1

Limiting Currency Volatility toStimulate Goods Market Integration: A Price Based Approach

byDavid Parsley and Shang-Jin Wei

Vanderbilt University Brookings

Institution

2

Introduction

Does exchange rate stabilization affect goods market integration?

We distinguish two types of stabilization Instrumental - reducing exchange rate

volatility through intervention in the fx market

Institutional - reducing volatility through an explicit currency board or common currency

3

Introduction

Two strands of empirical research into goods market integration: studies examining actual flows of goods price-based studies (e.g., PPP, LOP)

We adopt a price-based approach a unique multi-country data set on

prices of very disaggregated products (e.g., light bulbs & onions)

4

Studies of observed trade flows

McCallum (95), Wei (96), Heliwell (98) conclusion: observed volume of trade

across national boundaries are much less than within countries

Rose (2000), Frankel and Rose (2000), Rose and Engel (2000), Rose and van Wincoop (2001) conclusion: common currencies increase

bilateral trade by as much as 300%

5

Limitations of observed trade flows

Two countries may have similar endowments and autarkic prices low trade

two countries may trade extensively with a 3rd country but little w/each other

Wei (1996) argues welfare implications from observed trade flows need auxiliary assumptions

6

Intuition

If two countries produce similar (highly substitutable) output, increased trade may raise welfare only marginally

Alternatively, if two countries have distinct comparative advantages, a slight rise in trade may substantially raise welfare

7

Economist Intelligence UnitPrice Data

Local currency price comparisons for > 160 goods and services from up to 122 cities

We select 95 goods and 83 cities

8

1. Apples (1 kg) (supermarket) 49. Onions (1 kg) (supermarket)2. Aspirin (100 tablets) (supermarket) 50. Orange juice (1 l) (supermarket)3. Bacon (1 kg) (supermarket) 51. Oranges (1 kg) (supermarket)4. Bananas (1 kg) (supermarket) 52. Peaches, canned (500 g) (supermarket)5. Batteries (two, size D/LR20) (supermarket) 53. Peanut or corn oil (1 l) (supermarket)6. Beef: filet mignon (1 kg) (supermarket) 54. Peas, canned (250 g) (supermarket)7. Beef: ground or minced (1 kg) (supermarket) 55. Pork: chops (1 kg) (supermarket)8. Beef: roast (1 kg) (supermarket) 56. Pork: loin (1 kg) (supermarket)9. Beef: steak, entrecote (1 kg) (supermarket) 57. Potatoes (2 kg) (supermarket)10. Beef: stewing, shoulder (1 kg) (supermarket) 58. Razor blades (five pieces) (supermarket)11. Beer, local brand (1 l) (supermarket) 59. Scotch whisky, six years old (700 ml) (supermarket)12. Beer, top quality (330 ml) (supermarket) 60. Sliced pineapples, canned (500 g) (supermarket)13. Butter, 500 g (supermarket) 61. Soap (100 g) (supermarket)14. Carrots (1 kg) (supermarket) 62. Spaghetti (1 kg) (supermarket)15. Cheese, imported (500 g) (supermarket) 63. Sugar, white (1 kg) (supermarket)16. Chicken: fresh (1 kg) (supermarket) 64. Tea bags (25 bags) (supermarket)17. Chicken: frozen (1 kg) (supermarket) 65. Toilet tissue (two rolls) (supermarket)18. Cigarettes, local brand (pack of 20) (supermarket) 66. Tomatoes (1 kg) (supermarket)19. Cigarettes, Marlboro (pack of 20) (supermarket) 67. Tomatoes, canned (250 g) (supermarket)20. Coca-Cola (1 l) (supermarket) 68. Tonic water (200 ml) (supermarket)

Sample of Economist Price Data

9

Abu Dhabi, UAE Colombo, Sri Lanka London, United Kingdom San Francisco, United StatesAmman, Jordan Copenhagen, Denmark Los Angeles, United States San Jose, Costa RicaAmsterdam, Netherlands Dakar, Senegal Luxembourg, Luxembourg Santiago, ChileAsuncion, Paraguay Detroit, United States Madrid, Spain Sao Paulo, BrazilAthens, Greece Douala, Cameroon Manila, Philippines Seattle, United StatesAtlanta, United States Dublin, Ireland Mexico City, Mexico Seoul, South KoreaAuckland, New Zealand Guatemala City, Guatemala Miami, United States Singapore, SingaporeBahrain, Bahrain Helsinki, Finland Montevideo, Uruguay Stockholm, SwedenBangkok, Thailand Hong Kong, Hong Kong Moscow, Russia Sydney, AustraliaBeijing, China,P.R. Honolulu, United States Mumbai, India Taipei, TaiwanBerlin, Germany Houston, United States Nairobi, Kenya Tehran, IranBogota, Colombia Istanbul, Turkey New York, United States Tel Aviv, IsraelBoston, United States Jakarta, Indonesia Oslo, Norway Tokyo, JapanBrussels, Belgium Johannesburg, South Africa Panama City, Panama Toronto, CanadaBudapest, Hungary Karachi, Pakistan Paris, France Tunis, TunisiaBuenos Aires, Argentina Kuala Lumpur, Malaysia Pittsburgh, United States Vienna, AustriaCairo, Egypt Kuwait, Kuwait Port Moresby, Papua New Guinea Warsaw, PolandCaracas, Venezuela Lagos, Nigeria Prague, Czech Republic Washington DC, United StatesCasablanca, Morocco Libreville, Gabon Quito, Ecuador Zurich, SwitzerlandChicago, United States Lima, Peru Riyadh, Saudi Arabia

Cities Included

10

Let be the U.S. dollar price of good k in city i at time t. For a given city pair (i,j) and a given good k at a time t, we define the common currency percentage price difference as:

tkjPtkiPtkijQ ,,ln,,ln,,

Our Approach

11

Our Approach

We study all bilateral price comparisons the data allow.

There are 3403 city pairs (=(83x82)/2) – each with 11 (annual) time periods.

Thus, for each of the 95 prices the vector of price deviations will contain 37,433 (3403x11) observations without missing values.

12

Dispersion in Price Differences

We focus on the cross sectional dispersion (across goods) of common currency price differentials for each city-pair and time period

Any particular realization of the common currency price differential, Q(ij,k,t) can be either positive or negative without triggering arbitrage as | Q(ij,k,t) | < the cost of arbitrage

13

Dispersion and Market Integration

The existence of arbitrage costs implies that must fall within a range

Any reduction to barriers to trade (i.e., movements toward market integration) should reduce the no-arbitrage range. Therefore the strategy we adopt is to study a measure of the dispersion of Q(ij,k,t) through time

14

Table 1: Percentage Price Deviations in Absolute Value (averaged over all years)

Asuncion-TaipeiLight Bulbs 65.4Onions 115.0

Paris-Vienna (1990-1998, pre-euro)Light Bulbs 13.4Onions 45.3

Paris-ViennaLight Bulbs 11.4Onions 40.1

Chicago-HoustonLight Bulbs 8.9Onions 42.7

15

Table 2: Dispersion and its DeterminantsAverages across city pairs and time

Observations Dispersion Distance Exch Rate Vol TariffsAll City Pairs 36531 0.638 8215 0.067 22.3

Hard Peg City Pairs 454 0.576 8602 0.001 9.8

U.S. Only City Pairs 975 0.378 2681 0.0 0.0

CFA Only City Pairs 110 0.629 3139 0.027 41.9

Euro City Pairs 110 0.419 1273 0.0 0.0

16

Intercity Price Dispersion

1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 20000.30

0.35

0.40

0.45

0.50

0.55

0.60Chicago-Houston

Chicago-Paris

Paris-Vienna

17

Regression analysis

We estimate the following baseline equation:

tijij

ijijij

dummiestimeandcityTariff

EuroUSCFAHPeg

xrvoldistdisttijqV

,8

7654

32

11 )()ln()ln(),(

18

Table 3: Benchmark Regression Results

Equation 1 Equation 2 Equation 3Log Distance 0.1267 0.1320 0.1216

(0.0213) (0.0230) (0.0229)

Log Distance Squared -0.0060 -0.0063 -0.0057(0.0014) (0.0015) (0.0015)

Nominal Exchange 0.0393 0.0362 0.0542Rate Variability (0.0114) (0.0116) (0.0100)

Hard Peg -0.0438 -0.0325 -0.0248(0.0065) (0.0070) (0.0067)

CFA -0.0149 -0.0102 -0.0090(0.0148) (0.0156) (0.0152)

U.S. -0.1104 -0.1015 -0.0955(0.0044) (0.0048) (0.0047)

Euro -0.0342 -0.0247 -0.0165(0.0056) (0.0059) (0.0055)

Weighted Avg. Tariff 0.0044 0.0041 0.0043(0.0001) (0.0001) (0.0001)

Absolute Wage 0.0021 0.0302Difference (0.0014) (0.0038)

Absolute Wage -0.0025Difference Squared (0.0003)

Adjusted R2 .23 .22 .23Number of Observations 27406 21863 21863

Robust standard errors are in parenthesis. All equations include city and time fixed effects.

19

Discussion

Dispersion increases with distanceExchange rate variability increases

dispersion reducing it to zero from the sample

average, reduces dispersion by .26% (=.067*.039*100)

Participating in a Hard Peg reduces dispersion by 4.4% - an order of

magnitude bigger

20

Discussion (continued)

Being a member of the CFA has no effect

Being a member of the euro ~ to hard peg

Being in a political union (US) has the largest institutional effect

21

Tariff Equivalents

Effect of the euro: ~ 4 percentage point reduction in tariffs on the same order of magnitude as the elimination of tariffs under the common market program

Effect of reducing xr volatility to zero for any random pair of countries is only 0.3%

Effect of political & economic Union (U.S.) ~13 percentage point reduction in tariffs

22

Summary

Institutional exchange rate stabilization has a much larger effect than instrumental stabilization

reducing xr vol < hard peg < full economic & political integration

The effect is non-trivial. On the order of the common market effect

A non-credible peg (CFA) has no effect

23

Robustness & Extensions

Additional explanatory variablesre-definitions of explanatory

variablesdifferent measures of dependent

variablealternative econometric

specifications

24

Conclusions

Institutional exchange rate stabilization matters for goods market integration.

The economic benefits of currency unions (Hard pegs) are an order of magnitude larger than simply reducing exchange rate vol to zero

Our results suggest that further economic and political integration can have an additional substantial impact on goods market integration