Deutscher Tropentag 2004: Conference on Rural Poverty Reduction through Research for Development and Transformation

Transcription

1 Deutsher Tropentag 2004: Conferene on Rural Poverty Redution through Researh for Development and Transformation Otober 5-7, 2004, Humboldt-Universität zu Berlin, Agriultural and Hortiultural Faulty Agriultural Trade Liberalisation in the WTO and its Poverty Impliations: the Case of Rural Households in Northern Vietnam Daude a, Sabine a Universität Hohenheim, Institut für Agrar- und Sozialökonomie in den Tropen und Subtropen, Lehrstuhl für landwirtshaftlihe Entwiklungstheorie und -politik (490a), Stuttgart, Germany. daude@uni-hohenheim.de Abstrat The objetive of the paper is to analyze poverty impliations of the agriultural liberalization proess disussed within the WTO Doha Negotiations for rural households in Northern Vietnam. In preparation of the WTO Ministerial Meeting in Canún, Mexio, in September 2003, the WTO suggested in the Harbinson Paper further liberalization of the agriultural setor to break up the deadlok in the negotiations. The impat of these suggestions for Vietnam s agriulture is simulated. Vietnam is urrently not a WTO member but it is in the proess of aession negotiations. The objetive of the Vietnamese Government is to aede to the WTO within the Doha Round by Two agriultural trade liberalization senarios are simulated: one with Vietnam being member of the WTO and a seond with Vietnam not being a member of the WTO. The hosen methodology is a maro-miro approah whih uses first a general equilibrium setting (standard GTAP model) to assess and analyse the prie hanges of the Harbinson suggestions. The data for the maro simulation are from the GTAP database (2001). In a seond step, to estimate poverty impliations for the households, data from a household survey from Northern Vietnam (2001) are used together with the prie hanges resulting from the maro simulation. Within this post-simulation analysis household data are mapped to the GTAP data. In order to estimate the net inome and poverty effets of trade liberalization the inome side (via fator prie hanges) and the onsumption side (via onsumption prie hanges) are onsidered. Households are lassified aording to their inome struture and their onsumption shares for different produt ategories. Results show how rural farmers in Northern Vietnam may be affeted by agriultural trade liberalisation and the ountry s aession to the WTO; the effets on farmers net inome and thus their poverty situation are assessed. 1 Bakground and aim of the study Within the liberalization efforts during the last years the World Trade Organization (WTO) played an important atalyti role by establishing a multilateral trading system where up to January member ountries were aepted by the Ministerial Sabine Daude 1/12 DTT Otober 2004

2 Conferene. However, during the present Negotiation Round, also alled the Doha Development Round, some drawbaks ourred and sheduled deadlines ould not be met, espeially not for the agriultural setor. In preparation of the Ministerial Meeting in Canún, Mexio, in September 2003, the WTO suggested further liberalization for the agriultural setor in the so alled Harbinson Paper to break up the deadlok in the negotiations. This was meant to lay a ground for further disussions in the sensitive area of agriultural liberalization. Although the partiipating ountries ould not agree during the Ministerial Conferene in Canún on further ommitments to liberalize the multilateral trading system, efforts are still under way to bring the Doha Development Round to a suessful end with further agreements to be reahed. An important step was the agreement in July 2004 on a framework for modalities whih frees the way for the further Doha Work Programme. Vietnam is urrently not a WTO member but is in the proess of aession negotiations. The desire to join the WTO is one step within the opening of the Vietnamese eonomy during the last years. The ountry started a reform proess (Doi Moi) from a entrally planned eonomy to a market-oriented eonomy in During this transition Vietnam experiened extraordinary eonomi growth mainly based on trade. After the first years of reform the proess lost some of its momentum due to the Asian risis but also to the hallenges to implement domesti reforms in Vietnam. The objetive of the Vietnamese Government is to aede to the WTO within the Doha Round by The inentive to fully join the trade system of the WTO will largely be driven by the possible gains that an be expeted from the aession besides the wish to promote domesti reforms. The disussion about trade and its role in the development proess an be traed bak over deades. The last years put more emphasis not only on development but on the expliit eradiation of poverty as a preondition for further development within ountries. Theories of the effets of trade on the poor and their poverty situation emerged and attempts to quantify the poverty impats of trade liberalization beame prominent. The hallenge is to trae the effets of international trade liberalization on the individual household level where poverty ours. The objetive of this paper is to analyze impliations of trade liberalization, as disussed within the Doha Round on rural households in mountainous regions of Northern Vietnam. The paper will examine how poverty rates for households in Northern Vietnam are affeted when pries hange due to agriultural trade liberalization. 2 Methodology and data A two-step miro-maro synthesis approah is hosen ombining the GTAP general equilibrium model with a post-simulation analysis to examine the effets of multilateral trade liberalization on rural households in Northern Vietnam. The term miro-maro synthesis is a reent term in the literature examining the relationship between trade and poverty. Used in this ontext it means the sequential linking of a model based on miro-level data with a model based primarily on maro-level data (Reimer, 2002). 2.1 Miro-maro synthesis first step: general equilibrium simulation with the GTAP Model The GTAP modelling framework employs a standard CGE model based on the neolassial theory of firm and household behaviour assuming perfet ompetition, rational and utility optimizing behaviour. The standard model is a omparative stati model whih means that after introduing an exogenous shok like a poliy hange the model works out a new equilibrium in all markets and determines new values for the Sabine Daude 2/12 DTT Otober 2004

3 endogenous variables. The model and its harateristis are extensively doumented in the literature (Hertel, 1997; Brokmeier, 1996). Following the mandate from the Doha Ministerial Delaration and in preparation of the WTO Ministerial Meeting in Canún, Mexio a proposal for further liberalization in the agriultural setor was submitted by the hairperson of the agriultural negotiations, Mr. Harbinson, in February It tried to bridge the immense gaps between parties in the negotiation proess. A slightly revised version of this so-alled Harbinson paper (Marh 2003) is the basis for the following two GTAP simulations. Table 1 summarizes the senarios: Table 1: Content of agriultural trade liberalization senarios Simulation 1: AgrLib Simulation 2: AgrLib+AVn Agriultural trade liberalization following Harbinson II without Vietnam s aession to the WTO - export subsidies are redued to zero - import tax are redued along the shedule following tax bands For industrialized ountries: Tax band > 90% 15% - 90% < 15% redution 60% 50% 40% For developing ountries: Tax band > 120% 60% - 120% 20% - 60% < 20% redution 40% 35% 30% 25% Agriultural trade liberalization following Harbinson II with Vietnam s aession to the WTO - the same export subsidy and import tax redutions as in senario 1 plus: Vietnam implements the agriultural liberalization itself as all other WTO members Vietnam enjoys the lower tariff rates that are applied between all WTO members 2.2 Miro-maro synthesis seond step: post-simulation analysis Having obtained the prie hanges from trade liberalization senarios from the GTAP general equilibrium model the post-simulation analysis aims at showing the impliations at the household level. Households are affeted by prie hanges in the ommodity and the fator markets beause adjustments are transmitted through these two hannels (Reimer, 2002). The net inome effet an thus be obtained as follows: Net inome effet = onsumption effet + earnings effet. The aim of the post-simulation analysis is to math the GTAP prie hanges with the household data in order to obtain the net inome effet and the impliations for poverty. The post-simulation analysis is done in five steps (Friedman, 2000): 1. Mathing the simulated prie hanges to household onsumption 2. Measuring the level-of-living 3. Measuring hanges in the ost-of-living (ompensating variation) 4. Measuring hanges in household earnings 5. Estimating the net effets of trade liberalization The following desription of the methodologial approah is based on these five steps and based on alulations suggested by Neil MCulloh (MCulloh, 2003). Sabine Daude 3/12 DTT Otober 2004

4 2.2.1 Mathing the simulated prie hanges to household onsumption From the GTAP results ommodity prie hanges and prie hanges for different prodution fators are taken. The onsumption information is derived from household surveys. Speial attention is given to the onsumption of self-produed goods. The value of onsumption of the households for different goods is alulated following the ommodity aggregates of the global equilibrium framework. Then mean expenditure shares for different ommodities and different household ategories are derived for different household ategories. The expenditure shares for onsumption goods reveal the importane that prie hanges for these goods have: The bigger the expenditure share the bigger the influene from a prie hange of this produt Measuring the level-of-living In order to measure the level-of-living a binary index poor/non-poor is used (Friedman, 2000). The hoie to ategorize a household as poor or non-poor depends ruially on the definition of a poverty line: if a household falls under this line he is deemed to be poor Measuring hanges in the ost-of-living (ompensating variation) The onsumption effet an be alulated by multiplying the prie hanges with the onsumption expenditure shares that are spent for the different onsumption goods. Friedman suggests to additionally taking into aount onsumption of self-produed agriultural goods. In the ase of dereasing pries for agriultural goods households with self-onsumption are relatively worse off ompared to households that buy their food (Friedman, 2000). Therefore, the imputed values for self-produed goods are treated as negative expenditures to alulate an adjusted measurement. The perentage effet on the ost-of-living an be written as (adapted from MCulloh, 2003): CL p jb = CS jb CS CL p jb with CL : ost-of-living CS jb js p p js js : onsumption expenditure share of onsumption good j bought (1) p jb : prie of onsumption good j bought CS js : onsumption expenditure share of onsumption good j self-produed p js : prie of onsumption good j self-produed Equation (1) shows that the hange in the ost-of-living for a household depends on the differenes in relative pries among the bundle of onsumed goods as well as on the onsumption expenditure budget share spent on eah good. In the ase of dereasing pries both terms beome negative and the seond term is then added and thus inreases the ost-of-living of self-produing households relative to the rest. In the ase of inreasing pries households that onsume their own-produed goods are relatively better of. In this ase the two terms are positive and the seond term is then deduted and thus redues the ost-of-living of self-produing households. So far, this analysis only inorporates the onsumption side. With the next steps the earning side will also be inluded in the analysis. Sabine Daude 4/12 DTT Otober 2004

5 2.2.4 Measuring hanges in household earnings From the GTAP simulations hanges in returns to prodution fators an be derived. These hanges are brought together with the available household data on earnings generation. From the household data different soures of earnings are derived like earnings from wages and salaries, earnings from agriultural prodution, earnings from non-agriultural own business and other minor soures. The share of eah ategory shows the relative importane for total household earnings (MCulloh, 2003). The household earning hange an thus be alulated as (adapted from Friedman, 2000, MCulloh, 2003): h EARN h p i = ES i (3) h EARN pi with EARN : earning h ES i : earning share of prodution fator i for household h h p i : prie of prodution fator i for household h With the different assumptions and the above equation (3) it is possible to alulate the earnings effet of the simulated agriultural trade liberalization via the prie hanges of the prodution fators and the ommodity prie hanges Estimating the net effets of trade liberalization When both the onsumption and the earnings effets are determined the net effet of trade liberalization is easily obtained by subtrating the onsumption effet from the earnings effet to get the net inome effet for eah household (Friedman, 2000). It has to be kept in mind that the results obtained are worst ase senarios beause no quantitative response is assumed neither through substitution effets in onsumption nor through earning generation hanges (MCulloh, 2003). For poverty alleviating poliy measures it is important to know the worst ase to help the poor to manage this transition proess. With the net inome effets alulated it is possible to determine the new extent of poverty and ompare it with the original poverty rates alulated before. 2.3 Data The database used in the maro model is the GTAP dataset, version 6.1 with referene year 2001 (Dimaranan and MDougall, 2003). It onsists of 85 ountries and regions, 57 eonomi setors and 5 prodution fators (skilled labour, unskilled labour, land, apital and natural resoures). An aggregation of regions and setors whih follows the needs of the researh question is neessary and is shown in Table 2: Table 2: Aggregation of regions and setors (12x10) Regions Setors Vietnam (VNM) Paddy rie (PDR) ASEAN Vegetables & fruits (V_F) China, Taiwan, Korea (CHTWKO) Cereals, oils seeds, raw sugar, fib (COSF) Japan (JPN) Other rops (OCR) Australia, New Zealand (AUSNZ) Animals & animal produts (AnimP) United States of Ameria (USA) Primary produts (PrimP) European Union (EU15) Meat (MTP) Sabine Daude 5/12 DTT Otober 2004

6 CEE Countries (CEEC) Proessed rie (PRR) Rest of developing ountries (RDC) Other proessed food (OPRF) Rest of industrialized ountries Textile & loth manufaturing (TCMNFC) (RIC) Other manufaturing (OMNFC) Servies (SVCES) The EU15, CEEC, USA, Japan, Australia & New Zealand, and the rest of industrialized ountries (RIC) are defined as industrialized ountries and the remaining regions are sub summarized under developing ountries. Household data are from 247 Vietnamese households from a survey that was done in two subprojets (F1/F2) of a researh projet of the University of Hohenheim alled The Uplands Program Researh for Sustainable Land Use and Rural Development in Mountainous Regions of Southeast Asia (Dufhues and Wirth, 2002). This program is an interdisiplinary Thai-Vietnamese-German Collaborative Researh Program. Households are loated in several villages in Son La provine (distrit Yen Chau) and Ba Kan provine (distrit Ba Be) in Northern Vietnam. Both regions give a good piture of the situation of poor rural farmers in the mountainous areas of Northern Vietnam. They are onsidered among the group of the poorest distrits of the ountry due to the geographial onditions and the marginalization of ethni minorities within the population. The survey was done in two rounds from November 2001 until April 2002 and relates to the time period Results and their disussion The first part gives the prie results for the maro-simulation before proeeding to the presentation of inome and poverty results for the households of the survey following the different steps of the post-simulation analysis in the seond subhapter. 3.1 Prie effets of maro-simulation When the multilateral trading system ontinues liberalization in the agriultural setor while Vietnam is still no member of the WTO this results in delining pries in Vietnam over all setors as an be seen in Table 3. The magnitude in the agriultural setors is greater (-1% to -3%), whereas the manufaturing and servie setor is only affeted by prie delines of under 1%. The results of simulation 2, when Vietnam is a member of the WTO, reveal prie inreases over all setors. The strongest prie inrease happens in the primary prodution setor (+3.5%) whereas the weakest inrease for the agriultural setor an be observed for the other proessed food (OPRF) setor with an inrease of less than half a perent. Also the manufaturing and servie setors experiene only small prie inreases. Table 3: Prie effets for Vietnam (%) Sim 1 Sim 2 hange Produts/setors AgrLib AgrLib+AVn PDR V_F COSF OCR AnimP Sabine Daude 6/12 DTT Otober 2004

7 PrimP MTP PRR OPRF TCMNFC OMNFC SVCES Soure: GTAP data, own alulations. The prie effets on the labor market are modest (Table 4). In the ase that Vietnam is not in the WTO and agriultural trade liberalization is pursued wages for Vietnamese unskilled and skilled labour go down equally by -0.8%. The downward trend holds even when Vietnam beomes a member of the WTO but is then muh smaller. Wages for unskilled labor nearly stay unhanged with a slight deline of -0.1% and skilled labor suffers from a wage derease of -0.3%. Table 4: Wage effets for skilled and unskilled labor for Vietnam of simulation 1 and 2 (%) Sim 1 Sim 2 hange AgrLib ArgLib+AVn Unskilled labour Skilled labour Soure: GTAP data, own alulations. 3.2 Inome and poverty effets for households in Northern Vietnam To examine the inome and poverty effets of the prie hanges of the trade liberalization senarios the outomes of the maro model are linked with the household post-simulation analysis. The onsumption and earning effets are determined to finally derive the net inome and poverty effets Consumption effet For rural households in Northern Vietnam subsistene agriulture and espeially the ultivation of rie for own-onsumption plays a dominant role. For that reason subsistene onsumption is inluded in the alulation of onsumption shares, evaluated at their loal market value (see also World Bank, 2003). When inluding subsistene onsumption into total onsumption minimum and maximum per apita onsumption are at VND/year (equivalent to 29,12 USD) and VND/year (equivalent to 257,07 USD) respetively with an average of VND/year (equivalent to 94,76 USD). The share of subsistene onsumption an attain up to 92 perent of total onsumption with an average of around 50 perent over all households of the survey. The per apita onsumption teriles and the distribution of onsumption expenditure shares taking into aount subsistene onsumption are illustrated in Table 5. Table 5: Consumption expenditure shares by terile of per apita onsumption inluding subsistene onsumption (%) 1 n=165 2 n=70 3 n=12 avg n=247 FOOD TOTAL 84,0 83,8 83,7 83,9 FOOD total bought 29,4 35,5 39,5 31,6 Sabine Daude 7/12 DTT Otober 2004

8 Husked rie 0,2 0,1 0,0 0,2 Rie 6,2 7,4 5,2 6,5 Maize, assava, sweet potato 0,9 1,0 1,3 1,0 Fruits, vegetables, beans 2,1 2,8 4,1 2,4 Meat, eggs 12,0 15,5 17,4 13,3 Oil, salt, sugar, tea, igarettes, alohol 8,1 8,8 11,5 8,4 FOOD total subsist. 54,5 48,4 44,3 52,3 Husked rie subsistene 40,1 36,0 33,9 38,7 Fruits, vegetables, beans, subsistene 6,8 5,,3 4,5 6,3 Maize, assava, sweet potato, 2,4 0,9 0,0 1,8 subsistene Meat, eggs 5,2 6,2 5,9 5,5 subsistene NON-FOOD TOTAL 16,0 16,2 16,3 16,1 Wood 0,1 0,4 0,6 0,2 Soap, petrol, lamp oil 4,5 4,8 6,8 4,7 Eletriity 0,8 1,4 1,4 1,0 Clothing 10,1 7,3 7,4 9,2 Building maintenane 0,6 2,1 0,1 1,0 TOTAL Soure: Household survey SFB 564 projet F1/F2, own alulations. Rie from self-prodution is the dominant onsumption share with an average of nearly 40 perent of total onsumption expenditures followed by meat and egg onsumption (bought and self-produed) with expenditure shares between 17 and 23 perent. For non-food items lothing expenditures dominate with their relative importane in total onsumption being 10 perent for the poorest teriles and around 7 perent for the upper teriles. Other manufatured produts (soap, petrol, lamp oil, et.) attain a share of around 5 perent with slightly inreasing tendenies over the teriles. Expenditure shares for onsumption goods reveal the importane that relative prie hanges for different goods have: the more important the share of a produt ategory within total onsumption and the bigger the share of bought onsumption goods the more is the household influened by prie hanges of this produt ategory. Assuming that the level-of-living is measured via onsumption expenditures the next step is to hose a poverty line. Households falling with their expenditures below this line are defined as being poor. Three different approahes to poverty lines are used in Vietnam. 1. Poverty lines of the Ministry of Labour, Invalids and Soial Affairs (MOLISA) 2. Poverty lines of the General Statistial Offie (GSO) 3. Poverty lines of the World Bank Taking the per apita onsumption expenditures of the household survey in Son La and Ba Kan Provine and assuming a binary index who defines those households that fall below a poverty line as poor and the others as non-poor the initial extent of poverty Sabine Daude 8/12 DTT Otober 2004

9 before any liberalization in this regions following the different poverty lines an be determined. Taking the ommodity prie hanges from the two GTAP simulations together with the expenditure shares it is possible to alulate the short run hanges in the osts-of-living for the households whih orresponds to the onsumption effet. Results are displayed in Table 6: Table 6: Consumption effet (ost-of-living effet) from trade liberalization assuming onsumption expenditure shares with subsistene onsumption for households in Northern Vietnam (%) Changes in the ost-of-living 1 n=165 2 n=70 3 n=12 avg n=247 AgrLiberalization (Sim 1) 0,4 0,1 0,0 0,3 AgrLib+Aession Vn (Sim 2) -0,8-0,4-0,3-0,7 Soure: Own alulations. In ase that Vietnam does not aede to the WTO (simulation 1) the figures indiate that households in Northern Vietnam would suffer from a slight inrease in their osts-ofliving of up to 0.4 perent. The poorest terile suffers from a bigger inrease in its ostsof-living than the upper teriles. In the ase of Vietnam s aession to the WTO and further multilateral liberalization (simulation 2) osts-of-living for households in Northern Vietnam generally deline by an average of -0.7 perent. The poorest terile benefits from a stronger deline in its osts-of-living ompared to the upper teriles. The reason for the results lays in the alulated GTAP prie hanges and the onsumption shares whih are different for the teriles Earning effet So far the effets of prie hanges on the onsumption side have been analyzed, but the earnings side has been negleted. Trade liberalization does not only affet pries of onsumption goods: agriultural households who are selling their produts are affeted by the same prie hanges through the effet on their earnings. Moreover, pries of prodution fators like wages an hange, and their share in total earnings determines how muh a household is affeted. For rural households agriultural earnings are in most ases the most important inome soure. In order to get a better piture how this agriultural earnings are generated the shares of different rop and livestok ativities are determined. Other earning ategories inlude earnings from non-agriultural own-business, wages and salaries as well as a ategory transfers. Table 7 shows the ontributions of different ategories to total earnings for teriles of per apita onsumption (inluding subsistene onsumption): Table 7: Earning shares by terile of per apita onsumption (%) 1 n=165 2 n=70 3 n=12 avg n=247 AGRIC INCOME 81,4 79,4 73,6 80,4 Rie ultivation 32,4 31,6 31,4 32,1 Veg & fruits ultivation 8,6 9,1 6,9 8,6 Maize, assava, sweet potato ultivation 22,9 18,9 10,4 21,2 Livestok 17,5 19,8 24,9 18,5 OFF-FARM SMALL OWN-BUSINESS 6,3 6,0 6,4 6,2 WAGES & SALARIES 8,6 9,0 20,0 9,3 TRANSFERS, PENSIONS 3,7 5,5 0,1 4,0 TOTAL Sabine Daude 9/12 DTT Otober 2004

10 Soure: Household survey SFB 564 projet F1/F2, own alulations. The figures onfirm that the households are predominantly agriultural households with rie being the most important earning soure. Differenes between onsumption teriles an be observed within the agriultural earnings for ultivation of other ereals and livestok prodution. Wages and salaries are the seond important inome soure followed by inome from off-farm self-employment. Transfers and pensions play a minor role but are nonetheless present, espeially in the lower teriles. In order to relate the prie hanges from the maro-simulations to the earning information from the household survey several assumptions are made how to math the two and then the hange in earnings is alulated (Table 8): Table 8: Earning hange results for households in Northern Vietnam (%) Changes in earnings 1 n=165 2 n=70 3 n=12 avg n=247 AgrLib (Sim 1) -1,8-1,7-1,6-1,8 AgrLib+AVn (Sim 2) 2,3 2,2 1,9 2,2 Soure: Own alulations. The earning hanges for all teriles reveal that the differenes between them are relatively evenly distributed. For simulation 1 earnings deline on average by -1.8 perent with the strongest effet in the lower terile. For simulation 2 earnings inrease by an average of 2.2 perent. Again, the bottom terile shows the biggest effet with an inrease in earnings of 2.3 perent. Again, the results are determined by the GTAP prie hange results and the different earning shares for the teriles from the household data. The results from Table 8 only show the impats on earnings of rural households in Northern Vietnam following multilateral trade liberalization. In order to get a omplete piture of the effets the next part will bring together the onsumption effet and the earnings effet to derive the net inome effet Net inome effet Combining both effets, i.e. a household s onsumption pattern and its earning pattern, we alulate the net inome effet. For simulation 1 the onsumption and earning effets go both in the same negative diretion: osts-of-living go generally up and the earnings go down thus resulting learly in a negative net inome position. In the ase of simulation 2 households are positively affeted in their net inome both by delining osts-of-living as well as by inreased earnings. The net inome effets for the two simulations are shown in Table 9. Table 9: Net inome effets from trade liberalization for households in Northern Vietnam (%) Changes in net inome 1 n=165 2 n=70 3 n=12 avg n=247 AgrLib (Sim 1) -2,2-1,9-1,7-2,1 AgrLib+AVn (Sim 2) 3,1 2,6 2,3 2,9 Soure: Own alulations. In the ase of simulation 1 households in Northern Vietnam suffer from an average net inome derease of around 2 perent due to the same fators that determined the separate onsumption and earning effets. Households in the lower terile are more negatively affeted than the upper teriles. For simulation 2 when Vietnam is in the WTO the results look different: households in Northern Vietnam benefit from net Sabine Daude 10/12 DTT Otober 2004

11 inome inreases with an average inrease of nearly 3 perent. Households in the poorest terile an grasp higher net inome inreases than the upper teriles Poverty effet Having obtained the perentage hanges in net inome for different household teriles in Northern Vietnam following two agriultural liberalization senarios the question remains how this affets the poverty situation in this region. A omparison between the extent of poverty before and after agriultural trade liberalization is given in Table 10. Table 10: Poverty rates before and after agriultural trade liberalization for households in Northern Vietnam (%) MOLISA Rie poverty line, 1997 GSO Food poverty line, 1996 World Bank Food poverty line, 1998 World Bank Total poverty line, 1998 Before liberalization 10,0 % 48,2 % 51,8 % 76,5 % Sim 1 (AgrLib) 10,8 % 49,4 % 52,6 % 76,9 % Sim 2 (AgrLib+AVn) 9,6 % 46,6% 49,0 % 74,9 % Soure: Own alulations. If agriultural liberalization is pushed further within the multilateral trading system and Vietnam is not member of the WTO households in Northern Vietnam are negatively affeted, although moderately. The extent of poverty as measured by the MOLISA rie poverty line and the WB food poverty line inreases by 0.8 perent; the GSO poverty line experienes the highest inrease with 1.2 perent and the WB total poverty line the smallest inrease with 0.4 perent. The rise of poverty by MOLISA s measure an be interpreted that more people will live in absolute poverty and are hronially undernourished. In the ase of the seond simulation when agriultural liberalization is going on and Vietnam is member of the WTO households in Northern Vietnam are positively affeted. Delines in poverty an be observed measured by all poverty lines. Taking the MOLISA rie poverty line poverty goes down by 0.4 perent whereas following the GSO food poverty line it delines by 1.6 perent. The strongest redution in poverty is shown if the food poverty line as defined by the World Bank is applied; poverty shows a redution of 2.8 perent falling under 50 perent to 49.0 perent. For the total poverty line of the World Bank the poverty redution amounts to 1.6 perent ompared to the situation before liberalization and amounts to 74.9 perent. 4 Conlusions and poliy reommendations Developing ountries inluding Vietnam need to pay speial attention to the impat of agriultural trade liberalization on their poor population; agriulture is still the most important inome soure of the poor in developing ountries. To analyse the linkages between trade and poverty is therefore an important task and the outomes of suh analysis are important for national poliies and espeially for targeted poverty redution programmes. If Vietnam is not member of the WTO and agriulture is further liberalized in the multilateral trading system remote areas suh as Son La and Ba Kan in Northern Vietnam are adversely affeted. Due to their onsumption and earning patterns osts-ofliving go up and earnings deline, thus leading to a net inome derease of -2 perent. This leads to an inrease of extremely poor households as measured by a Vietnamese rie poverty line by 0.8 perent. Other poverty lines show an inrease in the extent of poverty between 0.4 and 1.2 perent. In the ase of Vietnam s aession to the WTO pries inrease over all setors, espeially in the agriultural setor. Combining this Sabine Daude 11/12 DTT Otober 2004

12 information with the household data it an be seen that osts-of-living go down due to the high share of subsistene prodution. On the other hand, earnings go up beause households in Northern Vietnam get most of their inome from agriulture. Both ostsof-living and earning developments lead to a raise in net inomes of 3 perent and delines in poverty rates of 0.4 to 2.8 perent. For rural households in Northern Vietnam it is more benefiial if the ountry aedes to the WTO. However, what might be good for them might be less good for other poor groups with different onsumption and earning patterns and a lesser degree of subsistene prodution. Therefore, it is neessary for the Vietnamese government to determine the impat of international poliies on various groups of the population, like the rural and the urban poor. The results of these analyses have to be integrated in national poverty redution poliies. Speial attention has to be given to the role of subsistene prodution and how it is inluded in the analysis as this has a great influene on the final outome. Moreover, the definition and use of different poverty lines an help to determine the poorest parts of the population and poverty redution efforts should start there. The first hannel to redue poverty is through an inrease in earnings and less through a redution in osts-of-living. Therefore the hallenge for the government is to reate better inome opportunities in those setors where the poor have their highest inome share, i.e. in the agriultural setor. This an happen through the reation of funtioning markets to sell agriultural produts. By ameliorating infrastruture or extension servies to produe a higher quality produt whih is espeially relevant for remote areas like the mountainous regions in Northern Vietnam, poor households an be linked to market opportunities reated through liberalization. Moreover, sustainable agriultural systems need to be promoted whih allow for more inome and whih do not harm the environment. In mountainous regions in Northern Vietnam with high rates of erosion and delining soil fertilities this is of speial onern as the environment is the basis for inome generation for the poor. Referene List Brokmeier, Martina A Graphial Exposition of the GTAP Model. GTAP Tehnial Paper No. 8. Purdue, USA: Purdue University, Center for Global Trade Analysis. Dimaranan, B. and MDougall, Robert GTAP data base 6.1. Purdue, USA: Purdue University. Dufhues, Thomas and Wirth, Thomas Joint Questionnaire for Rural Households: subprojet F1 "Land Tenure and Natural Resoure Management in the Upland Areas of Vietnam" and subprojet F2 "Rural Finanial Market Development in Northern Vietnam". SFB 564 The Uplands Program, First Phase Hohenheim, D: University of Hohenheim. Friedman, Jed Differential Impats of Trade Liberalization on Indonesia's Poor and Non-poor. Paper prepared for the Conferene on International Trade and Poverty, Stokholm, Ot 20-21, Hertel, Thomas W Global Trade Analysis - Modeling and Appliations. Cambridge, USA: University Press. MCulloh, Neil The Impat of Strutural Reforms on Poverty: A Simple Methodology with Extensions. Poliy Researh Working Paper Washington D.C., USA: World Bank. Reimer, Jeffrey J Estimating the Poverty Impats of Trade Liberalization. Purdue, USA: Purdue University. World Bank World Development Indiators Washington D.C., USA: World Bank. Sabine Daude 12/12 DTT Otober 2004