Is Free Trade Always Best

1
Is Free Trade Always Best?

• Edward F. Buffie
• FAO workshop on Staple Food Trade and Market
  Policy Options
• for Promoting Development
  in Eastern and Southern Africa
• March 1-2, 2007

2
Ground Rules

• What is the underlying distortion or market
  imperfection?
• Does protection/export promotion counteract the
  distortion?
• Does the argument for protection have a
  satisfactory rejoinder to the first-best solution
  recommended by the Principle of Targeting?
• Complicated tax-subsidy schemes are impractical.
• The welfare loss from certain byproduct
  distortions, such as the distortion of consumer
  choice, is quite small (a little triangle).
• Once administrative costs are thrown into the
  mix, protection can compete for the honor of the
  first-best policy (Buffie, 2001).

3
Food Security/Poverty Reduction

• Arguments for protection based on concerns about
  food security are suspect.
• Since most food products are tradable on world
  markets, food security depends mainly on
  purchasing power, not management of domestic
  supply. Food insecurity is a byproduct of
  poverty.
• The debate usually centers around the impact of
  policy on real income of the poor. Is the notion
  of food security excess conceptual baggage?
• The first-best remedy of fighting poverty with
  direct transfers to the poor has worked in some
  countries.
• Decoupled support payments to farmers hurt by
  NAFTA in Mexicos PROCAMPO program.
• Experience in SSA? Large-scale transfer schemes
  administratively impractical, vulnerable to
  corruption? (Direct cash transfers to 50 of the
  labor force?)

4
Food Security/Poverty Reduction (cont.)

• Assuming large-scale direct transfers are
  infeasible, the critical issue is how protection
  affects poverty.
• Typical analysis in the literature is
  inconclusive and unsatisfying.
• Protection would seem to help the poor because
  poverty is overwhelmingly concentrated in
  agriculture.
• But higher food prices do little to help poor
  farmers, who often produce mainly for own
  consumption. In addition, they hurt landless
  laborers and the urban poor.
• Some analyses stop here.
• Better analyses recognize that unskilled labor in
  agriculture may benefit from higher wages when
  food prices increase. But when the impact on
  other sectors is taken into account, it is not
  clear whether protection of agriculture raises
  overall demand for unskilled labor in the economy
  (McCalla and Nash, 2007, Reforming Agricultural
  Trade for Developing Countries).

5
Food Security/Poverty Reduction (cont.)

• Knowledge Gap
• Need more and better analysis of how protecting
  agriculture affects the poor.
• Static models push the line that growth in
  agriculture does more to reduce poverty than
  growth in other sectors (e.g., Foster and Valdez,
  2003). Reason agriculture is highly
  labor-intensive.
• With respect to management of the supply chain,
  it is often asserted that protection of food
  crops worsens poverty by attracting resources
  away from more labor-intensive food processing
  and nontraditional export industries (Roberts,
  Buetre, and Jotzo, 2007 McCalla and Nash, 2007).
• Not clear that the data support this claim.
• Dearth of analysis based on dynamic models.
• Dynamics in GTAP, Linkage, and other large CGE
  models either import assumptions about how factor
  supplies and productivity grow or rely on ad hoc
  specifications of saving and investment. Not
  grounded in good theory.
• First, most of the issues that arise in the
  popular debate over the impacts of trade policy
  are fundamentally comparative static in nature.
  (Hertel and Winters, p.7, Poverty and the WTO
  Impacts of the Doha Development Agenda.)

6
Food Security/Poverty Reduction (cont.)

• If the policy debate is to be better informed, we
  need input from medium-sized dynamic trade models
  in which private
• agents solve intertemporal optimization problems.
• First thoughts on the general structure of the
  dynamic trade model.
• 4 sectors industry, primary agriculture,
  food-processing/agro-industry, services/nontradabl
  es.
• 5 factors of production capital, skilled labor,
  unskilled labor, land, and infrastructure.
  Stocks of land and infrastructure treated are
  exogenous.
• Heterogeneous agents. Capitalists invest only in
  the capital in their own sector.
• Nature of the labor market? Sectoral wage gaps?
  Degree of intersectoral labor mobility over
  different time horizons?

7
Learning Effects and Infant Industry Arguments

• The classic infant industry argument (IIA)
  asserts that firms become more efficient as they
  become
• more experienced.
• During the infancy/learning phase, firms suffer
  losses when competing at free trade prices.
• After firms grow up and join the adult world,
  they can compete profitably without protection.
• Mill-Bastable Criterion If the present value
  gains reaped after the industry is mature exceed
  the present value losses incurred during the
  infancy phase, then the IIA is potentially valid.
• Underlying imperfection in financial markets
  For some reason, firms cannot borrow to cover
  temporary losses during the learning phase.
• Quite plausible. Banks face severe moral hazard
  and adverse selection problems. Many borrowers
  lack marketable collateral.
• Are banks excessively risk averse in SSA?

8
Learning Effects and Infant Industry Arguments
(cont.)

• Is There a Second-Best Argument for Protection?
• Is protection justified if the laissez-faire
  equilibrium is not constrained Pareto-efficient
  and the government cannot lend
• to the targeted firms? Not necessarily.
• Precommitment to temporary protection may be
  critical.
• -- Despite brave announcements to the
  contrary, the government may be reluctant to
  remove protection for
• firms that cannot compete at free trade
  prices.
• -- Learning is probably not deterministic.
  If firms expend real resources to learn and
  suspect that the
• government is soft, then there is an
  incentive to balance the gains from learning
  against the costs of losing
• protection (Matsuyama, 1990 Tornell,
  1991).

9
Learning Effects and Infant Industry Arguments
(cont.)

• Is There a Second-Best Argument for Protection?
  (cont.)
• If learning effects are external to the firm,
  protection may be justified even when financial
  markets operate perfectly.
• The external effect gives rise to multiple
  equilibria and a coordination problem (Krugman,
  1991). The coordination problem can be resolved
  by imposing a tariff that makes it profitable for
  each firm to produce even when all other firms do
  not.
• Imperfect information about the potential
  viability of the industry may be the source of
  the externality (Hoff, 1997).
• -- Initial entrants provide information
  through the success/failure of their venture that
  reduces uncertainty for
• future entrants. (Note Initial entrants
  cannot appropriate the social value of the
  information they generate.
• Once it becomes known that success is
  possible, new firms enter rapidly and compete
  away rents by
• bidding up wages.)
• -- The social value of information that
  reduces uncertainty is greatest in the poorest
  LDCs where the costs of
• bearing risk are high.

10
Learning Effects and Infant Industry Arguments
(cont.)

• Empirical Evidence on Learning Effects
• Is there any evidence of learning effects that
  might justify temporary protection or temporary
  export
• promotion?
• Mixed, mainly anecdotal evidence of learning in
  import-competing industries.
• Most of the recent literature has concentrated on
  testing for learning in export industries. Key
  issue Does the greater productivity of
  exporters reflect self-selection or learning by
  exporting (LBE)?
• Majority of studies do not find evidence of LBE.
• But there is support for LBE in a minority of
  studies, including those done for SSA.
• Only two studies (Clerides, Lach, and Tybout,
  1998 Alvarez and Lopez, 2006) investigate
  learning externalities.

11
Raising Revenue for Public Sector Investment

• Given the limited number of tax handles, the
  relatively low administrative costs of trade
  taxes, and the high social return to public
  investment in social and physical infrastructure,
  there is a sound case for including tariffs in
  the set of optimal taxes levied to finance
  government expenditure.
• Aid-for-trade can, of course, lessen the need to
  use tariffs as a revenue-raising device.

12
Factor Market Distortions

• Imperfections that cause underemployment and
  misallocation of labor across sectors
• Unions, payroll taxes, and minimum wage laws that
  make labor more expensive in the formal sector
  than in other sectors.
• Efficiency wages in agro-industry and other
  branches of the formal sector.
• Wage premium need not reflect a compensating
  differential for working harder in the formal
  sector. The nature of the monitoring technology
  matters. (Why isnt the wage gap eliminated by
  competition from lower-paid workers in the
  informal sector? Answer the promise to work
  hard at a lower wage is not credible.)
• Absence of well-defined land tenure conflates the
  return to work in agriculture with rents from
  land. This impedes intersectoral labor mobility
  -- if you move away, you lose your claim to the
  land and the rents in confers.

13
Factor Market Distortions (cont.)

• Factors that cause under-investment
• Share tenancy
• Communally-owned land
• Insecure, unclear property rights
• Taxation of profits
• Lack of credit for smallholders
• Sens isolation paradox . . . the market
  cannot express the collective demand for
  investment to benefit the future. (Feldstein,
  1964)

14
Factor Market Distortions (cont.)

• Implications for Policy
• Principle of Targeting Attack the distortions
  directly with suitable employment and investment
  subsidies financed by taxes that do not interfere
  with free trade.
• Labor market imperfections favor protection of
  manufacturing and agro-industry.
• Agro-industry generally pays higher wages than
  traditional agriculture (World Bank, 2006,
  Agricultural Growth for the Poor).
• Abundant casual evidence that exporters pay
  higher wages to both skilled and unskilled labor
  than non-exporters.
• Market imperfections that cause underinvestment
  operate to varying degrees in all sectors.

15
Factor Market Distortions (cont.)

• In combination, underemployment and
  underinvestment argue for an escalated structure
  of protection and export promotion that lowers
  the real price of imported capital goods and
  intermediate inputs.
• Support for this conjecture in the solutions for
  optimal trade policy presented in Buffie (2001).
• Variance in protection/export promotion across
  sectors (primary agriculture vs. agro-industry
  vs. other manufacturing) likely to be less when
  distributional objectives are taken into account.
  See Buffie (2001).
• Need optimal policy solutions or at least a sense
  of the tradeoffs that accompany different policy
  packages in dynamic trade models with a more
  detailed representation of the agricultural
  sector.
• An escalated structure of protection plus export
  promotion does not necessarily imply that the
  economy is less open or that there is less trade
  than under free trade. The policy package alters
  the composition of trade fewer imports of
  consumer goods, but more exports and more imports
  of intermediate inputs and capital goods.

16
A Customs Union for Agriculture?

• Analysis of the welfare outcome usually focuses
  on three effects
• Trade creation vs. trade diversion (Viner).
• More competition, higher output per firm, and
  lower unit costs in industries where increasing
  returns to scale exist.
• Distribution of gains and losses within the
  customs union (CU).
• If the CU improves allocative efficiency, will
  intra-union transfer payments be needed to ensure
  that all members benefit?
• In SSA another effect is important
• Impact on government revenue and the ability to
  finance high-return investments in social and
  physical infrastructure.

17
A Customs Union for Agriculture? (cont.)

• Trade Creation vs. Trade Diversion
• Textbook treatments are deceiving in that,
  following Viner, they assume the countrys trade
  partners in the union operate with
  infinitely-elastic supply curves.
• This ensures that intra-union imports completely
  displace imports from outside the union. Hence
  the CU generates some trade creation that might
  offset trade diversion.
• But with positively-sloped supply curves,
  intra-union imports may not completely displace
  imports from the outside.
• -- If they do not, then the CU reduces
  efficiency. It results only in trade diversion.
• Zero trade creation is a serious concern in the
  African context.
• Land is an important quasi-fixed factor. Given
  this, the shortage of credit, and numerous severe
  infrastructural bottlenecks, is it naïve to hope
  for an elastic supply response?
• Volume of intra-regional trade is generally very
  low. Is this true for the most important
  agricultural products?

18
A Customs Union for Agriculture? (cont.)

• Intra-Regional Distribution of Gains and Losses
• Distribution of intra-regional trade is very
  lopsided.
• Kenya and S. Africa are the only countries that
  export a substantial amount to the region.
  Neither country, however, imports much from the
  region.
• Is this also true of intra-regional trade in
  agriculture? (Maybe not. Isnt S. Africas
  agricultural sector relatively small?)
• If it is, then Kenya and S. Africa, the two most
  developed countries in the region, will reap
  gains as
• beneficiaries of trade diversion while
  everyone else suffers losses. In short, the CU
  operates like a
• scheme where everyone hands over their
  tariff revenue to Kenya and S. Africa.

19
A Customs Union for Agriculture? (cont.)

• Impact on Trade Tax Revenue
• Trade taxes constitute a significant share of
  total revenue in many countries.
• Limited replacement ( 33) of lost trade tax
  revenue in low-income LDCs (Baungaard and Keen,
  2005).
• Public investment in infrastructure is likely to
  bear the brunt of adjustment to the revenue loss.
  The welfare costs of this are sure to exceed any
  small triangle welfare gains that might
  materialize should the benefits of trade creation
  exceed the losses from trade diversion.

20
A Customs Union for Agriculture? (cont.)

• Bottom Line
• Maybe it is best to go slow on regional
  integration until major supply-side constraints
  have been alleviated and regional
• trade is a greater proportion of total trade.
• Enhances the likelihood that the CU will generate
  some trade creation.
• Reduces the welfare costs associated with lower
  revenue from trade taxes.