Figure 1: Average Tariff Rate and Tariff Exposure
Notes: This figure shows the monthly average tariff rate
(left panel) and tariff exposure (right panel) over the sample period
for products in the Numerator panel. For each year-month, we calculate
the mean across all observations. The tariff rate is the ratio of duties
collected to import value (Equation 1). Tariff
exposure multiplies the tariff rate by import penetration (Equation 3). The red
dashed vertical line marks February 2025, when tariffs were
implemented.
Two-panel line chart showing monthly averages from January 2024 through December 2025. The left panel plots the tariff rate in percent; the right panel plots tariff exposure in percent. A red dashed vertical line marks February 2025 in both panels. The area under each line is shaded, red in the left panel and blue in the right panel.
Left panel (Tariff Rate): The series is flat at roughly 3.4 percent from January 2024 through January 2025. It rises sharply beginning in February 2025, passing 10 percent by around May 2025 and continuing to climb through the second half of the year. It peaks at approximately 17 percent in late 2025 and ends near 16 percent in December 2025.
Right panel (Tariff Exposure): The series is flat at roughly 0.9 percent from January 2024 through January 2025. It rises sharply after February 2025, reaching approximately 3.2 percent by mid-2025 and continuing upward to a peak near 4.7 percent in late 2025, ending at approximately 4.5 percent in December 2025.
Figure 2: Event Study: Price, Spending, and Quantity Responses to Tariffs
Notes: The figure plots coefficients from event study
versions of Equations (4) and (5) run on Numerator
panel data, interacting the treatment variable with year-month
indicators (January 2025 omitted as reference). The top row uses the
tariff rate; the bottom row uses the tariff exposure measure. Left
panels show price (log Fisher Price Index); right panels show quantity
and spending (both in logs and spending is real spending). All
specifications include year-month and panel-unit-by-calendar-month
interaction fixed effects, with panel-specific linear pre-trends in the
price equation. Shaded areas indicate 95% confidence intervals based on
standard errors clustered at the HS4 code level. The vertical dashed
line marks February 2025, the tariff implementation date.
Four-panel chart arranged in a 2x2 grid, with coefficients measured in percentage points on the vertical axes and year-months from January 2024 through December 2025 on the horizontal axes. The top row uses the tariff rate specification; the bottom row uses the tariff exposure specification. Left panels show the price response (log Fisher Price Index); right panels show spending and quantity together, spending in orange and quantity in green. A dark red dashed vertical line marks February 2025 in all panels. Shaded bands indicate 95 percent confidence intervals.
Top-left (Price, Tariff Rate): Coefficients fluctuate around zero through the pre-tariff period, dipping to roughly negative 0.5 in some months. Around February 2025 the point estimate falls to about negative 0.3, then turns positive from spring 2025 onward and settles at approximately 0.1 to 0.2 percentage points through the end of the sample, with confidence bands that narrow considerably in the post period.
Top-right (Spending and Quantity, Tariff Rate): Pre-period coefficients for both series move around zero, with quantity above spending in early 2024 and a local peak just before February 2025. After the tariff date both series move below zero and stay there, settling at roughly negative 0.5 to negative 0.7 percentage points by late 2025. Spending and quantity track each other closely throughout the post period.
Bottom-left (Price, Tariff Exposure): The vertical scale is wider, running from about negative 5 to positive 2.5. Pre-period coefficients are noisy with very wide confidence bands, including a dip to roughly negative 1.7 at the February 2025 reference window. After February 2025 the coefficients rise to slightly positive values near 0.2 and the confidence bands collapse sharply.
Bottom-right (Spending and Quantity, Tariff Exposure): The vertical scale runs from about negative 3 to positive 6. Pre-period coefficients are imprecisely estimated with wide bands. After February 2025 both spending and quantity move negative, settling at approximately negative 1 to negative 1.5 percentage points through December 2025, with much tighter confidence bands than in the pre period.
Figure 3: Welfare Losses as Share of Approximate Income
Notes: The figure presents three welfare measures: net
spending effect, deadweight loss, and Sato-Vartia cost-of-living index,
as a percentage of approximate household income. See Equations (10)–(16) for measure
definitions. Income is approximated using bracket midpoints: $20,000
(low), $82,500 (middle), $187,500 (high). The high-income midpoint is a
lower bound on mean income in this group. The deadweight loss and
Sato-Vartia bars are visually almost indistinguishable, reflecting the
close agreement between the two price-anchored welfare measures.
Grouped bar chart showing three welfare measures (Net Spending Effect in gray, Deadweight Loss in red, and Sato-Vartia in dark navy) as a percentage of approximate household income, for three income groups (Low, Middle, High). Bars extend downward from zero, indicating welfare losses, and each bar is labeled with its value.
Low-income group: The Net Spending Effect bar reaches 1.38 percent of income. The Deadweight Loss bar reaches 0.72 percent. The Sato-Vartia bar reaches 0.69 percent. The Deadweight Loss and Sato-Vartia bars are visually almost indistinguishable.
Middle-income group: The Net Spending Effect bar reaches 0.48 percent. The Deadweight Loss bar reaches 0.17 percent. The Sato-Vartia bar reaches 0.17 percent.
High-income group: The Net Spending Effect bar reaches 0.28 percent. The Deadweight Loss bar reaches 0.07 percent. The Sato-Vartia bar reaches 0.07 percent.
The chart shows that the welfare burden is sharply regressive across all three measures, with low-income households bearing the largest cost as a share of income.
Figure 4: Household Sentiment and Intentions from the Survey
Notes: Panel A reports the share of respondents in the
Numerator survey selecting each Q5 concern item in each of the six
biweekly waves. Panel B reports the share selecting each Q7
anticipated-behavior item.
Two-panel grouped bar chart showing survey responses across six biweekly waves (weeks of 4/21, 5/5, 5/19, 6/2, 6/16, and 6/30), each wave shown in a different shade running from dark purple to yellow. Items are ordered from most to least frequently selected.
Panel A (Tariff Concerns, Q5): "Higher prices on everyday goods" is the most common concern at approximately 64 percent in every wave. "General inflation/Rising cost of goods" follows at approximately 59 to 61 percent. "Higher prices on non-essential items" is next at approximately 49 percent, then "Limited availability of certain products" at approximately 44 to 47 percent. "Impact on my job or industry" declines across waves from approximately 42 to 32 percent. "Impact on the stock market/investments" is approximately 30 to 33 percent, "Increased cost of healthcare or insurance" approximately 26 to 30 percent, and "Slower economic growth" approximately 18 to 20 percent. "No concerns" is selected by approximately 9 percent. Responses are broadly stable across the six waves.
Panel B (Anticipated Reactions, Q7): "Cut back spending on non-essentials" is the most common at approximately 46 to 49 percent. "Look for sales or coupons" follows at approximately 42 to 43 percent, then "Delay non-essential/big-ticket purchases" at approximately 31 to 34 percent. "Switch to lower-priced retailers" and "Buy fewer imported goods" are each around 28 to 32 percent, and "Delay purchases until prices stabilize" around 26 to 30 percent. "Switch to U.S.-made alternatives" is approximately 21 to 25 percent, "Buy/stock up earlier" approximately 20 to 22 percent, and "Reallocate budget to essentials" approximately 18 to 19 percent. "Increase budget for certain items" is approximately 8 to 9 percent and "I don’t expect to make any changes" approximately 7 to 8 percent. Responses are broadly stable across the six waves.
Figure A.1: Numerator Panel Demographics
Notes: Numerator Total Commerce Static Panel weighted as of
March 2025. Values shown as percentages. Blue bars show Numerator
panelists before applying demographic weights; red bars show the panel
after demographic weights are applied; yellow bars shows the American
Community Survey (ACS) implied demographics in 2024.
Eight-panel grouped horizontal bar chart comparing three series: Raw (Pre-Weighting) Numerator panelists in blue, Weighted Numerator panelists in red, and American Community Survey (ACS) targets in yellow, across eight demographic dimensions as of March 2025. Values are shown as percentages and labeled on each bar. In most dimensions the weighted and ACS bars carry identical values, so the red and yellow bars are the same length.
Age: Six groups. 18-24 is 4.4 weighted and ACS versus 2.0 raw; 25-34 is 15.8 versus 14.1; 35-44 is 17.4 versus 21.2; 45-54 is 16.2 versus 20.6; 55-64 is 17.6 versus 21.4; and 65+ is 28.6 versus 20.6 raw.
Income: Six brackets. Under $20K is 10.2 weighted and ACS versus 10.3 raw; $20k-40k is 13.2 versus 13.7; $40k-60k is 13.9 versus 15.4; $60k-80k is 11.8 versus 13.6; $80k-125k is 19.7 versus 23.9; and $125k+ is 31.7 versus 23.0 raw.
Ethnicity: Five categories. White/Caucasian is 64.0 weighted and ACS versus 68.2 raw; Black or African American is 13.7 versus 10.6; Hispanic/Latino is 14.9 versus 12.3; Asian is 5.9 versus 6.1; and Other is 1.5 versus 1.9 raw.
Gender: Female is 73.0 weighted and ACS versus 81.2 raw; Male is 27.0 versus 18.8 raw.
Children Under 17: Has Child is 27.2 weighted and ACS versus 34.5 raw; Has No Child is 72.8 versus 65.5 raw.
Household Size: Five groups. One-person households are 29.1 weighted and ACS versus 18.8 raw; two-person are 34.7 versus 36.1; three-person are 15.0 versus 18.0; four-person are 12.4 versus 15.5; and 5+ are 8.8 versus 11.6 raw.
Urbanicity: Rural is 27.2 weighted and ACS versus 29.6 raw; Suburban is 37.9 versus 38.6; Urban is 34.9 versus 31.8 raw.
Census Division: Nine divisions. South Atlantic is the largest at approximately 20.6 weighted versus 21.6 raw and 19.7 ACS. The remaining divisions range from roughly 4 to 16 percent, with the weighted panel close to ACS targets throughout.
Figure A.2: Retail sales, total excluding motor vehicles and parts (seasonally
adjusted) constructed using Census Bureau estimates and Numerator panel
data
Notes: The red solid line is the Census Bureau’s
seasonally-adjusted retail sales series excluding motor vehicles and
parts. The blue dashed line is the retail spending measure we construct
from the Numerator data, which we seasonally adjust. The gray
dash-dotted line is the non-seasonally adjusted Census series that we
then seasonally adjust. The data are monthly from January 2018 through
December 2025. Authors’ seasonal adjustment method uses X13-ARIMA-SEATS.
All three lines are indexed to 100 in 2019.
Source: Census Bureau and Numerator, and authors’
own calculations.
Line chart with three series from January 2018 through December 2025, indexed to 100 in 2019. The red solid line is the Census series, seasonally adjusted. The black dash-dotted line is the Census series using the authors’ seasonal adjustment methods. The blue dashed line is the Numerator series using the authors’ seasonal adjustment methods.
All three series move together closely. They begin near 96 in early 2018 and rise gradually to about 100 by 2019. There is a sharp dip to approximately 84 in April 2020, followed by a rapid rebound and a jump to roughly 116 in early 2021, with the Numerator series running slightly above the Census series through 2021. The series then flatten near 113 to 116 through 2022 and 2023 before resuming a steady upward drift, ending near 125 at the end of 2025.
Figure A.3: Numerator Fisher Price Index versus the PCE deflator for food
Notes: Left panel: Numerator price index for January 2018
through December 2025; BEA PCE price index for food for January 2018
through November 2025. Fisher Price Index constructed using Numerator
data in solid blue line, BEA’s PCE for food in dotted green line;
January 2018=100. The PCE data is seasonally adjusted by the BEA and the
Fisher Price Index is seasonally adjusted using X13-ARIMA-SEATS. Right
panel: Data for January 2019 through November 2025; the year-over-year
percent change in the Fisher Price Index is constructed using Numerator
data and shown as the solid blue line, the year-over-year percent change
for the PCE Price Index for food is shown as the dotted green
line.
Source: Numerator FPI calculated using Numerator
Panel Data and based on authors’ calculations. PCE Deflator for Food is
from the Federal Reserve Bank of St. Louis FRED, series
DFXARG3M086SBEA.
Two-panel chart. In both panels the Numerator FPI is a solid blue line and PCE Food is a dotted green line.
Panel A (Numerator FPI): Index levels from January 2018 through December 2025, with January 2018 equal to 100. Both series are flat near 100 through 2019, begin rising in 2020, and climb steeply from 2021 through 2022. The PCE food series runs slightly above the Numerator FPI through the middle of the sample, and the two converge toward the end. The Numerator FPI ends at 130.7.
Panel B (12-month Percent Change): Year-over-year percent changes from January 2019 through November 2025. Both series are near zero in 2019, spike to roughly 5 percent in 2020, fall back in 2021, then rise sharply to a peak near 13 percent in 2022 before decelerating through 2023 to around 1 to 2 percent. Both show a modest pickup in 2025, and the Numerator FPI ends at 1.9 percent.
Figure A.4: Income-Specific Fisher Price Indices
Notes: The figure plots the Fisher Price Index constructed
separately for each income group (Low, Middle, High) from January 2024
through December 2025, expressed as cumulative percentage change from
January 2024. The vertical dashed line marks February 2025, when tariffs
were implemented. Income groups are defined as in Section 3.3.
Line chart showing cumulative percentage change from January 2024 for three income groups (Low in red, Middle in blue, High in green) from January 2024 through December 2025. A red dotted vertical line labeled "Tariffs begin" marks February 2025.
All three series begin at zero in January 2024. The middle- and high-income series dip slightly below zero through the first half of 2024 while the low-income series stays at or above zero. From mid-2024 onward all three rise, with the low-income series consistently highest. By January 2025 the low-income series is at roughly 2.4 percent, with middle and high near 1.7 percent.
After February 2025 all three continue climbing, and the low-income series stays above the other two for most of the post-tariff period. By December 2025 the low-income series reaches approximately 4.5 percent, middle approximately 4.5 percent, and high approximately 3.7 percent cumulative price growth.
Figure B.1: Country-of-Origin Validation: Actual vs. Proxy Tariff Measures
Notes: Each point is one HS4 code (\(N = 158\)), using all products with an
observed country of origin from the merged Label Insight and Open Facts
sources. Panel A shows tariff rates: the \(x\)-axis is the proxy tariff rate (duties
over total imports for each HS4 code, from trade data); the \(y\)-axis is the actual tariff rate built
from products with verified origin using spending-weighted origin shares
(Equation B.1). Panel B
shows tariff exposure: the \(x\)-axis
multiplies the proxy rate by import penetration (Equation 3); the \(y\)-axis multiplies the actual rate by the
same import penetration (Equation B.1). The
dashed line is the 45-degree line and the solid line is the OLS fit. The
subtitle of each panel reports the Spearman rank correlation between the
actual and proxy measures. Panel A: Spearman \(= 0.51\). Panel B: Spearman \(= 0.71\).
Two-panel scatter plot, each point an HS4 code, with 158 codes in each panel. A red dashed 45-degree line and a solid black OLS fit line are shown in both panels. Each panel subtitle reports the sample size and the Spearman rank correlation.
Panel A (Validation A: Actual vs. Proxy Tariff Rate): The horizontal axis is the proxy tariff rate at the HS4 level, running from 0 to about 0.35; the vertical axis is the actual tariff rate from origin-known products, running from 0 to about 0.5. Points are widely dispersed, with a dense cluster below 0.1 on both axes and a scattering of points above 0.3 on the vertical axis. The OLS fit lies below the 45-degree line over most of the range, indicating attenuation. The subtitle reports Spearman equal to 0.505.
Panel B (Validation B: Actual vs. Proxy Tariff Exposure): The horizontal axis is the proxy tariff exposure, running from 0 to about 0.20; the vertical axis is the actual tariff exposure, running from 0 to about 0.25. Points cluster more tightly around the 45-degree line than in Panel A, and the OLS fit tracks the 45-degree line closely at low values before falling slightly below it at higher values. The subtitle reports Spearman equal to 0.712.
Figure C.1: Average Tariff Rate by Country
Notes: This figure shows the monthly average tariff rate
over the sample period by country. For each year-month, we calculate the
mean tariff rate across all observations. The tariff rate is measured in
percent.
Four-panel line chart showing monthly average tariff rates in percent from January 2024 through December 2025 for Canada, China, Mexico, and Rest of World (RoW). Each panel has its own vertical scale.
Canada: The tariff rate is flat at roughly 0.8 percent through January 2025, jumps to about 3.4 percent in the spring of 2025, drifts up to a peak near 5 percent in late 2025, and ends around 4.7 percent.
China: The tariff rate is roughly 19 to 20 percent through January 2025, reflecting pre-existing tariffs. It rises steeply after February 2025 to a peak of approximately 52 percent in mid-2025, falls back to around 44 to 45 percent, and declines further to approximately 37 percent by December 2025.
Mexico: The tariff rate is flat below 1 percent through January 2025, jumps to about 3.7 percent in spring 2025, rises to a peak near 4.4 percent in late 2025, and ends around 4.2 percent.
RoW (Rest of World): The tariff rate is flat at roughly 2 percent through January 2025, then rises steadily through 2025 to a peak of approximately 17 percent in late 2025, ending near 16 percent.
Figure C.2: Average Tariff Rate and Exposure by Sector
Notes: This figure shows the monthly average tariff rate
(left axes) and exposure (right axes) over the sample period by sector.
For each year-month, we calculate the mean tariff rate and exposure
across all observations. Both are measured in percent.
Seventeen small-multiple panels, one per sector, each showing the monthly average tariff rate (blue line, left axis) and tariff exposure (orange line, right axis) from January 2024 through December 2025. Each panel has its own axis scales. The sectors shown are Apparel, Footwear, and Accessories; Automotive; Baby; Books; Electronics; Entertainment; Grocery; Health & Beauty; Home & Garden; Household; Office; Party & Occasions; Pet; Sports; Tobacco Products and Accessories; Tools & Home Improvement; and Toys.
All sectors share a common pattern: both series are low and flat through January 2025, then rise sharply from early 2025 and plateau or drift higher through the end of the sample.
Sectors with the largest tariff rate increases include Apparel, Footwear, and Accessories (from roughly 10 to about 25 percent), Party & Occasions (from roughly 5 to about 21 percent), Sports and Toys (each rising from low single digits to roughly 20 percent), Home & Garden and Tools & Home Improvement (each rising to roughly 20 percent), and Pet (rising to roughly 20 percent).
For most sectors the tariff rate and tariff exposure lines move nearly on top of one another, indicating high import penetration. Grocery is the clearest exception: its tariff rate rises to roughly 11 percent while its exposure stays near 1 to 2 percent, reflecting a large domestic production share. Health & Beauty and Household show a similar but smaller gap. Books is the one sector where the exposure line sits above the rate line on its own scale.
Figure C.3: Average Tariff Rate and Exposure by Essentiality
Notes: This figure shows the monthly average tariff rate
(left axis) and exposure (right axis) over the sample period by
essentiality. Essential: Grocery, Household, Health & Beauty.
Non-essential: everything else (Automotive, Electronics, Tools &
Home Improvement, Home & Garden, Office, Baby, Pet, Tobacco,
Apparel/Footwear/Accessories, Sports, Toys, Books, Entertainment, Party
& Occasions). For each year-month, we calculate the mean tariff rate
and exposure across all observations. Both are measured in percent.
Two-panel line chart showing the monthly average tariff rate (blue line, left axis) and tariff exposure (orange line, right axis) from January 2024 through December 2025, separately for Essential and Non-essential goods.
Essential panel: The tariff rate is flat near 2.5 percent through January 2025, then rises to about 10 percent by mid-2025 and peaks near 14 percent in late 2025 before easing to roughly 13 percent. Tariff exposure runs well below the rate throughout, rising from under 1 percent to a peak near 3 percent, so a visible wedge opens between the two lines after February 2025.
Non-essential panel: The tariff rate and tariff exposure lines lie almost exactly on top of each other for the entire sample. Both are flat at low levels through January 2025, then rise steeply to roughly 13 percent by mid-2025 and to a peak near 17 percent in late 2025, ending around 15 to 16 percent.
Non-essential goods therefore face substantially higher tariff rates and exposure than essential goods, and for non-essentials the two measures coincide because import penetration is close to complete.
Figure D.1: Leave-one-sector-out estimates of β
(a) Tariff Rate
(b) Tariff Exposure
Notes: Each row reports the benchmark tariff coefficient
\(\beta\) from Equation (4) when the indicated sector is
excluded from the estimation sample. The row labeled (Full) reports the
full-sample estimate, also indicated by the red dashed vertical line in
each panel. Whiskers denote 95% confidence intervals with standard
errors clustered at the HS4 code level. Columns correspond to the three
outcome variables. Panel A uses the realized tariff rate \(\tau_{h(i),t}\) as the treatment variable;
Panel B uses the tariff exposure measure defined in Equation (7).
Two stacked sets of forest plots. Panel A is labeled Tariff Rate and Panel B is labeled Tariff Exposure. Each panel contains three side-by-side columns of estimates, headed Price, Spending, and Quantity. Each row corresponds to the sector dropped from the estimation sample, with the row labeled (Full) at the top, and reports a point estimate with a 95 percent confidence interval. A red dashed vertical line marks the full-sample estimate in every column.
Panel A (Tariff Rate): The full-sample price estimate is approximately 0.14, and the leave-one-out estimates range from about 0.12 to 0.18, with dropping Books producing the largest upward movement and dropping Apparel, Footwear, and Accessories the largest downward movement. The full-sample spending estimate is approximately negative 0.46, with leave-one-out estimates between roughly negative 0.38 and negative 0.55; dropping Home & Garden gives the most negative estimate and dropping Apparel, Footwear, and Accessories the least negative. The full-sample quantity estimate is approximately negative 0.41, with leave-one-out estimates between roughly negative 0.35 and negative 0.48.
Panel B (Tariff Exposure): The full-sample price estimate is approximately 0.20, with leave-one-out estimates from about 0.17 to 0.23. The full-sample spending estimate is approximately negative 0.62, with leave-one-out estimates from roughly negative 0.53 to negative 0.72; dropping Pet gives the most negative estimate and dropping Grocery the least negative. The full-sample quantity estimate is approximately negative 0.55, with leave-one-out estimates from roughly negative 0.45 to negative 0.64, again with Grocery as the largest attenuator.
In both panels every leave-one-out confidence interval overlaps the full-sample estimate, and no single sector’s removal overturns the sign, statistical significance, or order of magnitude of any coefficient.
Figure D.2: Average Spending and Quantity Responses: Stockpiling versus Sustained
Response
Notes: The figure reports average monthly responses of log
real spending and log quantity to tariffs, separately for a three-month
transition window (February–April 2025) and a steady-state window
covering the remainder of the sample period (May–December 2025).
Estimates are constructed as linear combinations of event-study
coefficients from Equations (4) and (5), averaging the
coefficients within each window. Standard errors are computed from the
full variance-covariance matrix of the estimated event-study
coefficients to account for correlation across event-time dummies, and
are clustered at the HS4 code level. The transition window captures any
anticipatory or stockpiling behavior that would unwind shortly after
tariff implementation; the steady-state window captures the sustained
response once such dynamics have dissipated. Horizontal bars report 95%
confidence intervals. Rows correspond to outcomes (spending, quantity);
columns correspond to treatment definitions (tariff rate, tariff
exposure).
Forest plot with two columns (Tariff Rate on the left, Tariff Exposure on the right) and two row blocks of outcomes (Spending on top, Quantity on bottom). Within each block there are two horizontal point-and-interval estimates: Steady state (May-December 2025) in blue on the upper line and Transition (February-April 2025) in orange on the lower line. The horizontal axis is the average monthly response in percentage points, and a vertical line marks zero. Horizontal bars report 95 percent confidence intervals.
Tariff Rate column, Spending: The steady-state average is approximately negative 0.45 with an interval running from about negative 0.75 to negative 0.15, excluding zero. The transition average is approximately 0.05 with a much wider interval spanning roughly negative 0.7 to positive 0.75.
Tariff Rate column, Quantity: The steady-state average is approximately negative 0.40 with an interval that excludes zero. The transition average is approximately 0.1 with a wide interval spanning zero.
Tariff Exposure column, Spending: The steady-state average is approximately negative 0.7 with an interval running from about negative 1.0 to negative 0.4, excluding zero. The transition average is approximately negative 0.3 with a very wide interval spanning roughly negative 2 to positive 1.4.
Tariff Exposure column, Quantity: The steady-state average is approximately negative 0.65 with an interval that excludes zero. The transition average is near zero with a wide interval spanning roughly negative 2 to positive 1.4.
Across all four outcome-by-treatment combinations the transition-window averages are small and indistinguishable from zero while the steady-state averages are large, negative, and precisely estimated.
Figure E.1: HS6 Event Study for the 2018–2019 Tariff Episode
Notes: The figure plots coefficients from event-study
versions of Equations (4) and (5) estimated on the
2018–2019 Numerator panel rebuilt at the HS6 level, interacting the
treatment variable with year-month indicators (the month before onset
omitted as reference). The top row uses the tariff rate; the bottom row
uses the tariff exposure measure. The left column shows the price
response (log Fisher Price Index); the right column shows spending and
quantity (both in logs, with spending in real terms). All specifications
include year-month and panel-unit-by-calendar-month interaction fixed
effects, with panel-specific linear pre-trends in the price equation.
Shaded areas are 95% confidence intervals from standard errors clustered
at the HS6 code level. The vertical dashed line marks September 2018,
the tariff onset.
Four-panel chart arranged in a 2x2 grid, with coefficients in percentage points on the vertical axes and year-months from January 2018 through December 2019 on the horizontal axes. The top row uses the tariff rate specification and the bottom row uses the tariff exposure specification. The left column shows the price response in blue; the right column shows spending in orange and quantity in green. A dark red dashed vertical line marks September 2018. Shaded bands indicate 95 percent confidence intervals.
Top-left (Price, Tariff Rate): Coefficients hover near zero through early 2018, rise to roughly 0.25 in mid-2018, then fall after the September 2018 onset to a trough near negative 0.6 in early 2019. They recover through 2019, turning positive from around July 2019 and reaching roughly 0.25 before ending near negative 0.15 in December 2019.
Top-right (Spending and Quantity, Tariff Rate): Both series start near positive 1 in early 2018, decline to roughly negative 2 by late 2018, and then recover gradually through 2019 toward zero, ending slightly positive. Confidence bands are wide throughout, and spending and quantity move together with no clean break at the September 2018 line.
Bottom-left (Price, Tariff Exposure): Coefficients are near zero in early 2018 with wide bands, drift to roughly negative 0.4 after the onset, reach a trough near negative 0.6 in early 2019, then return toward zero and end slightly negative.
Bottom-right (Spending and Quantity, Tariff Exposure): Both series begin near zero, fall to roughly negative 2.5 around late 2018, then rise steadily through 2019 to fluctuate around zero, ending slightly positive. As in the top row, there is no clean or consistent break at the September 2018 onset.
Figure F.1: Expenditure Share Changes by Tariff Exposure Quartile
Notes: The figure plots mean expenditure share changes (in
percentage points) from the pre- to post-tariff period, by tariff
exposure quartile and income group. Q1 corresponds to HS codes with the
lowest tariff exposure; Q4 corresponds to the highest. Negative values
indicate households shifted spending away from those categories.
Grouped bar chart showing the mean expenditure share change in percentage points from the pre- to post-tariff period, by tariff exposure quartile on the horizontal axis and income group by color (Low in red, Middle in blue, High in green). A dashed horizontal line marks zero.
Q1 (Low exposure): All three groups show share gains. Low is approximately positive 0.006, Middle approximately positive 0.003, and High approximately positive 0.007.
Q2: All three groups again show gains, larger than in Q1 for Middle and High. Low is approximately positive 0.007, Middle approximately positive 0.008, and High approximately positive 0.010.
Q3: All three bars are essentially zero, at roughly positive 0.0004 or less.
Q4 (High exposure): All three groups show share declines. Low is approximately negative 0.013, Middle approximately negative 0.011, and High approximately negative 0.017.
The pattern is consistent with households reallocating spending away from the most tariff-exposed categories, with high-income households showing both the largest gains in the low-exposure quartiles and the largest declines in the highest-exposure quartile.