Estimating the impacts of US tariffs on UK exports of single malt Scotch whisky

This discussion paper aims to estimate the impact of US tariffs on UK exports of single malt Scotch whisky between Q4 2019 – Q4 2020 using the novel synthetic control method.


Annex D – Final Results

Monthly Specifications

The table below shows results for all tested specifications without population interpolation or consumption disaggregation. Generally, results were similar whether this was done or not.[58]

Table D1. Final set of results for monthly specifications
Dependent
variable*
Lag Estimated monthly
tariff effect
p-value 26/29 confidence set
(~90%)
Unit(a) Prop. Lower Upper
% % %
Value No lag -0.01 -7.1 0.517 -87.9 +19.4
Value First lag -0.02 -17.4 0.207 -87.0 +1.2
Value Seasonal lag -0.02 -18.1 0.172 -88.8 +1.4
Value First and seas. lag -0.02 -17.2 0.207 -86.4 +1.5
Quantity No lag -0.40* -16.5 0.103 -62.5 -0.8
Quantity First lag -0.24 -10.7 0.448 -34.9 +16.9
Quantity Seasonal lag -0.21 -11.4 0.655 -32.5 +26.6
Quantity First and seas. lag -0.24 -12.9 0.448 -34.8 +17.5
Price No lag -2.98 -6.2 0.517 -38.0 +29.6
Price First lag -5.72 -11.9 0.414 -56.6 +3.8
Price Seasonal lag -5.88 -12.2 0.414 -57.5 +3.8
Price First and seas. lag -5.72 -11.9 0.414 -56.7 +4.0

Without population interpolation or consumption disaggregation (results for these specifications were similar if not identical). Key: (a) Logged value (£/cap), quantity (millilitres of pure alcohol / cap) and price (£/LPA); *** p < 2/29; ** p < 3/29; * p < 4/29.

Quarterly Specifications

The same synthetic control method was applied on a quarterly trade series, with the tariff assumed to be introduced in Q4 2019. Using a quarterly series may limit the effects of seasonality and would more closely align with the quarterly consumption series. The monthly interest rate and exchange rates were averaged over the quarter, while trade value/quantity was summed. Quarterly population interpolation was also explored but not found to have a major impact.

While the quarterly series may limit seasonality, it would also reduce the power of any hypothesis test by reducing the number of observations. The total number of observations for each country decreases from 132 months to 44 quarters, with only 5 post-tariff observations (2019 Q4 to 2020 Q4).

Although the estimated quarterly effect of the tariff on the export value and quantity appears to be strictly negative at a 10% significance level, the lower bound for the value confidence set is still somewhat non-informative at [-100.0%, -4.7%].

Table D2. Final set of results for quarterly specifications

Dependent variable Lag Estimated quarterly
tariff effect
p-value 26/29 confidence set (~90%)
Unit(a) Prop. Lower Upper
% % %
Value No lag -0.03 -8.5 0.448 -100.0 (a)
Value First lag -0.06** -18.3 0.069 -100.0 -4.7
Value Seasonal lag -0.06** -18.9 0.069 -100.0 -5.2
Value First and seas. lag -0.06** -18.3 0.069 -100.0 -4.7
Quantity No lag -1.38** -20.2 0.069 -69.4 -4.6
Quantity First lag -0.64** -10.3 0.069 -19.6 -9.5
Quantity Seasonal lag -0.57** -9.0 0.069 (b) (b)
Quantity First and seas. lag -0.64*** -10.2 0.034 -19.2 -9.8
Price No lag -3.89 -8.3 0.552 -49.7 +32.8
Price First lag -4.86 -10.2 0.345 -64.7 +17.9
Price Seasonal lag -4.27 -9.1 0.483 -62.5 +25.5
Price First and seas. lag -4.86 -10.2 0.345 -64.7 +17.9

Without population interpolation (results for these specifications were similar if not identical). Key: (a) Logged value (£/cap), quantity (millilitres of pure alcohol / cap) and price (£/LPA); (b) no lower/upper bound was found due to lower power; *** p < 2/29; ** p < 3/29; * p < 4/29.

Table D3. Comparison between different first-lag frequency specifications
Frequency Dependent
variable
Estimated monthly or quarterly tariff effect p-value 26/29 confidence set
(~90%)
Unit(a) Prop. % Lower % Upper %
Quarterly Value -0.06** -18.3 0.069 -100.0 -4.7
Monthly Value -0.02 -17.4 0.207 -87.0 +1.2
Quarterly Quantity -0.64** -10.3 0.069 -19.6 -9.5
Monthly Quantity -0.24 -10.7 0.448 -34.9 +16.9
Quarterly Price -4.86 -10.2 0.345 -64.7 +17.9
Monthly Price -5.72 -11.9 0.414 -56.6 +3.8

Without population interpolation or consumption disaggregation. Key: (a) Logged value (£/cap), quantity (millilitres of pure alcohol / cap) and price (£/LPA); ** p < 3/29.

Contact

Email: agric.stats@gov.scot

Back to top