.131)
154.4 %

4.377
(3.499)
69.3 %

0.215
(0.151)
95.2 %

-0.324
(1.121)
-18.7 %

-0.000
(0.024)
-0.42 %

0.171
(0.958)
8.26 %

0.346
(0.239)
293.4 %

-0.086
(0.642)
-4.00 %

0.016
(0.022)
20.4 %

0.290
(0.312)
44.5 %

0.013*
(0.008)
94.2 %

0.798
(0.885)
160 %

0.009
(0.019)
18.4 %

-0.072
(1.711)
-1.60 %

0.057
(0.056)
31.4 %

0.187
(0.675)
13.7 %

0.014
(0.016)
44.4 %

-0.050
(0.303)
-5.22 %

0.045
(0.083)
58.7 %

0.445
(0.629)
19.0 %

0.058
(0.039)
56.5 %

0.608
(0.482)
77.6 %

0.015
(0.010)
74.1 %

0.714
(0.798)
145.2 %

-0.022
(0.017)
-45.1 %

1.071
(0.924)
27.4 %

0.091*
(0.047)
54.3 %

-0.191
(0.397)
-16.0 %

0.007
(0.009)
25.5 %

0.218
(0.289)
24.4 %

0.135**
(0.067)
160.7 %

-0.007
(0.476)
-0.38 %

0.015
(0.052)
11.0 %

2.789
0.119*
0.538
0.018
(1.846)
(0.066)
(0.787)
(0.018)
38.8 %
48.1 %
30.8 %
48.7 %
β/Mean
Standard Errors clustered at the household level. Household composition regressions

Table 6
Heterogeneous Household’s and the Distribution of Intra-Household Consumption of Clothing and Shoes
STH_Day
Clothing-Men
Clothing-Women
Clothing-Children
Married-Young Children
β/Mean
Income $30,000-$70,000
β/Mean
Income > $70,000
β/Mean

STH_Day
Married-Young Children
β/Mean
Income $30,000-$70,000
β/Mean
Income > $70,000

$ Spent
0.992
(0.994)
77.8 %

# Purchased
0.039
(0.031)
85.1 %

$ Spent
1.822
(1.224)
87.1 %

# Purchased
0.053
(0.042)
64.2 %

$ Spent
3.078*
(1.824)
116.7 %

# Purchased
0.174**
(0.082)
123.2 %

0.244
(0.445)
24.1 %

0.008
(0.016)
22.0 %

0.614
(0.583)
33.6 %

0.029
(0.024)
40.7 %

0.288
(0.408)
32.1 %

0.049*
(0.027)
95.3 %

0.341
(0.798)
18.2 %

0.009
(0.018)
16.6 %

0.288
(0.881)
8.70 %

-0.001
(0.026)
-0.70 %

2.097*
(1.178)
129.6 %

0.109**
(0.051)
136.3 %

Shoes -Men

Shoes -Women

Shoes -Children

$ Spent
0.070
(0.491)
14.1 %

# Purchased
0.005
(0.008)
63.8 %

$ Spent
-0.237
(0.393)
-41.4 %

# Purchased
0.003
(0.012)
22.5 %

$ Spent
1.931**
(0.965)
294.7 %

# Purchased
0.048*
(0.026)
234.9 %

-0.270
(0.204)
-63.9 %

-0.001
(0.004)
-9.73 %

-0.016
(0.235)
-3.03 %

0.005
(0.006)
36.1 %

0.096
(0.208)
40.9 %

0.003
(0.005)
40.2 %

-0.508*
-0.006*
0.602
0.008
0.444
0.016
(0.271)
(0.003)
(0.528)
(0.011)
(0.479)
(0.013)
-89.1 %
-68.3 %
73.9 %
43.7 %
122.3 %
156.3 %
β/Mean
Errors clustered at the household level. Household composition regressions include income group controls and vice-versa

Table 7
Average Daily Expenditures in August on Non-STH Days
Household Type
Clothing
Shoes
School
Supply

Clothing + Shoes
+ School

Household Composition
One Parent

$ Spent
6.32
(43.4)

$ Spent
2.51
(16.8)

$ Spent
1.83
(18.0)

$ Spent
10.70
(56.5)

$ Sales Tax
0.53
(3.02)

Married-No Children

3.32
(27.3)

0.99
(10.2)

0.83
(16.2)

5.15
(36.8)

0.26
(2.01)

Married-Young Children

7.60
(46.9)

2.14
(17.0)

2.50
(24.5)

12.20
(66)

0.62
(3.50)

Married-Old Children

5.98
(33.4)

1.88
(12.8)

4.69
(44.3)

12.60
(63.5)

0.62
(3.26)

Single Households

1.93
(17.7)

0.70
(9.33)

0.66
(12.9)

3.31
(25.2)

0.17
(1.28)

“Other” Households

4.32
(28.5)

1.49
(1.03)

1.18
(17.9)

7.00
(40.3)

0.35
(2.08)

2.24
(19.0)

0.88
(9.92)

0.80
(13.8)

3.94
(29.7)

0.20
(1.56)

Income $30,000-$70,000

4.21
(27.4)

1.41
(12.5)

1.35
(20.3)

6.97
(40.6)

0.36
(2.13)

Income > $70,000

7.21
(47.0)

2.03
(16.0)

2.69
(28.4)

12.0
(65.3)

0.59
(3.45)

Income Groups
Income < $30,000

Table 8
The Effect of STH on Non-Exempt Items
Non-Exempt Items¹

Food at Home

$ Spent
# Purchased % Days
$ Spent
# Purchased
% Days
STH_Day -3.302
-0.066
-0.016
-0.153
0.235
0.010
(13.578)
(0.121)
(0.015)
(0.625)
(0.179)
(0.012)
-2.86 %
-2.25 %
-2.38 %
-1.29 %
6.86 %
3.00 %
β/Mean
All regressions include controls for age, sex, and race of reference head, household composition and income
controls, and state and calendar date fixed effects. Standard Errors clustered at the household level
1. Excluding food purchased at home

Table 9
Does STH cause shifting of consumption? Days before and after STH
Time
Clothing
Levels
15-21 Days Before
8-14 Days Before
Sales Tax Holiday
8-14 Days After
15-21 Days After

$ Spent
-0.161
(0.785)
0.306
(0.824)
1.150*
(0.662)
0.701
(0.610)
-0.563
(0.446)

# Purchased
-0.031*
(0.018)
0.002
(0.019)
0.084***
(0.029)
0.051**
(0.025)
-0.021
(0.018)

Number of Observations
600,591
600,591
Standard Errors clustered at the household level.

Shoes

School Supply

% Days
-0.014***
(0.005)
0.005
(0.006)
0.015**
(0.007)
0.007
(0.006)
0.001
(0.006)

$ Spent
-0.145
(0.368)
0.211
(0.284)
0.297
(0.322)
0.262
(0.242)
0.013
(0.259)

# Purchased
-0.008*
(0.005)
0.006
(0.005)
0.013*
(0.007)
0.008
(0.006)
-0.005
(0.004)

% Days
-0.008***
(0.003)
0.004
(0.004)
0.011**
(0.005)
0.006
(0.004)
-0.002
(0.003)

$ Spent
-0.081
(0.111)
0.170
(0.276)
0.385
(0.300)
0.325
(0.631)
0.387
(0.575)

# Purchased
-0.002
(0.006)
0.003
(0.008)
0.043*
(0.023)
0.015
(0.016)
0.005
(0.012)

% Days
-0.001
(0.004)
0.004
(0.004)
0.005
(0.006)
0.005
(0.005)
0.005
(0.005)

600,591

600,591

600,591

600,591

600,591

600,591

600,591

Table 10 Simulated Sales Tax Holidays
STH_Day
Actual STH
Total
Clothing
Shoes
School Supplies
Men’s
Clothing
Shoes
Women’s
Clothing
Shoes
Children
Clothing
Shoes

Modal-STH

Recent-STH

$ Spent

# Purchased

$ Spent

# Purchased

$ Spent

# Purchased

1.106*
(0.655)
0.281
(0.317)

0.083***
(0.029)
0.013*
(0.007)

-0.170
(0.894)
0.767
(0.605)

-0.026
(0.027)
0.005
(0.008)

-0.194
(0.966)
0.281
(0.397)

-0.022
(0.029)
0.003
(0.008)

0.328
(0.304)

0.055*
(0.032)

0.259
(0.530)

0.005
(0.021)

0.383
(0.480)

0.017
(0.024)

0.081
(0.270)
-0.146
(0.123)
0.141
(0.342)
0.107
(0.174)
0.910**
(0.394)
0.319
(0.207)

0.005

-0.447

-0.008

0.084

-0.003

(0.008)

(0.350)

(0.009)

(0.415)

(0.008)

-0.001

0.062

0.001

-0.039

-0.000

(0.002)

(0.203)

(0.003)

(0.172)

(0.002)

0.010

0.246

-0.000

0.164

0.002

(0.012)

(0.633)

(0.015)

(0.818)

(0.019)

0.003
(0.004)

0.235
(0.393)

-0.001
(0.004)

0.242
(0.292)

0.001
(0.004)

0.068***

-0.053

-0.017

-0.510*

-0.022

(0.020)

(0.475)

(0.015)

(0.304)

(0.015)

0.010**

0.471

0.006

0.078

0.002

(0.005)

(0.420)

(0.006)

(0.171)

(0.005)

All regressions include controls for age, sex, and race of reference head, household composition and income
controls, and state and calendar date fixed effects. Standard Errors clustered at the household level

Working Paper Series
A series of research studies on regional economic issues relating to the Seventh Federal
Reserve District, and on financial and economic topics.
Risk Taking and the Quality of Informal Insurance: Gambling and Remittances in Thailand
Douglas L. Miller and Anna L. Paulson

WP-07-01

Fast Micro and Slow Macro: Can Aggregation Explain the Persistence of Inflation?
Filippo Altissimo, Benoît Mojon, and Paolo Zaffaroni

WP-07-02

Assessing a Decade of Interstate Bank Branching
Christian Johnson and Tara Rice

WP-07-03

Debit Card and Cash Usage: A Cross-Country Analysis
Gene Amromin and Sujit Chakravorti

WP-07-04

The Age of Reason: Financial Decisions Over the Lifecycle
Sumit Agarwal, John C. Driscoll, Xavier Gabaix, and David Laibson

WP-07-05

Information Acquisition in Financial Markets: a Correction
Gadi Barlevy and Pietro Veronesi

WP-07-06

Monetary Policy, Output Composition and the Great Moderation
Benoît Mojon

WP-07-07

Estate Taxation, Entrepreneurship, and Wealth
Marco Cagetti and Mariacristina De Nardi

WP-07-08

Conflict of Interest and Certification in the U.S. IPO Market
Luca Benzoni and Carola Schenone

WP-07-09

The Reaction of Consumer Spending and Debt to Tax Rebates –
Evidence from Consumer Credit Data
Sumit Agarwal, Chunlin Liu, and Nicholas S. Souleles

WP-07-10

Portfolio Choice over the Life-Cycle when the Stock and Labor Markets are Cointegrated
Luca Benzoni, Pierre Collin-Dufresne, and Robert S. Goldstein

WP-07-11

Nonparametric Analysis of Intergenerational Income Mobility
with Application to the United States
Debopam Bhattacharya and Bhashkar Mazumder

WP-07-12

How the Credit Channel Works: Differentiating the Bank Lending Channel
and the Balance Sheet Channel
Lamont K. Black and Richard J. Rosen

WP-07-13

Labor Market Transitions and Self-Employment
Ellen R. Rissman

WP-07-14

First-Time Home Buyers and Residential Investment Volatility
Jonas D.M. Fisher and Martin Gervais

WP-07-15

1

Working Paper Series (continued)
Establishments Dynamics and Matching Frictions in Classical Competitive Equilibrium
Marcelo Veracierto

WP-07-16

Technology’s Edge: The Educational Benefits of Computer-Aided Instruction
Lisa Barrow, Lisa Markman, and Cecilia Elena Rouse

WP-07-17

The Widow’s Offering: Inheritance, Family Structure, and the Charitable Gifts of Women
Leslie McGranahan

WP-07-18

Incomplete Information and the Timing to Adjust Labor: Evidence from the
Lead-Lag Relationship between Temporary Help Employment and Permanent Employment
Sainan Jin, Yukako Ono, and Qinghua Zhang

WP-07-19

A Conversation with 590 Nascent Entrepreneurs
Jeffrey R. Campbell and Mariacristina De Nardi

WP-07-20

Cyclical Dumping and US Antidumping Protection: 1980-2001
Meredith A. Crowley

WP-07-21

Health Capital and the Prenatal Environment:
The Effect of Maternal Fasting During Pregnancy
Douglas Almond and Bhashkar Mazumder

WP-07-22

The Spending and Debt Response to Minimum Wage Hikes
Daniel Aaronson, Sumit Agarwal, and Eric French

WP-07-23

The Impact of Mexican Immigrants on U.S. Wage Structure
Maude Toussaint-Comeau

WP-07-24

A Leverage-based Model of Speculative Bubbles
Gadi Barlevy

WP-08-01

Displacement, Asymmetric Information and Heterogeneous Human Capital
Luojia Hu and Christopher Taber

WP-08-02

BankCaR (Bank Capital-at-Risk): A credit risk model for US commercial bank charge-offs
Jon Frye and Eduard Pelz

WP-08-03

Bank Lending, Financing Constraints and SME Investment
Santiago Carbó-Valverde, Francisco Rodríguez-Fernández, and Gregory F. Udell

WP-08-04

Global Inflation
Matteo Ciccarelli and Benoît Mojon

WP-08-05

Scale and the Origins of Structural Change
Francisco J. Buera and Joseph P. Kaboski

WP-08-06

Inventories, Lumpy Trade, and Large Devaluations
George Alessandria, Joseph P. Kaboski, and Virgiliu Midrigan

WP-08-07

2

Working Paper Series (continued)
School Vouchers and Student Achievement: Recent Evidence, Remaining Questions
Cecilia Elena Rouse and Lisa Barrow
Does It Pay to Read Your Junk Mail? Evidence of the Effect of Advertising on
Home Equity Credit Choices
Sumit Agarwal and Brent W. Ambrose

WP-08-08

WP-08-09

The Choice between Arm’s-Length and Relationship Debt: Evidence from eLoans
Sumit Agarwal and Robert Hauswald

WP-08-10

Consumer Choice and Merchant Acceptance of Payment Media
Wilko Bolt and Sujit Chakravorti

WP-08-11

Investment Shocks and Business Cycles
Alejandro Justiniano, Giorgio E. Primiceri, and Andrea Tambalotti

WP-08-12

New Vehicle Characteristics and the Cost of the
Corporate Average Fuel Economy Standard
Thomas Klier and Joshua Linn

WP-08-13

Realized Volatility
Torben G. Andersen and Luca Benzoni

WP-08-14

Revenue Bubbles and Structural Deficits: What’s a state to do?
Richard Mattoon and Leslie McGranahan

WP-08-15

The role of lenders in the home price boom
Richard J. Rosen

WP-08-16

Bank Crises and Investor Confidence
Una Okonkwo Osili and Anna Paulson

WP-08-17

Life Expectancy and Old Age Savings
Mariacristina De Nardi, Eric French, and John Bailey Jones

WP-08-18

Remittance Behavior among New U.S. Immigrants
Katherine Meckel

WP-08-19

Birth Cohort and the Black-White Achievement Gap:
The Roles of Access and Health Soon After Birth
Kenneth Y. Chay, Jonathan Guryan, and Bhashkar Mazumder

WP-08-20

Public Investment and Budget Rules for State vs. Local Governments
Marco Bassetto

WP-08-21

Why Has Home Ownership Fallen Among the Young?
Jonas D.M. Fisher and Martin Gervais

WP-09-01

Why do the Elderly Save? The Role of Medical Expenses
Mariacristina De Nardi, Eric French, and John Bailey Jones

WP-09-02

3

Working Paper Series (continued)
Using Stock Returns to Identify Government Spending Shocks
Jonas D.M. Fisher and Ryan Peters

WP-09-03

Stochastic Volatility
Torben G. Andersen and Luca Benzoni

WP-09-04

The Effect of Disability Insurance Receipt on Labor Supply
Eric French and Jae Song

WP-09-05

CEO Overconfidence and Dividend Policy
Sanjay Deshmukh, Anand M. Goel, and Keith M. Howe

WP-09-06

Do Financial Counseling Mandates Improve Mortgage Choice and Performance?
Evidence from a Legislative Experiment
Sumit Agarwal,Gene Amromin, Itzhak Ben-David, Souphala Chomsisengphet,
and Douglas D. Evanoff

WP-09-07

Perverse Incentives at the Banks? Evidence from a Natural Experiment
Sumit Agarwal and Faye H. Wang

WP-09-08

Pay for Percentile
Gadi Barlevy and Derek Neal

WP-09-09

The Life and Times of Nicolas Dutot
François R. Velde

WP-09-10

Regulating Two-Sided Markets: An Empirical Investigation
Santiago Carbó Valverde, Sujit Chakravorti, and Francisco Rodriguez Fernandez

WP-09-11

The Case of the Undying Debt
François R. Velde

WP-09-12

Paying for Performance: The Education Impacts of a Community College Scholarship
Program for Low-income Adults
Lisa Barrow, Lashawn Richburg-Hayes, Cecilia Elena Rouse, and Thomas Brock
Establishments Dynamics, Vacancies and Unemployment: A Neoclassical Synthesis
Marcelo Veracierto

WP-09-13

WP-09-14

The Price of Gasoline and the Demand for Fuel Economy:
Evidence from Monthly New Vehicles Sales Data
Thomas Klier and Joshua Linn

WP-09-15

Estimation of a Transformation Model with Truncation,
Interval Observation and Time-Varying Covariates
Bo E. Honoré and Luojia Hu

WP-09-16

Self-Enforcing Trade Agreements: Evidence from Antidumping Policy
Chad P. Bown and Meredith A. Crowley

WP-09-17

Too much right can make a wrong: Setting the stage for the financial crisis
Richard J. Rosen

WP-09-18

4

Working Paper Series (continued)
Can Structural Small Open Economy Models Account
for the Influence of Foreign Disturbances?
Alejandro Justiniano and Bruce Preston

WP-09-19

Liquidity Constraints of the Middle Class
Jeffrey R. Campbell and Zvi Hercowitz

WP-09-20

Monetary Policy and Uncertainty in an Empirical Small Open Economy Model
Alejandro Justiniano and Bruce Preston

WP-09-21

Firm boundaries and buyer-supplier match in market transaction:
IT system procurement of U.S. credit unions
Yukako Ono and Junichi Suzuki
Health and the Savings of Insured Versus Uninsured, Working-Age Households in the U.S.
Maude Toussaint-Comeau and Jonathan Hartley

WP-09-22

WP-09-23

The Economics of “Radiator Springs:” Industry Dynamics, Sunk Costs, and
Spatial Demand Shifts
Jeffrey R. Campbell and Thomas N. Hubbard

WP-09-24

On the Relationship between Mobility, Population Growth, and
Capital Spending in the United States
Marco Bassetto and Leslie McGranahan

WP-09-25

The Impact of Rosenwald Schools on Black Achievement
Daniel Aaronson and Bhashkar Mazumder

WP-09-26

Comment on “Letting Different Views about Business Cycles Compete”
Jonas D.M. Fisher

WP-10-01

Macroeconomic Implications of Agglomeration
Morris A. Davis, Jonas D.M. Fisher and Toni M. Whited

WP-10-02

Accounting for non-annuitization
Svetlana Pashchenko

WP-10-03

Robustness and Macroeconomic Policy
Gadi Barlevy

WP-10-04

Benefits of Relationship Banking: Evidence from Consumer Credit Markets
Sumit Agarwal, Souphala Chomsisengphet, Chunlin Liu, and Nicholas S. Souleles

WP-10-05

The Effect of Sales Tax Holidays on Household Consumption Patterns
Nathan Marwell and Leslie McGranahan

WP-10-06

5