COMESA-EAC-SADC Tripartite FTA Impact analysis

1. COMESA-EAC-SADC Tripartite FTA Impact analysis Regional Integration Research Network, TradeMark Southern Africa 30 October 2013

2. Monday, 28 October 2013, at the 8th African Economic Conference ……… • The UNECA Deputy ES said: • Africa is making considerable effort in order to integrate.  The successful establishment of a CFTA in Africa would see intra-African trade rising to 22% of total African Trade and, in line with similar efforts worldwide, would add approximately US$ 1 trillion to the global economy. ……. trade can serve as a powerful engine for job creation, inclusive growth, poverty reduction, and sustainable development. • The AU Commission Chairperson said: • The EAC/COMESA/SADC Tripartite that will see 27 countries uniting into a free trade area, …..represent a positive step in terms of integration.  However, the Tripartite is not moving as fast as expected and we should urge that we speed up the implementation of this agreement. • The President of the African Development Bank said: • We are not here to make a case for regional integration. That case was made long ago. We are here because Africa is at a critical juncture, determined to convert the turning point into the tipping point. ….These are times for strategic decisions on where Africa goes next. ……If we are to deepen integration more quickly, countries will have to rethink the zero-sum calculus that says "You will gain/I will lose" which is a sort of prisoners’ dilemma as to what the other side will do if I am the first to make a move. Evidence is ample that along the way it is a win-win for each and every one and, as we can see from the euro zone, variable geometry comes with costs Africa cannot afford.

3. Outline • TMSA’s approach to poverty • Measuring its impact on poverty • Globally & on country level others are measuring it, e.g. WB, AfDB, countries …; • TMSA is measuring TFTA impacts through CGE modelling: • 1st round – aggregate gains & distribution between countries • 2nd round – macro-micro simulations • TMSA specific intervention measurement, e.g. surveys of small-scale traders; border posts, NTBs, case stories

4. How does TMSA focus on poverty? The high-level story(1) • TMSA & other similar regional integration programmes typically describes the long route from programme outputs to growth and poverty, i.e. how programme components are stimulating trade performance and /or reducing trade transaction costs in the region. Both are positively associated with increased economic growth. Economic growth, in turn, is crucial for poverty reduction. It affects the lives of the poor through 3 primary channels. • increases in GDP create employment which expands the number of people able to directly benefit from growth; • growth stimulates other areas of the economy through firms spending on inputs and employees spending wages; • growth increases the resources available to governments for spending on poverty reduction through increasing tax revenue. Growth has the strongest effect on poverty when all of these channels operate (labour intensive growth with proceeds spent throughout the economy which generates high tax receipts used for welfare spending)

5. How does TMSA focus on poverty? The high-level regional agenda (2) • The time between action and impact is long and the pathways are characterised by programmes, projects and activities which are far removed from working directly with households and firms. • Programmes are about • coordination / cooperation / harmonisation / reforms of policies and strategies between RECs and countries; • On specific programme areas, e.g. regional infrastructure , the focus is on networks rather than infrastructure services for poor people per se (e.g. in the area of power, it is about regional power markets and power interconnectors rather than electricity distribution services; and in the area of transport, it is about regional transport corridors rather than feeder / urban roads) Despite the regional (supra-national) nature of the TMSA programme, a deliberate effort has been made to build a pro-poor bias into programmes / projects / activities …….

6. …. mainstreaming of poverty in TMSA programmes: focus on poorest countries (3) 17 LDCs in the Tripartite Region – more than a third of world’s LDCs • the WTO LDC Group - negotiating DFQF market access and an LDC package in the Doha round. A World Bank paper (2011) Opening Markets for the Poorest Countries: Assessing the Effects of Duty-free Quota-free Access to the G20 estimated that full DFQF access could increase national incomes in LDCs by some 0.5% of GDP on average; a G20 DFQF access initiative could lift 3 million people in LDCs above the poverty line; and if DFQF is complemented by actions to reduce trade costs for LDCs by just 2%, average income gains could rise by another 0.6 percent of GDP. • LDCs in the Tripartite Region e.g.in terms of EIF proposals (Lesotho and Uganda examples) and DTIS (Mozambique, Malawi, Zambia). TMSA’s LDC workstream is complimentary to its regional work e.g. TF assessments; harmonising implementation of TF programmes requested under the DTISs and Tier 2 programmes with the Tripartite trade facilitation programme.

7. …. mainstreaming of poverty in TMSA programmes: focus on landlocked countries (4) • A WB paper on the cost of being landlocked (WP4258, June 2007) • found that a large proportion of the least developed countries are landlocked and their access to world markets depends on the availability of a trade corridor and transit systems. • proposes a microeconomic quantitative description of logistics costs, based on empirical evidence from WB projects and assessments in Africa, Central Asia and elsewhere • theoretically and empirically highlights that landlocked economies are primarily affected not only by a high cost of freight services, but also by the high degree of unpredictability in transportation time. The main sources of costs are not only physical constraints but widespread rent activities and severe flaws in the implementation of the transit systems, which prevent the emergence of reliable logistics services. • TMSA is not only focusing on landlocked countries through its support for the well-known NSC Aid-for-Trade pilot programme (Botswana, Zimbabwe, Zambia, Malawi and Eastern part of DRC which are all land-locked), but is also promoting the design and implementation of comprehensive, robust and resilient transport and transit regimes. [One country, Zambia is deriving particular direct and indirect benefits from TMSA programmes: it is an LDC (WTO LDC Group and EIF programmes) and it is landlocked with eight adjacent neighbours, deriving benefit from TMSA’s trade facilitation and NSC activities.

8. …. mainstreaming of poverty in TMSA programmes: focus on people / firms (5) • 6 legacy Standards projects • introducing modalities into programmes that make it accessible to the poor, e.g. the sms (mobile phone text) facility in the NTB identification, notification and resolution system to supplement the use of computers, faxes and landlines

9. Measuring economic and poverty impacts • Tools & techniques • An Integrated framework • The micro side • The macro side • The meso level

10. A common framework to integrate/capturemacro-micro links • Trade-offs/links between stabilization, growth, and poverty alleviation • Impact of macro and growth policies • Changes in inflation and macro balances • Structural reforms • Exogenous shocks • Investment and employment generation • Impact of redistribution policies on poverty (and inequality) • Public expenditures • Taxation • Income and cash transfers/subsidies to households • Delivery of public services

11. Tools to evaluate impact of economic policies on poverty – micro side • Traditional benefit incidence analysis (BIR) of public expenditure and taxation • Recent extensions to improve public service delivery: • Marginal vs. average benefit incidence analysis • Geographical mapping of poverty and expenditure • Institutional aspects • public financial management at different governmental levels • inter-governmental fiscal dimensions • Tracking delivery of public services with quantitative survey methods • clients, service providers, facilities • How much of budget allocation filters down? • Basis: detailed household and poverty data

12. Tools to evaluate the distributional impact of macro policies – macro side • Traditional macro approaches • Cross-country analysis of growth and distribution; growth and poverty etc. • Country-specific macro-econometric and CGE models • Recent approaches • Policy determinants of growth from new growth theory • links between public expenditures via human capital (education, health etc.) to growth • New cross-country growth regressions • Disaggregated macro-econometric (IS-LM) models • Macro-augmented CGE models • Linking household data or micro simulations to CGE or macro models • Simple approaches – 123PRSP model • CGE or macro models + Micro simulations • Econometric model of occupational choices of income generation by households

13. Ex-ante Tools

14. Simplified Structure of a CGE Model

15. 1st round modelling of the TFTA: GLOBE Model • A ‘standard’global CGE model that has been calibrated using data derived from the GTAP database (v8.1), informed by the SAM approach to national accounting and the SAM approach to modelling. • The underlying approach to multi-region modelling is the construction of a series of single country CGE models that are linked through their trading relationships. Consequently each region in the model has its own numéraire price and a nominal exchange rate, while the model as a whole requires a numéraire. This is a fundamentally different philosophical approach to global modelling than in the GTAP model. • Trade is modelled using nested Constant Elasticity of Substitution – CES - (imports) and Constant Elasticity of Transformation - CET - (exports) functions with an adaptation to mitigate the impacts of small trade shares. Production is modelled using nested CES functions. The tax instruments are those in GTAP plus direct (income) taxes on the households who also save. • A distinctive feature of the model is the use of a 'dummy' region, known as Globe, that allows for the recording of inter regional transactions where either the source or destination are not identified. The Globe construct provides a general method for dealing with any transactions data where full bilateral information are missing.

16. 1st round modelling of the TFTA: GLOBE Model Features • Production, input demand & factor markets – standard neoclassical factor market closure with market-clearing factor prices • Final domestic demand by commmodity • International trade – Constant elasticity of substitution – imports & constant elasticity of transformation - exports • Macro closure – neutral / balanced set of macro closure rules • Model calibration to GTAP 8.1 data base; benchmark year 2007; production, trade & income elasticities drawn from GTAP behavioural data base • 19 countries and or country groups (15 of 26 Tripartite countries are separate) & 3 composite non-TFTA regions, namely Other Sub-Saharan Africa, the European Union, and the “Rest of the World”. • 22 commodity / sector groups including 5 agricultural sectors, 3 natural resource extraction sectors, 3 food-processing sectors, 8 non-food manufacturing sectors and 3 service categories

17. 1st round modelling of the TFTA: GLOBE Model Simulation Scenarios 8 Simulation Scenarios differing in the assumed level of ambition in terms of regional coverage, product coverage and trade facilitation effort: • S1: Elimination of remaining intra-COMESA and intra-SADC baseline tariffs • S2: Elimination of all intra-TFTA tariffs • S3: Elimination of intra-TFTA tariffs without participation of Angola, DR Congo and Ethiopia • S4: Elimination of intra-TFTA tariffs except tariffs on fossil fuels and sugar products • S5: Elimination of intra-TFTA tariffs without participation of Angola, DR Congo and Ethiopia, and except tariffs on fossil fuels and sugar products (Combination of S3 and S4:exclusions) • S6: Full liberalisation of capital goods, 80% tariff cuts on intermediate goods, 50% tariff cut on consumption goods • S7: Full liberalisation of non-sensitive commodity groups, partial (50%) liberalisation of “revealed” sensitive goods, i.e. goods with high (10% plus) tariff rates in 2007. • S8: Elimination of all intra-TFTA tariffs S2 and real transport / transaction cost reduction on intra-TFTA flows.

18. 1st round modelling of the TFTA: GLOBE Model Results: Welfare Impacts (US$ million)

19. 1st round modelling of the TFTA: GLOBE Model Results: Welfare Impacts (% change from baseline)

20. 1st round modelling of the TFTA: GLOBE Model Results: Welfare Impacts of S8 (%)

21. 1st round modelling of the TFTA: GLOBE Model Results: overall results • All 8 trade liberalization scenarios lead to positive net real income gains for the TFTA area as a whole. • S1 - removal of remaining tariff barriers to intra-COMESA and intra-SADC trade by 2014 in the absence of a TFTA agreement generates an estimated aggregate annual gain for the TFTA group of US$328m or 0.04% of TFTA 2014 baseline final demand for goods and services. • S2 – TFTA with full elimination of all tariffs on trade among all 26 potential partners will generate an annual welfare gain of US$578m or roughly 0.1% of total TFTA area 2014 baseline absorption. • If complete tariff liberalization within COMESA and SADC without any remaining exceptions for sensitive products will be achieved by 2014 prior to the implementation of TFTA, the additional welfare gain genuinely attributable to TFTA tariff liberalization among the three RECs is around US$250m p.a. for the TFTA group as a whole. • In absolute terms, South Africa enjoys the largest real income gains under full intra-FTA tariff liberalization whereas the largest gains relative to baseline absorption are projected for “Other SACU” (i.e. Swaziland and Lesotho) (+0.8%) and Namibia (+0.4%). • Zimbabwe and to a lesser extent Malawi, Zambia, Rwanda, South Central Africa (Angola and DR Congo), Botswana and Other East Africa suffer moderate welfare losses under this scenario as result of a terms-of trade deterioration that dominates the gains from lower consumer prices for TFTA imports.

22. 1st round modelling of the TFTA: GLOBE Model Results: overall results • S3: If Ethiopia, Angola and DR Congo choose not to participate in the TFTA, aggregate net welfare gain for the area as a whole drops by around US$260m versus the full participation scenario S2. The simulation results suggest that participation in the free trade agreement would be in Ethiopia’s own interest. • S4: Exclusion of fossil fuels and sugar products as sensitive products from tariff liberalization would reduce the total welfare gain for the TFTA group by roughly US$130m per annum compared to S2.

23. 1st round modelling of the TFTA: GLOBE Model Results: overall results • S5: Elimination of intra-TFTA tariffs without participation of Angola, DR Congo and Ethiopia, and except tariffs on fossil fuels and sugar products (Combination of S3 and S4:exclusions), the aggregate net welfare gain for the area as a whole drops by around US$ 338.5 million compared to the full participation scenario S2. Baseline tariffs on intra-TFTA fossil fuel trade are already generally moderate, while sugar products are sensitive products for a range of TFTA partners. Kenya, Uganda, Egypt and Other East Africa impose the highest average applied tariff rates on TFTA sugar imports; whereas Mozambique, OSACU, Ethiopia and SA face on average the highest TFTA import duties on their sugar product exports. Fossil fuels and sugar account for 13.1% and 1.6 % of total intra-TFTA baseline trade of goods and services and under full TFTA tariff liberalization (S2) the two product groups contribute 17% (around US$ 440 million) to the projected total increase in intra-TFTA trade volumes. In the S4 scenario the trade expansion for the two commodity groups is close to zero

24. 1st round modelling of the TFTA: GLOBE Model Results: overall results • S6: Partial tariff liberalization, which assumes full liberalisation of capital goods only, 80% tariff cuts on intermediate goods and 50% tariff cut on consumption goods, reduces the net aggregate welfare gain for the TFTA group by nearly US$150m and the increase in aggregate intra-TFTA trade flows is US$821m lower than under S2. • S7: In the least ambitious tariff liberalization scenario under consideration, only baseline tariffs with an ad valorem rate of up to 10% are removed completely, whereas tariffs with a higher rate are cut by 50% - the aggregate net welfare gain for the TFTA group projected by the model is a meagre 0.04% of baseline absorption.

25. 1st round modelling of the TFTA: GLOBE Model Results: overall results • S8: most ambitious TFTA scenario, combining complete tariff liberalization for intra-TFTA trade with a reduction in costs of border-crossing trade within the TFTA are equivalent to a 5 percentage points ad valorem tariff. • Aggregate net benefit for the TFTA group amounts to over US$3.3 billion p.a. or nearly 0.4% of aggregate baseline absorption and more than 5 times the gains resulting from full intra-TFTA tariff liberalization alone. • In contrast to the S2 scenario all TFTA regions enjoy a positive aggregate welfare gain in this case. The countries with the largest projected percentage increases in real absorption are Zimbabwe (+2.6%), Namibia (+2.4%), Mozambique (+2.2%), Botswana (+1.8%) and Other SACU (+1.5%). • Total volume of intra-TFTA trade is boosted by US$7.7 billion, an increase of nearly 20% relative to the 2014 baseline volume. • The simulation results do not suggest that TFTA leads to systematic increase in wage inequality. • Significant sectoral production effects with corresponding significant implications for sectoral employment are concentrated in a sub-set of sectors including primarily sugar products with backward linkage effects to sugar cane production, beverages and tobacco and light manufacturing, and to a lesser extent for some TFTA countries in textiles, metals and metal production, and chemicals.

26. Next Steps • Proceed with 2nd round modelling of macro-micro simulations with WB team & wide reference group in the region – process has already started • Build on the 1st round model, also policy scenarios, e.g. tariff changes, NTMs, transport / border costs, services liberalisation

27. 2nd Round Modelling: Macro-Micro Linkages to determine impact of TFTA? • Macro • Policies • Shocks Meso Disaggregated - more sectors, - factors of production CGE models? Linkage model (WB) Globe model (IDS) • Micro • household data • Micro simulation • Incidence analysis • Service delivery GIDD (WB)

28. 2nd Round Modelling: Macro-Micro Linkages to determine impact of TFTA?

29. Macro-Micro ModellingIssues • The difficulty of full integration • Top-down causality: macro to micro • Remaining problems • Introducing dynamics • Determinants of long-term growth • Firm and investment climate • Data limitations • Trade-offs: • Simplicity vs sophistication & more content vs black boxes • In any case ……..models are : • Tools not substitutes for policy • There is no single tool for all issues • Effective if they clarify policy by providing substance to the policy discussion Simulation models like these have to be understood as what-if scenarios, not forecasts

30. TMSA specific intervention measurement • Surveys of small-scale traders • Border post efficiency monitoring • NTBs and system use • Case stories Visit: www.trademarksa.org