THE EFFECTS OF INSTABILITY ON BILATERAL TRADE WITH IRAQ

This article investigates the effects of instability on Turkish and Iraqi bilateral foreign trade. It focuses on Turkish exports to Iraq and their fluctuations before, during and after periods of regional and global instability. We find the effects of both global and regional instability to be positive on Turkish exports to Iraq. Based on our results, trade costs of instability in Iraq need to be evaluated carefully. Policy implications suggest that a long-term and sustained stability in Iraq might be preferred to a hasty and ad hoc solution.

Ayla Oğuş∗ and Can Erbil∗∗ ∗

Ayla Oğuş is Assistant Professor at the Department of Economics of the Izmir University of Economics. Can Erbil is Assistant Professor at the Department of Economics and International Business School of Brandeis University, USA; and Research Fellow for Global Economic Modeling Network (EcoMod), Brussels. ∗∗

1

T

o focus on Turkish and Iraqi bilateral foreign trade, we look at Turkish exports to Iraq and inspect its fluctuations before, during and after periods of regional and global instability. 1 Although our analysis has connections to the literature in the study of trade and conflict, we diverge from the existing pool of papers in several ways. Firstly, we are not interested in how trade affects conflict, but rather the opposite. 2 We are also not interested in the causality question. 3 Neither does this paper intend to measure the costs of Iraqi wars for Turkey. 4 Lastly, while most of the existing literature is concerned with the direct effects of instability on trade, we are looking at the effects on Turkey of Iraq’s wars and conflict with other countries. 5 Our main motivation is similar to Barbieri and Levy’s (1999) “Sleeping with the Enemy: The Impact of War on Trade” 6 and Anderton and Carter’s (2001) “The Impact of War on Trade: An Interrupted Time-Series Study.” 7 The entire region in general and Iraq in particular, has experienced repeated and prolonged periods of conflict and war. To quantify “instability” in a practical way, we define the period of regional instability as the duration of the Iran-Iraq war (1980-1988) and the periods of global instability as the duration of the two Gulf wars (1990-1991 and 2003-2004). We find the effect of both global and regional instability to be positive on Turkish exports to Iraq. The UN embargo and changes that have been implemented on it from time to time, along with the mismanaged “Oil-For-Food” program of the UN make it more difficult to build a complete dataset and identify the narrow definition of “conflict”. 8

Data and Methodology We investigate the effects of instability on Turkish and Iraqi bilateral foreign trade using a panel data set of exports in 60 sectors in the years 1980 through 2004. Data on total exports of Turkey and exports to Iraq are from the United Nations Comtrade Database. We have used data by Standard International Trade Classification (SITC) Revision 1 since sectoral trade 1

The import data is believed to be less reliable (the investigation about the corruption of the UN’s “Oil-ForFood” program and the unrecorded border trade between Iraq and Turkey are some of the reasons). 2 Philippe Martin, Thierry Mayer and Mathias Thoenig, in “Make Trade not War?,” Centre for Economic Policy Research Discussion Papers, No.5218, (September 2005) question if the intuition that trade promotes peace is true. They claim that “multilateral trade openness increases more the probability of war between countries which are close to each other”. 3 James D. Morrow in “How Could Trade Affect Conflict?,” Journal of Peace Research, Vol.36, No.4 (1999), pp. 481-489, finds that “if higher trade flows reduce both sides’ resolve for war, then the effect of trade on the likelihood of conflict is indeterminate”. 4 See Scott Wallsten and Katrina Kosec, “The Economic Costs of the War in Iraq,” AEI-Brookings Joint Center for Regulatory Studies Working Paper Series, No.05-19, (September 2005) and William D. Nordhaus, “The Economic Consequences of a War with Iraq,” NBER Working Paper Series, No.9361, (December 2002). 5 See Han Dorussen, “Balance of Power Revisited: A Multi-Country Model of Trade and Conflict,” Journal of Peace Research, Vol.36, No.4 (1999), pp. 443-462, for extending the model for more than two players and emphasizing that intra-alliance trade may be more efficient. 6 Katherine Barbieri and Jack S. Levy, “Sleeping with the Enemy: The Impact of War on Trade,” Journal of Peace Research, Vol.36, No.4 (1999), pp. 463-479, claim that “in most instances war has no permanent longterm effect on trading relationships, in fact, trade often increases in the postwar period”. 7 Charles H. Anderton and John R. Carter, “The Impact of War on Trade: An Interrupted Times-Series Study,” Journal of Peace Research, Vol.38, No.4 (2001), pp. 445-457 disagree with Barbieri and Levy, (1999) pp. 463479 and find “reasonably strong evidence that major power war is associated with a decline in trade relative to pre- and postwar periods”. 8 See Kenneth Katzman and Christopher M. Blanchard, “Iraq: Oil-For-Food Program, Illicit Trade, and Investigations,” Congressional Research Service Report for the US Congress, Order Code RL30472, (14 June 2005).

2

data for a longer period was available under this classification. We extracted data for SITC 01-99. There are ten broad categories under this classification: food and live animals (00-09), beverages and tobacco (10-19), crude materials, inedible except fuel (20-29), mineral fuels, lubricants and related materials (30-39), animal and vegetable oils and fats (40-49), chemicals (50-59), manufactured goods classified chiefly by material (60-69), machinery and transport equipment (70-79), miscellaneous manufactured articles (80-89), commodities and transactions not classified according to kind (90-99). The real exchange rate index is from the Central Bank of Turkey database. Table 1. 2004 Export Figures Turkish Exports (US $) Food (0) Beverages and Tobacco (1) Crude materials (2) Minerals (3) Animal and vegetable oils (4) Chemicals (5) Manufacturing (6) Machinery and Transport (7) Misc. Manuf. (8) Other (9)

Share in Total Exports

Exports to Iraq (US $)

5,023,506,716 590,939,352

8 % 1%

248,500,765 43,979,261

Share in Total Exports to Iraq 14 % 2%

1,451,745,670 1,433,163,111 205,450,042

2% 2% 0%

4,318,343 378,002,648 11,086,95

0 % 21 % 0%

2,503,467,041 18,518,250,354 18,181,644,219

4% 29 % 29 %

118,264,658 539,969,652 357,029,272

7% 30 % 20 %

14,983,144,584 138,215,013

24 % 0%

119,591,618 1,089

7% 0%

The table above presents the volume of total Turkish exports and exports to Iraq for 2004 to present the relative weight of each sector in exports. Exports to Iraq mirror total exports in the largest export sectors. 29 percent of Turkish exports and 30 percent of exports to Iraq are in manufactured goods. Machinery and transport constitute 29 percent of total exports and 20 percent of exports to Iraq. Mineral exports constitute a minor percentage of total exports (2 percent) but with 21 percent, represent the second largest sectoral group in exports to Iraq. Share of food exports to Iraq (14 percent) is also greater than the share of food exports in total exports (8 percent). Even though we do not employ a full-fledged trade model, our choice of regressors captures the basic dynamics of trade models. We only look at exports because imports from Iraq are more sporadic in nature and they are concentrated in relatively few sectors which make statistical analysis less reliable due to small sample bias. Furthermore, unreported trade is a bigger issue for imports and thus more likely to taint statistical results. We include total Turkish sectoral exports as an indicator of size of trade, the real effective exchange rate index to measure changes in terms of trade, trend variables to capture the increasing trading volume due to liberalization policies, dummy variables for global and regional crises and a dummy variable for the Customs Union (CU) period which represents another breakthrough in Turkish foreign trade. The global instability dummy variable takes the value one for the two Gulf wars (1990-1991 and 2003-2004) and is zero elsewhere. The regional instability dummy variable is one for the duration of the Iran-Iraq (1980-1988) war and is zero elsewhere. The CU dummy takes the value one starting in 1996. 3

As can be seen in the above table, sectoral exports vary significantly in volume, and this suggests that the appropriate estimation technique is the least-squares dummy-variable approach also known as the covariance or the fixed-effects model. This model is usually denoted as: Y it = ∑ β j X it, j + ∑ γ j Dit, n + ε it j n

where Xs are the explanatory variables and Ds are the dummies for each sector. It is common practice, when dealing with pooled cross-section and time-series data, to also estimate what is known as the error component or random-effects model and compare the relative efficiency of the two models 9 (Hsiao 1989). The random effects model failed to provide an increase in the efficiency of the parameter estimates and therefore is not presented here.

Results In the tables below we see the effects of instability on Turkish exports to Iraq. The effect of both global and regional instability is positive on Turkish exports to Iraq. Only results that are statistically significant at the 10 percent level are presented. The estimates for all sectors are results based on 60 subsectors of the 10 broad sectors defined above. We had enough data to analyze food (SITC Rev. 1 Code 0), crude materials (SITC Rev. 1 Code 2), chemicals (SITC Rev. 1 Code 5), manufactured goods (SITC Rev. 1 Code 6) and miscellaneous manufactured goods (SITC Rev. 1 Code 8) separately. During periods of regional instability sectoral exports to Iraq increase by approximately 20 million dollars per year. Since the average sectoral export to Iraq is around 7 million dollars (See Table A1), this represents a huge increase. However, we should point out that the average sectoral exports are low since there are years where exports to Iraq are zero in some or all subsectors. Trade in certain subsectors is sporadic over the 1980-2004 period where we may see exports for several years followed by no trade for a few years. During the embargo, there are no exports to Iraq as reported in the Comtrade database although some other sources will report positive export figures for this period.

9

The reasoning behind the random-effects model is that, since the error term is generally considered to represent the effect of omitted variables, and some of the omitted variables could represent factors peculiar to individuals or time periods or both, the error term should be treated as consisting of three components. Cheng Hsio, Analysis of Panel Data (Cambridge, UK: Cambridge University Press, Paperback edition, 1989).

4

Table 2. Regression Results Coefficient Turkish Xs Xrate Global Regional CU Trend Regression F Number of Observations

All 0.00157 (0.0008) -1030 (35889) 9087663 (1529285) 19778843 (2011429) -6999149 (2329304) 1165564 (214896) 9.4957 1435

*

*** *** ***

Crude 0.0294 (0.0048) 1031 (13502) -759665 (600255) 2827733 (522347) -1109526 (551407)

*** 14.7581 225

***

*** **

Food 0.0053 (0.005) 56666 (73161) 6659703 (3113627) 15358766 (4086003) -11034696 (4749499) 980123 (435431) 11.6946 250

** *** ** **

Chemicals 0.0286 (0.0083) -132918 (43759) 3470823 (1891102) 11595757 (2473627) -4822725 (2841213) 613742 (271388) 7.4537 225

*** *** * *** * **

Manuf 0.0082 (0.0042) -106109 (165758) 21536998 (7097135) 59628015 (9371412) -15341021 (10726724) 3308491 (1029315) 8.3037 225

**

*** ***

***

Misc Manuf -0.0012 (0.0004) 603 (33448) 5555696 (1423842) 12285349 (1870864) -3706565 (2171404) 888101 (201007) 8.5980 175

***

*** *** * ***

Notes: What we see when we look at the exports figures to Iraq is that in subsectors where there was trade, levels declined after the first Gulf war and only managed to pickup in the last couple of years. In the sectors Turkey was not exporting a lot in the earlier years, we see major leaps in the last couple of years. One might argue that the positive effect of global instability is due to end to the embargo in which case it is not related to instability but to the embargo. However, the CU dummy can be regarded as a dummy for the Oil-For-Food Program and thus we are accounting for the embargo. Furthermore, when we repeated the analysis on a sample restricted to years where there was trade more than 100,000 dollars (real dollars-base 2004), the effect of regional and global instability had positive signs but the magnitude of the effect of global instability was larger. Therefore, we argue that we have captured a positive effect on instability on exports which goes beyond an increase merely due to the lifting of the embargo in 2003.

5

Table 3. The Effect of Instability (Global and Regional) Effect of global instability on Iraq (US $)

Effect of regional instability on Iraq (US $)

All sectors Food

9,087,664 6,659,703

19,778,843 15,358,766

Crude materials

Not statistically significant 3,470,823 21,536,998 5,555,696

2,827,733

Effect of a million $ increase in Turkish sectoral exports on sectoral exports to Iraq (US $) 1,578.5 Not statistically significant 29,400

11,595,757 59,628,015 12,285,349

28,600 8,214 879

Chemicals Manufacturing Misc. Manuf.

The effect of regional instability ranges form approximately 3 million to 60 million dollars per year. The lower end is for crude materials and the high end is for manufacturing. As noted above, manufacturing represents 30 percent of Turkish exports to Iraq and totaled 540 million dollars in 2004. The effect of regional instability for a subsector is over 10 percent of the total exports. The effect on total exports to Iraq would be around 90 percent. Global instability also has a positive effect on exports to Iraq. The effect on average sectoral exports is 9 million dollars per year which is still more than the sample average of 7 million dollars. The effect of global instability ranges from 3.5 million to 21.5 million dollars. The lower end is for chemical exports. For all other sectors analyzed except for manufactured goods, the effect is not much larger: 5.5 million dollars for miscellaneous manufacturing and 6.7 million for food. The effect on manufacturing is 21.5 million dollars per year which represents 5 percent of total exports. This would indicate a 50 percent increase per year in periods of global instability. The effect of increasing Turkish exports by sector is also reported. This effect is largest for crude materials and chemicals. Every million dollar increase in Turkish sectoral exports leads to a 29,000 dollars increase in sectoral exports to Iraq. For example, if Turkish exports in chemicals were to increase by 1 percent from the 2004 level of 2.5 billion dollars, this would lead to a 725,000 dollars increase in chemical exports to Iraq which represents a 0.6 percent increase.

7

Table 4. Sectors where Iraq is a major trade partner (2003-2004) Share of Sectoral Exports to Iraq 76 % Gas, natural and manufactured (34) 66 % Electric energy (35) 33 % Beverages (11) 33 % Wood and cork manufactures excluding furniture (63) 19 % Miscellaneous food preparations (09) 15 % Animal and vegetable oils and fats, processed (43) 15 % Sugar, sugar preparations and honey (06) 11 % Cereals and cereal preparations (04) 10 % Feed. Stuff for animals excl. Unmilled cereals (08) 10 % Perfume materials, toilet & cleansing preperations (55) 10 % Dairy products and eggs (02) Iraq is a major destination for Turkish exports in the above subsectors. In the food sector, exports to Iraq represent at least 10 percent of Turkish sectoral exports in subsectors 02, 04, 06, 08 and 09. The food sector, by 2004 figures is a 5 billion dollar export industry hence the share of exports to Iraq is noteworthy. Natural gas and electric energy where Iraq receives over two thirds of Turkish exports are subsectors of the minerals sector which accounts for 1.4 billion dollars of exports in 2004. In Turkey’s largest export industry, manufacturing, Iraq is a major export destination in only wood and cork manufacturing accounting for 33 percent of sectoral exports. Conclusions and Policy Implications There are important policy implications of our results. Our analysis shows that during periods of conflict there is an increase in the average exports of Turkey to Iraq. This is not entirely puzzling if one keeps in mind that although conflict makes foreign trade more costly by destabilizing trade routes, trade that Iraq might otherwise have with other trading partners, is diverted to Turkey. One reason that the trade route between Turkey and Iraq is less sensitive to conflict is proximity. There are no other major economies among the bordering countries with the exception of Iran, and unlike Iran, Turkey and Iraq were never engaged in direct military conflict. 10 Based on our results, trade costs of instability in Iraq, need to be evaluated carefully. The losses in the embargo period and the potential losses due to economic strains of instability on the Iraqi economy need to be contrasted with the positive effect of instability on Turkey’s exports to Iraq in the pre and post-embargo period. Conflicts Iraq had at the regional or global level may be helping to strengthen the trade relations between Iraq and Turkey as Turkey has to compete with less of Iraq’s trading partners in times of conflict. As the market share of Turkish exports in Iraq increases, Turkey might be gaining the advantage of becoming the

10

Relations have been sour from time to time, however. For a discussion see Ibrahim Al-Marashi, “Middle Eastern Perceptions of US-Turkey Relations After the 2003 Iraq War,” Turkish Policy Quarterly, Vol.4 , No. 1, (Spring 2005) and Michael Rubin, “A Comedy of Errors: American-Turkish Diplomacy and the Iraq War,” Turkish Policy Quarterly, Vol.4 , No.1 (Spring 2005).

8

largest and most comprehensive trading partner in many sectors in Iraq. These sectors are, and will be, rebuilding over the next decade. 11 If it is not clear that Turkey is losing in one of its major export markets in times of instability then, purely from a trade point of view, one might be indifferent between stabilization in Iraq “from within” and “from above”. 12 Stabilization in Iraq from within, although it will take longer, does not necessarily constitute an inferior alternative. 13 It is undisputable that a stable and growing economy in the region is more beneficial to Turkey than a country crippled by war and conflict. However, our findings suggest that Turkey can afford to support a policy standpoint which allows Iraq enough time to reach this point at its own pace, since Turkey might have some advantages in trade during the process of stabilization in Iraq. The reader should be reminded once more that this is purely from a foreign trade point of view based on an empirical analysis of officially reported foreign trade data from the UN sources. It is widely acknowledged that a portion of the trading activity goes unreported at the TurkishIraqi border. Hence our analysis is based on conservative estimates of Turkish exports to Iraq.

11

For a discussion of economic reconstruction of Iraq see Ayad Allawi, “The Democratization Process in Iraq,” Turkish Policy Quarterly, Vol.4 , No. 2, (Summer 2005). 12 During January-September 2005, the largest increase in exports was experienced with Iraq. Turkish exports increased by 52.9% up to 2 billion dollars. In 2005, Iraq became Turkey’s 7th largest export market. (from the Undersecreteriat of the Prime Ministry for Foreign Trade). 13 For a detailed discussion see Arend Lijphart , Patterns of Democracy : Government Forms and Performance in Thirty-Six Countries, (Yale University Press, 1999) and Guillermo O'Donnell, The Quality Of Democracy: Theory And Applications, (Helen Kellogg Institute for International Studies, 2004).

9

11