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 B – Intermediate Synthetic Control Outputs

Synthetic Control Weights

Table B1 and Table B2 below show the country and predictor weights for the different lag specifications using value and quantity as the dependent variable at a quarterly frequency. Results for specifications using the price dependent variable and/or a monthly frequency are available upon request.

Table B1. Optimal country weights with different lag specifications
Country Value Quantity
No Lag First Lag Seas. Lag First & S. Lag No Lag First Lag Seas. Lag First & S. Lag
1 Australia .. 0.164 0.161 0.164 0.172 0.098 0.096 0.099
2 Austria 0.132 .. .. .. .. .. .. ..
3 Belgium .. .. .. .. .. .. .. ..
4 Canada 0.577 0.248 0.276 0.248 0.430 0.560 0.563 0.557
5 Chile .. .. .. 0.009 .. 0.003 0.002 0.003
6 Czechia .. 0.017 .. 0.017 .. .. .. ..
7 Denmark .. .. .. .. .. .. 0.021 ..
8 Finland 0.086 .. .. .. .. .. .. ..
9 France .. 0.228 0.235 0.228 .. 0.033 0.013 0.031
10 Germany .. .. .. .. .. .. .. ..
11 Hungary .. .. .. .. .. .. .. ..
12 Iceland .. 0.015 .. 0.015 .. .. 0.005 ..
13 Ireland .. .. .. .. .. .. .. ..
14 Italy 0.095 .. .. .. .. .. .. ..
15 Japan .. .. .. .. .. 0.063 0.038 0.061
16 Lithuania .. 0.019 .. 0.019 .. 0.153 0.164 0.156
17 Netherlands .. 0.039 0.047 0.039 .. 0.048 0.048 0.048
18 N. Zealand .. 0.016 .. 0.016 .. 0.021 0.025 0.023
19 Norway 0.109 .. .. .. .. .. .. ..
20 Poland .. .. .. .. 0.398 .. .. ..
21 Portugal .. 0.105 0.126 0.105 .. 0.004 0.009 0.004
22 Slovakia .. .. .. .. .. .. .. ..
23 Slovenia .. .. .. .. .. .. .. ..
24 South Africa .. .. .. .. .. .. .. ..
25 South Korea .. .. 0.013 .. .. .. .. ..
26 Spain .. 0.081 0.114 0.081 .. .. .. ..
27 Sweden .. 0.057 0.027 0.057 .. 0.003 .. 0.004
28 Switzerland .. .. .. .. .. 0.013 0.017 0.014

.. : no weight placed (0.000); using the quarterly specifications with no population interpolation. Rounded to three decimal places.

Table B2.Optimal predictor weights with different lag specifications
Dependent Variable Independent
Variables
No Lag First Lag Seas. Lag First & S. Lag
Logged Export Value per Capita First Lagged Dependent Variable .. 1.000 .. 1.000
Seasonal Lagged Dependent Variable .. .. 0.051 ..
Long-Term Interest Rate .. .. .. ..
GBP Exchange Rate 0.032 .. 0.898 ..
Logged Private Final Consumption (per cap) (Local) .. .. .. ..
Logged Private Final Consumption (per cap) (UK) 0.323 .. 0.051 ..
Monthly Dummy 0.323 .. .. ..
Average Alcohol Consumption (per cap) (2010-18) 0.323 .. .. ..
Distance between Capitals .. .. .. ..
Logged Export Quantity per Capita First Lagged Dependent Variable .. 0.212 .. 0.124
Seasonal Lagged Dependent Variable .. .. 0.078 0.033
Long-Term Interest Rate 0.027 .. 0.001 ..
GBP Exchange Rate 0.321 .. .. ..
Logged Private Final Consumption (per cap) (Local) 0.006 .. .. ..
Logged Private Final Consumption (per cap) (UK) 0.321 .. 0.078 0.124
Monthly Dummy 0.321 .. .. ..
Average Alcohol Consumption (per cap) (2010-18) .. 0.787 0.843 0.719
Distance between Capitals 0.005 .. .. ..

.. : no or little weight placed (0.000 or < 0.0005); Using the quarterly specifications with no population interpolation. Rounded to three decimal places.

Post- to Pre-Treatment RMSPE Ratios

The following tables show ratios of the post-tariff fit to the pre-tariff fit (for the US). A ratio of > 1 would show that our synthetic control has a worse fit after the tariff, while a ratio of < 1 would show the opposite. These ratios themselves do not have much meaning without also looking at the ratios obtained from countries unaffected by the tariff (the placebo studies). Comparing these ratios with those obtained by the US would provide us with a p-value.

Table B3. Ratio of post- to pre-treatment RMSPEs with different monthly lag specs.
Dependent Variable Period* No Lag First Lag Seas. Lag First & S. Lag
Logged Export Value per Capita US Pre-Tariff 0.23836 0.19924 0.20298 0.19917
US Post-Tariff 0.26765 0.30455 0.30911 0.30151
US Ratio** 1.12291 1.52859 1.52289 1.51386
Countries with Ratio ≥ US 15/29 6/29 5/29 6/29
Logged Export Quantity per Capita US Pre-Tariff 0.22948 0.16004 0.16304 0.16004
US Post-Tariff 0.35667 0.20361 0.18330 0.20361
US Ratio* 1.55429 1.27221 1.12426 1.27221
Countries with Ratio ≥ US 3/29 13/29 19/29 13/29
Logged Export Price US Pre-Tariff 0.14120 0.12383 0.12499 0.12383
US Post-Tariff 0.13269 0.14933 0.15302 0.14933
US Ratio* 0.93975 1.20595 1.22430 1.20595
Countries with Ratio ≥ US 15/29 12/29 12/29 12/29

Using the monthly specifications with no interpolation or disaggregation. Rounded to five decimal places. * Pre-tariff is Jan 2010 to Oct 2019, post-tariff is Nov 2019 to Dec 2020; ** Larger ratios show larger differences between post- and pre-tariff fit, see equation (4).

Table B4. Ratio of post- to pre-treatment RMSPEs with different quarterly lag specs.
Dependent Variable Period* No Lag First Lag Seas. Lag First & S. Lag
Logged Export Value per Capita US Pre-Tariff 0.17703 0.11943 0.12009 0.11943
US Post-Tariff 0.21071 0.28206 0.28489 0.28206
US Ratio** 1.19022 2.36168 2.37217 2.36168
Countries with Ratio ≥ US 13/29 2/29 2/29 2/29
Logged Export Quantity per Capita US Pre-Tariff 0.14457 0.06657 0.06709 0.06656
US Post-Tariff 0.31616 0.17758 0.17290 0.17787
US Ratio* 2.18689 2.66753 2.57719 2.67245
Countries with Ratio ≥ US 2/29 2/29 2/29 2/29
Logged Export Price US Pre-Tariff 0.10643 0.09301 0.09453 0.09301
US Post-Tariff 0.10326 0.12301 0.10881 0.12301
US Ratio* 0.97021 1.32256 1.15109 1.32256
Countries with Ratio ≥ US 16/29 10/29 14/29 10/29

Using the quarterly specifications with no interpolation. Rounded to five decimal places. * Pre-tariff is Q1 2010 to Q3 2019, post-tariff is Q4 2019 to Q4 2020; ** Larger ratios show larger differences between post- and pre-tariff fit, see equation (4).

Synthetic Control Predictor Values

The table below shows the average value of each variable for the particular weighted average of countries (synthetic control) for the pre-tariff period. Some of these values may be heavily skewed (for example, in the case of the exchange rate) due to certain countries being included in the weighted average. This is not necessarily a concern – only the difference in export value, quantity, or price is used in estimating the impact of tariff.

Table B5. Average synthetic control predictor values for the pre-tariff period

  Export Value, Quantity, or Price Local Private Final Consumption UK Private Final Consumption Population (b) Interest Rate GBP Exchange Rate Avg. Alcohol Consumption 2010-18 Distance Between Capitals
See (a) for unit £000's per capita £000's per capita Millions % Local currency / £ Litres per cap km
Monthly average for the Jan 2010 – Oct 2019 pre-tariff period
US
Value 0.06 6.4 4.7 319.4 2.42 1.48 8.8 5.9
Quantity 0.001 6.4 4.7 319.4 2.42 1.48 8.8 5.9
Price 39.02 6.4 4.7 319.4 2.42 1.48 8.8 5.9
Synthetic control predictor averages
Value 0.06 3.8 4.7 .. 2.27 46.10 9.6 5.1
Quantity 0.001 4.1 4.7 .. 2.10 137.34 8.9 6.7
Price 38.53 4.5 4.7 .. 1.59 10.94 9.2 3.0
Synthetic control weights (excl. lagged dep. vars. and/or monthly dummy)
Value .. 0.000 0.000 .. 0.003 0.001 0.000 0.000
Quantity .. 0.000 0.000 .. 0.001 0.000 0.000 0.000
Price .. 0.000 0.330 .. 0.000 0.011 0.000 0.000
Quarterly average for the Q1 2010 – Q3 2019 pre-tariff period
US
Value 0.17 6.4 4.7 319.3 2.42 1.48 8.8 5.9
Quantity 0.004 6.4 4.7 319.3 2.42 1.48 8.8 5.9
Price 39.17 6.4 4.7 319.3 2.42 1.48 8.8 5.9
Synthetic control predictor averages
Value 0.17 3.5 4.7 .. 2.53 45.57 9.9 4.9
Quantity 0.004 3.9 4.7 .. 2.24 267.11 8.8 6.8
Price 38.69 4.7 4.7 .. 2.18 21.99 9.8 5
Synthetic control weights (excl. lagged dep. vars. and/or monthly dummy)
Value .. 0.000 0.000 .. 0.000 0.000 0.000 0.000
Quantity .. 0.000 0.000 .. 0.000 0.000 0.787 0.000
Price .. 0.004 0.000 .. 0.000 0.033 0.000 0.000

(a) Export value is measured in £ per capita, export quantity is measured in litres of pure alcohol per capita, price is measured in £/LPA. (b) Population is not a predictor in itself – instead, it was used to convert export value or quantity and private final consumption into per-capita values. These variables were subsequently logged.

Contact

Email: agric.stats@gov.scot

Back to top