## _spn1007

## Source details

**Canonical URL:** [_spn1007](https://www.imf.org/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/spn/2010/_spn1007.pdf)

## Other formats

- [Markdown version](/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/spn/2010/_spn1007.pdf.md)
- [Structured JSON version](/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/spn/2010/_spn1007.pdf.json)

---

### Executive Summary — I. Introduction and the “Great Trade Collapse”
- Pace of trade reforms waned from the mid-2000s; with the global financial crisis, reform progress halted and began to reverse.
- Global trade volumes fell 17.5 percent between September 2008 and January 2009 (the “Great Trade Collapse”).
- Principal explanations for the collapse:
  - Compositional effects: demand for investment goods and consumer durables—“postponable” goods—fell most sharply.
  - Global supply chains and “just-in-time” production propagated demand shocks.
  - Reduced availability and higher cost of trade finance and credit, particularly affecting emerging market economies and smaller firms.
- Trade dynamics during the crisis:
  - Trade rose 25 percent between May and December 2009, returning to some 90 percent of September 2008 levels.
- Policy-relevant finding:
  - New restrictive measures, where implemented, had a strong negative impact on trade in targeted product categories.

### Executive Summary — II. Slowing Momentum of Trade Liberalization Holds Risks
- Long-term liberalization trends and remaining vulnerabilities:
  - Among major western European and North American countries, average tariffs fell from 15 to 4 percent during 1952–2005.
  - Brazil’s average tariff declined from 51 percent in 1987 to 12 percent today.
  - India’s average tariff fell from 71 percent in 1994 to 32 percent in 2002 and 13 percent today.
  - Progress includes removal of many non-tariff barriers (NTBs) in textiles, clothing, and footwear and reductions in agricultural export subsidies and NTBs, though tariffs remain relatively high in those areas.
- Scope for reversal under existing WTO bindings:
  - Many countries have average MFN tariffs well below their WTO-bound ceilings, creating room to increase tariffs without violating WTO commitments.
  - Examples of gaps between average applied tariffs and WTO-bound rates: Brazil 18 percentage points, Argentina 20 p.p., India 36 p.p.
- Measured restrictiveness ahead of the crisis:
  - Overall import restrictiveness (World Bank OTRI) was about 5 percent for advanced economies and about 15 percent for non-advanced, non-emerging market economies.
  - Exports from developing economies faced slightly higher restrictions in their export markets.

### Executive Summary — III. Awareness, Monitoring, and Factors Limiting Protectionism So Far
- Aggregate impact of new trade restrictions to date:
  - The aggregate impact of new restrictions is modest—about 0.25 percent of global trade—because most countries resisted widespread protectionism.
- High-level political awareness and commitments:
  - G-20 leaders pledged in November 2008 to “refrain from raising new barriers to investment or to trade in goods and services, imposing new export restrictions, or implementing WTO inconsistent measures to stimulate exports.”
  - In April 2009, G-20 leaders extended this pledge through 2010 and asked the WTO and other institutions to monitor adherence.
- Intensified monitoring efforts:
  - WTO monitoring reports cover trade measures for all countries.
  - Joint reports by WTO, OECD, and UNCTAD cover G-20 trade and investment measures in response to the G-20 request.
  - Additional monitoring, such as Global Trade Alert (GTA), supplements official monitoring.
- Factors that have limited resort to protectionism:
  - Multilateral rules and institutions (WTO) clarified responsible policy actions; very little resort to new WTO-inconsistent measures has been observed.
  - Trade fell more rapidly than overall economic activity, reducing import-to-GDP ratios; job losses were not widely blamed on trade.
  - Supportive macroeconomic and financial policies: highly expansionary monetary policy and expanded central bank balance sheets in key advanced economies, and additional fiscal stimulus in many countries. Coordinated fiscal stimulus helped sustain demand and, according to McKibbin and Stoeckel (2009), reduces protectionist sentiment more than unilateral packages.

### Risk of rising protectionism: context and channels (section 11)
- Contextual factors increasing protectionist risk:
  - Unemployment has risen and may remain high; more countries face limited fiscal space—this helps explain increased industry requests for trade remedy measures.
  - Unconventional and “behind-the-border” measures (“murky” protectionism) may already have had a considerable impact; GTA analysis has focused attention on their extent.
  - Extensive supply chains have increased reliance on trade and contributed to a “pro-trade” sentiment.
  - Some 99 percent of tariffs today are calculated as ad valorem rates, unlike the specific tariffs widely used in the 1930s.
  - Large fiscal stimulus packages spill over to imports; perceived “free riding” may provoke discriminatory trade measures (procurement or other).
  - Global coordination of proportionally similar stimulus packages helped reduce the risk of discriminatory responses.

### New trade restrictions and measured impacts
- Monitoring conclusions:
  - Monitoring activities differ on extent of protectionism; none suggests an extreme surge comparable to the 1930s.
  - GTA’s 4th Report argues reporting lags understate anti-foreigner discrimination and found new measures in Q4 2009 at rates comparable to the height of the crisis earlier in 2009.
- Empirical findings:
  - Newly implemented trade restrictions have already had a strong negative impact, though they cover only a small share of trade.
  - Trade analysis used monthly bilateral trade values at the 4-digit (HS) product level, with reporter coverage beginning in July 2008 and ending, for most reporters, with November 2009; import and export reporters include the EU and 14 other major G-20 traders, covering some 80 percent of global merchandise trade.
  - 184 measures identified by GTA as highly likely to be discriminatory (the “red measures”) were incorporated into the analysis.
  - Raw data comparisons (e.g., measures implemented Nov. 2008) show imports targeted by new import-restrictive measures declined more than world trade in the same products.
  - Export subsidies seemingly increased targeted exports relative to world trade of the same products; export restrictions as defined included measures that both reduce and intensify export activity.

### Econometric quantification of trade distortions (Henn and McDonald (2010))
- Data and controls:
  - Analysis examined 4-digit bilateral monthly trade flows and controlled for: (i) product-specific demand shocks from the crisis; (ii) country-specific income declines as the crisis progressed; and (iii) bilateral exchange rates, inflation differentials, and transport costs.
- Basic and refined estimates:
  - Basic specification estimate: a new restriction is associated with about an 8.5 percent distortion to trade.
  - Refined specification (after accounting for other trade determinants): a new restriction is responsible for a 3 percent distortion to trade.
    - Note: these effects apply to entire 4-digit product categories—most measures cover only a portion of those categories—implying potentially larger impacts on the specifically affected products.
- Aggregate estimates:
  - Measures distorted aggregate world trade by about 0.25 percent.
  - Estimates across specifications imply measures distorted aggregate global trade by between 0.2 and 0.7 percent.
  - Results are based only on the 184 measures for which GTA provided necessary product and country coverage information.

### ANNEX I — Econometric specification (summary)
- Estimation equation:
  - ln(Imports_ijpt)Δ = TVFE + ProtDummy_ijptΔ + ε_ijpt
  - ln(Imports)Δ is the 12-month change in the log of imports.
  - TVFE = one or more sets of time-varying fixed effects.
  - ProtDummyΔ = change in the 0–1 indicator variable for observations subject to a protectionist measure (dummy = –1 for discriminatory measures that support trade, such as export subsidies).
  - Indices: i = exporter, j = importer, p = 4-digit product, t = time.
- Fixed-effect specifications:
  - Specification 1: TV product fixed effects.
  - Specification 2: TV product & TV importer fixed effects.
  - Specification 3: TV product & TV country-pair fixed effects.

### ANNEX I — Main econometric results (aggregate protectionist dummy)
- Baseline results (all regressions have 8,806,027 observations and F-Statistic rejection Prob. 0.000):
  - Regression 1a (TV Product; Number of FE 20,569):
    - Protectionist Dummy coefficient: –0.089*** (T-statistic: –17.2)
    - R-squared (percent): 1.5
  - Regression 2a (TV Product & TV Importer; Number of FE 24,436):
    - Protectionist Dummy coefficient: –0.033*** (T-statistic: –4.6)
    - R-squared (percent): 2.0
  - Regression 3a (TV Product & TV Cty-pair; Number of FE 96,082):
    - Protectionist Dummy coefficient: –0.032*** (T-statistic: –4.8)
    - R-squared (percent): 3.3
- Interpretation:
  - Specification 1 estimated coefficient of 0.089 implies about an 8.5 percent distortion to trade from a new restriction.
  - Specifications 2 and 3 imply new measures distorted trade in targeted 4-digit product categories by about 3 percent.
  - Implementation of new measures is found to discriminate against targeted trade flows with statistical significance at the 0.1 percent level.

### ANNEX I — Disaggregated results by type of measure
- Table 2 results (all regressions have 8,806,027 observations and F-Statistic rejection Prob. 0.000):
  - Regression 1b (TV Product; Number of FE 20,569):
    - Import Restrictions: –0.127*** (T-statistic: –18.5)
    - Export Support: 0.096*** (T-statistic: 6.2)
    - Export Restrictions: –0.018* (T-statistic: –2.0)
    - R-squared (percent): 1.5
  - Regression 2b (TV Product & TV Importer; Number of FE 24,436):
    - Import Restrictions: –0.026*** (T-statistic: –2.8)
    - Export Support: 0.092*** (T-statistic: 4.0)
    - Export Restrictions: –0.025* (T-statistic: –2.0)
    - R-squared (percent): 2.0
  - Regression 3b (TV Product & TV Cty-pair; Number of FE 96,082):
    - Import Restrictions: –0.028*** (T-statistic: –3.2)
    - Export Support: 0.004 (T-statistic: 0.2)
    - Export Restrictions: –0.066*** (T-statistic: –5.1)
    - R-squared (percent): 3.3
- Interpretation (Specification 2 emphasized in text):
  - Import restrictions on average lead to a 2.5 percent import decline (Specification 2 coefficient: –0.026***).
  - Export support measures increase exports by 9.5 percent (Specification 1 coefficient: 0.096***; Specification 2: 0.092***).
  - Export restrictions are associated with a 2.5 percent decrease in exports (Specification 2 coefficient: –0.025*).

### ANNEX I — Aggregate impact on world trade (calculation and estimates)
- Specification 2 implies measures distorted aggregate global trade by 0.22 percent.
  - Specification 2 breakdown:
    - Restrictions lowering world trade by 0.18 percent.
    - Export support measures increasing world exports by 0.04 percent.
- Specification 3 yields a similar aggregate distortion of 0.28 percent.
- Specification 1 yields an aggregate distortion of 0.71 percent.
- Footnote calculation (as provided): The calculation is 0.05 (exp( 0.026) − 1) − 0.004 (exp(0.092) − 1) (0.022 (exp(0.025) − 1) = 0.22%.

### Why protectionist pressures may intensify after the crisis
- Key channels for intensification:
  - Persistent high unemployment; employment recovery may lag growth.
  - High unemployment combined with rising import growth and higher import market shares increases filing rates for anti-dumping petitions (evidence of higher filing rates in H2 2009).
  - Narrowing macroeconomic policy space and eroding political consensus for stimulus may lead to premature withdrawal of support that helped dampen protectionist pressures.
  - Higher commodity prices risk some countries taxing or restricting commodity exports.
  - Surges in capital inflows and rapid currency appreciation in some emerging markets may pressure export- and import-competing industries toward trade actions.
  - Widening global imbalances may trigger protectionism; past work suggests protectionist surges reduce global growth while leaving imbalances unresolved.
  - Retaliation risk: succumbing to protectionist pressures could provoke widespread retaliation.
- Historical and econometric evidence:
  - Trade protection demand depends positively on real exchange rate appreciation and unemployment (or negatively on production).
  - Example: a one percentage point increase in the U.S. unemployment rate (1980–1995 study) leads to an expected 54 additional anti-dumping or countervailing duty petitions.
  - Unemployment rates for advanced countries are projected to increase through 2010; protectionist demands typically subside only when unemployment begins to decline.

### How to avoid a return to widespread protectionism — policy recommendations
- Maintain sustained vigilance and enhanced monitoring of protectionist measures to prevent broader protectionism.
- Preserve high-level political awareness of macroeconomic risks associated with protectionism.
- Tighten multilateral trade commitments as the most reliable way to avoid a return to widespread protectionism:
  - Conclude the WTO Doha Round to tighten commitments; Doha is identified as a key way to tighten commitments and can be viewed as an integral part of the exit strategy from the global economic crisis.
  - Doha would bring actual policy changes across many areas and enhanced predictability and security of trade policies.
  - Binding gaps leave scope to reverse past liberalization; industrial product tariffs of several key emerging market countries could be raised on average by some 10 to 30 percentage points and yet remain within WTO commitments.
  - In agriculture, present commitments allow countries such as the United States to sharply raise domestic subsidies, distorting global agricultural commodity prices.
  - A Doha agreement could tighten or clarify commitments and rules across anti-dumping, food aid, services, subsidies, and trade facilitation.
- Senior macroeconomic policymakers should treat concluding Doha as part of their exit policy from the global financial crisis:
  - New restrictions have had strong negative effects on the covered products.
  - With unemployment remaining high and imports rebounding, and as monetary and fiscal stimulus are withdrawn, protectionist pressure may intensify.
  - Widespread succumbing to these pressures would jeopardize the pace, strength, and durability of economic recovery.

*Source: Executive Summary, IMF staff paper and ANNEX I in _spn1007 (sections I–III, 11, and ANNEX I as provided).*

### Executive Summary ......................................................................................................

### Executive Summary

### I. Introduction and the “Great Trade Collapse”

- The pace of trade reforms waned from the mid-2000s and, with the global financial crisis, reform progress halted and began to reverse.
- Global trade volumes fell 17.5 percent between September 2008 and January 2009 in the episode termed the “Great Trade Collapse.”
- The Great Trade Collapse is largely explained by:
  - Compositional effects (demand for investment goods and consumer durables—“postponable” goods—fell most sharply).
  - Global supply chains and “just-in-time” production propagating demand shocks.
  - Reduced availability and higher cost of trade finance and credit, which particularly affected emerging market economies and smaller firms.
- Trade dynamics during the crisis:
  - Trade rose 25 percent between May and December 2009, returning to some 90 percent of September 2008 levels.
- Policy-relevant finding:
  - Where new restrictive measures were put in place, they had a strong negative impact on trade in the targeted product categories.

### II. Slowing Momentum of Trade Liberalization Holds Risks

- Long-term liberalization trends and remaining vulnerabilities:
  - Among major western European and North American countries, average tariffs fell from 15 to 4 percent during 1952–2005.
  - Brazil’s average tariff declined from 51 percent in 1987 to 12 percent today.
  - India’s average tariff fell from 71 percent in 1994 to 32 percent in 2002 and 13 percent today.
  - Progress includes removal of many non-tariff barriers (NTBs) in sectors such as textiles, clothing, and footwear and reductions in agricultural export subsidies and NTBs, though tariffs remain relatively high in those areas.
- Scope for reversal under existing WTO bindings:
  - Many countries have average MFN tariffs well below their WTO-bound ceilings, creating room to increase tariffs without violating WTO commitments.
  - Examples of gaps between average applied tariffs and WTO-bound rates: Brazil 18 percentage points, Argentina 20 p.p., India 36 p.p.
- Measured restrictiveness ahead of the crisis:
  - Overall import restrictiveness (World Bank OTRI) was about 5 percent for advanced economies and about 15 percent for non-advanced, non-emerging market economies.
  - Exports from developing economies faced slightly higher restrictions in their export markets.

### III. Awareness, Monitoring, and Factors Limiting Protectionism So Far

- Aggregate impact of new trade restrictions to date:
  - The aggregate impact of new restrictions is modest—about 0.25 percent of global trade—because most countries resisted widespread protectionism.
- High-level political awareness and commitments:
  - G-20 leaders pledged in November 2008 to “refrain from raising new barriers to investment or to trade in goods and services, imposing new export restrictions, or implementing WTO inconsistent measures to stimulate exports.”
  - In April 2009, G-20 leaders extended this pledge through 2010 and asked the WTO and other institutions to monitor adherence.
- Intensified monitoring efforts:
  - WTO monitoring reports cover trade measures for all countries.
  - Joint reports by WTO, OECD, and UNCTAD cover G-20 trade and investment measures in response to the G-20 request.
  - Additional monitoring, such as Global Trade Alert (GTA), supplements official monitoring.
- Factors that have limited resort to protectionism:
  - a. Multilateral rules and institutions (WTO) clarified responsible policy actions; very little resort to new WTO-inconsistent measures has been observed.
  - b. Trade fell more rapidly than overall economic activity, reducing import-to-GDP ratios; job losses were not widely blamed on trade.
  - c. Supportive macroeconomic and financial policies, including highly expansionary monetary policy and expanded central bank balance sheets in key advanced economies, and additional fiscal stimulus in many countries. Coordinated fiscal stimulus helped sustain demand and, according to McKibbin and Stoeckel (2009), reduces protectionist sentiment more than unilateral packages.

### Key policy implications and recommended priorities (from the Executive Summary)

- Sustained vigilance and enhanced monitoring of protectionist measures are important to prevent a broader return to protectionism.
- Maintaining high-level political awareness of the macroeconomic risks associated with protectionism is critical.
- Tightening multilateral trade commitments is the most reliable way to avoid a return to widespread protectionism:
  - Completing the WTO Doha Round is identified as a key way to tighten commitments and can be viewed as an integral part of the exit strategy from the global economic crisis.

*Source: Executive Summary, IMF staff paper (sections I–III as provided).*

### 11.      There is a risk now that protectionism may increase. Unemployment has risen and

### _spn1007 - 11.      There is a risk now that protectionism may increase. Unemployment has risen and

### Risk of rising protectionism: context and channels
- Unemployment has risen and may remain high, and more countries now face limited fiscal space. This may help to explain the increased frequency of industry’s requests for trade remedy measures.
- Unconventional and “behind-the-border” measures (termed “murky” protectionism) may already have had a considerable impact; analysis by GTA has focused attention on their extent.
- From a political economy perspective, extensive supply chains have increased reliance on trade and contributed to a “pro-trade” sentiment.
- Some 99 percent of tariffs today are calculated as ad valorem rates, unlike the specific tariffs widely used in the 1930s.
- Large fiscal stimulus packages spill over to imports; a perception that others will “free ride” may provoke discriminatory trade measures (procurement or other).
- Global coordination of proportionally similar stimulus packages in response to the crisis helped to reduce the risk of discriminatory responses.

### New trade restrictions and measured impacts
- Monitoring activities have come to different conclusions about the extent of protectionism; none suggests an extreme surge comparable to the 1930s.
- GTA’s 4th Report argues that reporting lags understate anti-foreigner discrimination and found new measures in Q4 2009 at rates comparable to the height of the crisis earlier in 2009.
- Our analysis finds that newly implemented trade restrictions have already had a strong negative impact, though they have covered only a small share of trade.
- The trade analysis used monthly bilateral trade values at the 4-digit (HS) product level, with reporter coverage beginning in July 2008 and ending, for most reporters, with November 2009; import and export reporters include the EU and 14 other major G-20 traders, covering some 80 percent of global merchandise trade.
- 184 measures identified by GTA as highly likely to be discriminatory (the “red measures”) were incorporated into the analysis.
- The raw data comparisons (e.g., measures implemented Nov. 2008) show that imports targeted by new import-restrictive measures declined more than world trade in the same products.
- Export subsidies seemingly increased targeted exports relative to world trade of the same products; export restrictions as defined included measures that both reduce and intensify export activity.

### Econometric quantification of trade distortions
- Econometric analysis (Henn and McDonald, 2010) examined 4-digit bilateral monthly trade flows and controlled for: (i) product-specific demand shocks from the crisis; (ii) country-specific income declines as the crisis progressed; and (iii) bilateral exchange rates, inflation differentials, and transport costs.
- Basic specification statistical estimate: a new restriction is associated with about an 8.5 percent distortion to trade.
- Refined specification (after accounting for other trade determinants): a new restriction is responsible for a 3 percent distortion to trade.
  - Note: these effects apply to entire 4-digit product categories—most measures cover only a portion of those categories—implying potentially larger impacts on the specifically affected products.
- Aggregating product-level estimates across trade subject to measures yields:
  - Measures distorted aggregate world trade by about 0.25 percent.
  - Estimates across specifications imply measures distorted aggregate global trade by between 0.2 and 0.7 percent.
- Results are based only on the 184 measures for which GTA provided necessary product and country coverage information.

### Why protectionist pressures may intensify after the crisis
- Key reasons protectionist pressures may intensify:
  - Unemployment is likely to remain high for some time and is associated with protectionist pressures; employment recovery may lag growth.
  - High unemployment combined with rising import growth and higher import market shares increases the success rate and appeal of anti-dumping petitions (evidence of higher filing rates in H2 2009).
  - Narrowing macroeconomic policy space and eroding political consensus for stimulus may lead to premature withdrawal of support that helped dampen protectionist pressures.
- Specific country developments that could generate additional demands:
  - Higher commodity prices risk some countries taxing or restricting commodity exports (as in the 2007–08 food price crisis).
  - Surges in capital inflows and rapid currency appreciation in some emerging markets may pressure export- and import-competing industries toward trade actions.
  - Widening global imbalances may trigger protectionism as countries seek to remedy protracted current account deficits; past work suggests protectionist surges reduce global growth while leaving imbalances unresolved.
  - Retaliation risk: giving in to protectionist pressures could provoke widespread retaliation if measures become common.
- Historical and econometric evidence:
  - Econometric studies show trade protection demand depends positively on real exchange rate appreciation and unemployment (or negatively on production).
  - Example: a one percentage point increase in the U.S. unemployment rate (1980–1995 study) leads to an expected 54 additional anti-dumping or countervailing duty petitions.
  - Unemployment rates for advanced countries are projected to increase through 2010; protectionist demands typically subside only when unemployment begins to decline.

### How to avoid a return to widespread protectionism — policy recommendations
- Leaders and senior policymakers should remain on heightened alert: continue and further enhance monitoring of protectionist measures and maintain high-level political awareness of macroeconomic risks of protectionism.
- Conclude the WTO Doha Round as the surest way to avoid a widespread resort to protectionism with adverse macroeconomic consequences:
  - Doha would bring actual policy changes across many areas and enhanced predictability and security of trade policies.
  - Binding gaps leave scope to reverse past liberalization; industrial product tariffs of several key emerging market countries could be raised on average by some 10 to 30 percentage points and yet remain within WTO commitments.
  - In agriculture, present commitments allow countries such as the United States to sharply raise domestic subsidies, distorting global agricultural commodity prices.
  - A Doha agreement could tighten or clarify commitments and rules across anti-dumping, food aid, services, subsidies, and trade facilitation.
- Senior macroeconomic policymakers should treat concluding Doha as part of their exit policy from the global financial crisis:
  - New restrictions have had strong negative effects on the covered products.
  - With unemployment remaining high and imports rebounding, and as monetary and fiscal stimulus are withdrawn, protectionist pressure may intensify.
  - Widespread succumbing to these pressures would jeopardize the pace, strength, and durability of economic recovery.

*Source: IMF staff analysis in _spn1007 (chapter/section 11–VI) — text and figures as provided in the supplied content.*

### ANNEX I. ECONOMETRIC SPECIFICATION AND RESULTS

### ANNEX I. ECONOMETRIC SPECIFICATION AND RESULTS

### Econometric specification
- Estimation equation (Henn and McDonald (2010)):
  - ln(Imports_ijpt)Δ = TVFE + ProtDummy_ijptΔ + ε_ijpt
  - ln(Imports)Δ is the 12-month change in the log of imports.
  - TVFE stands for one or more sets of time-varying fixed effects.
  - ProtDummyΔ is the change in the 0–1 indicator variable for observations subject to a protectionist measure.
  - The dummy takes the value of –1 for discriminatory measures that support trade, such as export subsidies.
  - Indices: i = exporter, j = importer, p = 4-digit product, t = time.
- Fixed-effect specifications:
  - Specification 1: TV product fixed effects (accounts for product-specific seasonality and worldwide price and demand shocks).
  - Specification 2: TV product & TV importer fixed effects (adds importer income and demand variability).
  - Specification 3: TV product & TV country-pair fixed effects (interaction of TV importer with TV exporter effects; also accounts for exporter-specific supply shocks and bilateral determinants such as transport connections or the exchange rate).

### Main econometric results (aggregate protectionist dummy)
- Table 1 — Baseline results (all regressions have 8,806,027 observations and F-Statistic rejection Prob. 0.000):
  - Regression 1a (TV Product; Number of FE 20,569):
    - Protectionist Dummy coefficient: –0.089*** (T-statistic: –17.2)
    - R-squared (percent): 1.5
  - Regression 2a (TV Product & TV Importer; Number of FE 24,436):
    - Protectionist Dummy coefficient: –0.033*** (T-statistic: –4.6)
    - R-squared (percent): 2.0
  - Regression 3a (TV Product & TV Cty-pair; Number of FE 96,082):
    - Protectionist Dummy coefficient: –0.032*** (T-statistic: –4.8)
    - R-squared (percent): 3.3
- Interpretation:
  - Under Specification 1, the estimated coefficient of 0.089 implies that a new trade restriction is associated with about an 8.5 percent distortion to trade.
  - Specifications 2 and 3, which include additional time-varying importer or country-pair fixed effects, imply that new measures distorted trade in the targeted 4-digit product categories by about 3 percent.
  - The implementation of new measures is found to discriminate against targeted trade flows with statistical significance at the 0.1 percent level.

### Disaggregated results by type of measure
- Table 2 — Results by type of measure (all regressions have 8,806,027 observations and F-Statistic rejection Prob. 0.000):
  - Regression 1b (TV Product; Number of FE 20,569):
    - Import Restrictions: –0.127*** (T-statistic: –18.5)
    - Export Support: 0.096*** (T-statistic: 6.2)
    - Export Restrictions: –0.018* (T-statistic: –2.0)
    - R-squared (percent): 1.5
  - Regression 2b (TV Product & TV Importer; Number of FE 24,436):
    - Import Restrictions: –0.026*** (T-statistic: –2.8)
    - Export Support: 0.092*** (T-statistic: 4.0)
    - Export Restrictions: –0.025* (T-statistic: –2.0)
    - R-squared (percent): 2.0
  - Regression 3b (TV Product & TV Cty-pair; Number of FE 96,082):
    - Import Restrictions: –0.028*** (T-statistic: –3.2)
    - Export Support: 0.004 (T-statistic: 0.2)
    - Export Restrictions: –0.066*** (T-statistic: –5.1)
    - R-squared (percent): 3.3
- Interpretation of disaggregate coefficients (Specification 2 emphasized in text):
  - Import restrictions on average lead to a 2.5 percent import decline (Specification 2 coefficient: –0.026***).
  - Export support measures increase exports by 9.5 percent (Specification 1 coefficient: 0.096***; Specification 2: 0.092***).
  - Export restrictions are associated with a 2.5 percent decrease in exports (Specification 2 coefficient: –0.025*), albeit at a lower level of statistical confidence.

### Aggregate impact on world trade
- Calculation and aggregate estimates:
  - Using the refined product-level estimates and the amount of trade subject to measures, Specification 2 implies that measures distorted aggregate global trade by 0.22 percent.
  - Specification 2 breakdown:
    - Restrictions lowering world trade by 0.18 percent.
    - Export support measures increasing world exports by 0.04 percent.
  - Using coefficients from Specification 3 yields a similar aggregate distortion of 0.28 percent.
  - Specification 1 yields an aggregate distortion of 0.71 percent.
- Footnote calculation (as provided):
  - The calculation is 0.05 (exp( 0.026) − 1) − 0.004 (exp(0.092) − 1) (0.022 (exp(0.025) − 1) = 0.22%.

*Source: ANNEX I. ECONOMETRIC SPECIFICATION AND RESULTS*

---


_Source: https://www.imf.org/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/spn/2010/_spn1007.pdf_