The Macroeconomic Effects of Trade Tariffs: Revisiting the Lerner Symmetry Result Jesper Linde and Andrea Pescatori

EC March 2018

1

The views expressed herein are those of the author and should not be attributed to the IMF, its Executive Board, or its management".

Motivation

• Recent policy proposals include a border adjustment tax (BAT)

• Debate strongly influenced by an old trade theory result (Lerner 1936):

• The exchange rate appreciates (at same rate) • No effect on real allocation and other prices

• How general is the result? • What deviations are quantitatively relevant?

Import tariff + (equal rate) export subsidy

(Lerner Symmetry Theorem)

2

What we do

We use a medium-scale two-country DSGE model (SIGMA) to analyze and *quantify*:

• The effects of Import tariffs and export subsidies (Lerner symmetry result) • The role of international asset markets and risk sharing • The role of exchange rate adjustment and currency pegs • The role of anticipation • The role of alternative pricing mechanisms

• Trade wars • CM vs IM • Full symmetric retaliation

3

Literature

Trade literature:

No uncertainty: Two goods environment, Lerner (1936); multiple goods, McKinnon (1966), Grossman (1980);

Symmetry breaks down: permanent vs. transitory, Razin and Svensson (1983); Dynamic, RE: Eichengreen (1981,1983)

Previous trade theory results can be overturned when tariffs are uncertain

(Tariff) Uncertainty: Stockman and Dellas (1986) the role of asset markets Barari Lapan (1993);

More recently:

DSGE model, Erceg et al (2017), Amiti et al (2017) – Barbiero et al (2017),

Theory: Costinot Werning (2017)

4

Features of the Model

• Medium-scale quantitative DSGE model calibrated to match US data

• Two (identical) countries

• Capital accumulation, financial accelerator, liquidity constrained households

• Trade: Home bias in consumption and investment RER not constant (even with LOP)

• Sticky prices and wages (alternative pricing regimes) • LCP: sticky import prices (low exchange rate pass-through to import prices)

• LCP: pre- or post- tax (tariff/subsidy) pricing (alternative pricing)

• PCP: sticky export prices (full exchange rate pass-through to import prices)

• Real rigidities: consumption habit, investment and import and export adjustment costs

• trade assumed balanced in steady state (Costinot & Werning: not relevant for symmetry)

N.B.: LCP = local-currency invoicing; dollar-invoicing: LCP and PCP* 5

The Lerner Symmetry Result

• With post-tax pricing, Lerner symmetry holds exactly independently of

• nominal rigidities (sticky import/export prices, sticky wages, LCP or PCP)

• Real rigidities (investment, import/export adjustment costs, habit, liquidity constraint agents)

• Financial accelerator

• Monetary policy rule (except reacting to exchange rate)

• Country size

• Intuition:

• If import tariff and export subsidy have (move by) the same rate

• Full exchange rate adjustment offsets both the import tariff and export subsidy’s distortions (i.e., the ToT is unchanged)

• Assumption TB =0: No fiscal implications

6

Lerner Theorem: Mechanics

Deviations from LOP Home country Foreign country Import Phillips curve (Local Currency Pricing)

USD per foreign currency: up USD depreciation

Foreign export prices in foreign units

Foreign marginal costs 7

Lerner Theorem: Mechanics

Deviations from LOP Home country Foreign country Import Phillips curve

Import tariff

Export subs.

8

Lerner Theorem: Mechanics

Assume: deviations from LOP =0 Price stickiness irrelevant

Assume the USD fully adjusts

Exchange rate adjustment mechanism (UIP – consistent with IM) The “equilibrium” exchange rate adjusts, non need for interest rates to move! It is an equilibrium Long-term (equilibrium)

exchange rate

Equilibrium?

Outside the mode: coordination problem (economists can help) 9

Figure 1: Tariff and subsidy: perfect symmetry

Permanent and anticipated 10p.p. import tariff and export subsidy shock

10

Deviations from Symmetry: alter. pricing mechanisms

Lerner symmetry temporarily breaks down under pre-tax pricing (and sticky prices)

LCP: low exchange rate pass-through on import prices (in both countries)

• Pre-import tariff and pre-export subsidy pricing: both tariff and subsidy are like a sales tax (both full pass-through);

• Pre-import tariff pricing: tariff is like a sales tax (full pass-through);

Intuition: Asymmetric pass-through breaks the exchange rate offset on import prices

Low pass-through

full pass-through 11

Figure 5 Alternative pricing assumption

12

Deviations from Symmetry: alter. exch. rate mechanisms

Lerner symmetry permanently breaks down under

• Complete markets

• Wealth insured (no wealth effect from price distortions) …

• … but price distortions remain

Lerner symmetry temporarily breaks down under

• Slow nominal exchange rate adjustment:

• The exchange rate moves (appreciates) too slowly (or never: currency pegs)

• Currency pegs affect world GDP

• Anticipation of the trade policy shock

• The exchange rate moves (appreciates) too early

13

Lerner Theorem: Mechanics

Anticipation Exchange rate moves in anticipation

Slow exchange rate adjustment

Same mechanics but past level slows down the adjustment

UIP adj parameter

14

Lerner Theorem: Mechanics

Complete Asset Markets log-utility

real rate differential “surprise”

(real) UIP condition

Dollar-expenditures equated across countries RER moves values not volumes insured

Home bias, tariffs, … RER no constant

15

Deviations from Lerner

16

Deviations from Lerner: currency pegs

17

Deviations from the Lerner Symmetry: complete markets

Complete international asset markets

• The first welfare theorem does not hold

• CM do not guarantee efficiency

• CM insure a wealth effect but do not fix price distortions

• World GDP and trade unaffected (the pie does not change)

• Lower welfare at home (lower consumption) higher output and TB

• Nominal rigidities: short-run drop in output

• Exchange rate adjustment only 60% of the shock

18

Deviations from Lerner: Complete Markets

19

Summary

• Anticipation: • minor negative effect on home output; • global trade and output NOT affected

• Gradual exchange rate adj: • modest output boom at home • global trade and output NOT affected

• Currency peg: • Big drop in foreign output • Global trade and output decline

• Complete markets (permanent effects): • Fall (increase) in domestic consumption (output) , eventually • Global trade/output NOT affected

• Alter. Pricing: • Exchange rate under-shooting; domestic output temporarily gain

20

Trade Wars

Retaliation: 20 percent tariff in foreign country

IM: foreign suffers the most

CM: symmetric effects

Symmetric retaliation: BAT in foreign country: No effects

21

Trade Wars: Full retaliation

Deviations from LOP Home country Foreign country Equilibrium SHOW TAUM=TAUM* TAUX=TAUX* - IN DIFF – INITIAL LEVELS DON’T MATTER

22

Trade Wars: Asymmetric retaliation

Deviations from LOP Home country Foreign country Equilibrium SHOW 2*TAUM=TAUM* TAUX=taum; TAUX*=0

23

Trade Wars

24

Figure 4: Trade Wars

25

Summary

• Fully symmetric retaliation: No effects

• Retaliation: • Global trade and output fall

• Both countries’ output drop

• Permanent effects

• Short-run dynamics determined by nominal rigidities

26