poverty traps

by user








poverty traps
The impact of the 2008-09 global crisis on commodity-dependent
low-income countries: confirming the relevance of the concept of
poverty trap?
Alice Sindzingre
National Centre for Scientific Research (CNRS, Paris)-University Paris-10;
School of Oriental and African Studies (SOAS), University of London,
department of economics; Institute of Political Studies (Sciences Po, Paris).
Development Research Seminar
Institute of Social Studies (ISS), The Hague
14 June 2010
1. Context, questions
Many low-income countries = commodity-exporters, and many
commodity-exporters = commodity-dependent countries=> impact of the
crisis on countries exhibiting this specific market structure?
= the most vulnerable to price volatility and fluctuations of global demand for
their exports - fuels, mineral or agricultural products.
This volatility is an intrinsic feature of commodities: amplified during the
2003-08 commodity boom by the increasing linkages between commodities and
financial markets: commodities traded as financial assets.
Back to theories of the secular decline of commodity prices (R Prebisch, H
Singer, A Maizels): this volatility ensnares low-income countries and
explains slow growth.
+ concept of the poverty trap: feedback processes, increasing returns,
spillovers, multiple equilibria, irreversibility, threshold effects...
=> difficult for low-income countries to reach the ‘tipping point’ above
which they can trigger long-run growth .
Poverty traps reinforce themselves through endogenous processes: low
productivity, low value-added , export of commodities, representing a decreasing
share of unit of GDP due to technological progress (World Bank , 2009a)
But for other studies, different conceptual frameworks: e.g., growth
accelerations-decelerations, no ‘traps’ for commodity-dependent countries
+ industrial and emerging countries need certain commodities as inputs for their
industries => demand for these commodities, and therefore prices, follows
cycles can be investigated without the concept of trap.
The paper: a theoretical analysis of the concept of the poverty trap
Focus on low-income commodity-dependent SSA countries
Argument: these countries growth trajectories confirm the relevance of the
conceptual framework of poverty traps.
The 2008-09 crisis, another commodity boom, then slump (then boom?), the
most severe recession in 50 years =example of the processes underlying the
concept of ‘trap’.
1. Key theoretical features of the concept of poverty traps
2. Market structures of commodity dependence and price volatility have
long been viewed as key factors of poverty traps, + deepening of the linkages
between commodity and financial markets over 2003-08.
3. Theoretical critiques regarding the very existence of poverty traps:
competing explanations of poor countries’ growth profiles.
4. Against these critiques: explanatory power of the concept of poverty
traps; commodity-dependent countries’ growth trajectories exhibit the
definitional features of traps: low equilibria, diverging paths relatively to
other groups of countries, threshold and lock-in effects.
5. The assessment of causalities does not mean determinism: other factors
combine with the relationship between commodity-dependence and growth,
which counter or reinforce the formation of traps: in particular, domestic
2. What is a poverty trap? The concept’s main features
Low equilibria, cumulative causation, lock-in processes
Poverty trap: irreversibility, lock-in devices, multiple equilibria (high and low),
threshold effects, non-linearity, increasing returns.
Multiple equilibria or ‘traps’: B Arthur (1989, 1994a): positive feedbacks, path
dependence, “lock-in by historical small events”, self-reinforcing mechanisms,
cumulative causation, some equilibria able to lock in economies or individuals in
inefficient behaviour and low levels of income.
Increasing returns in growth, e.g., A Young (1930s), N Kaldor (1950s).
B Arthur: dynamic nature of increasing returns and positive feedbacks, their
stochastic character = random deviations from long-run tendencies
This property => multiple long-run states depending on initial conditions and random
fluctuations, and of ‘specialised’ outcomes (e.g. in geographical terms).
Individual level => learning, experience and the perception of success may lead to
the reinforcement of some processes, e.g., the transmission of some information at
the expense of others: such processes lock individuals in inefficient behaviour.
=> even with suitable initial conditions, the same mechanisms can lead to either
optimal or inefficient equilibria.
Lock-in (e.g. by technological choices) and positive feedback => path
dependence (David, 1985, 2000) = phenomena that have the dynamic property of
non-ergodicity in stochastic processes (i.e. not having the “ability eventually to
shake free from the influence of their past states”), and which, beyond market
failures and inefficiencies, imply the existence of “winners and losers”.
P David (2000) :‘lock-in’ = the “entry of a system into a trapping region”
= the basin of attraction that surrounds a locally (or globally) stable and selfsustaining equilibrium.
A dynamic system that enters into such regions needs, in order to escape from it,
external forces that alter its structure.
Locked-in equilibria may be optimal or detrimental
P David: whatever the equilibrium, individuals are happy doing something,
“even though they would be happier doing something else if everybody would
also do that other thing too”, because incomplete information prevent them from
coordinating and moving elsewhere collectively.
Alternatives paths are possible: path dependence does not mean determinism.
The recognition of traps in growth theories
Some growth models: when jointly considering income and growth rate,
non-linearity, multiple equilibria=salient feature of the growth process
(Fiaschi-Lavezzi, 2003)
Kelly (2001): Schumpeterian growth models: development =progress
through a space of commodities, from simple to complex goods (linkage
formation) =>thresholds: below a critical probability of linkage formation,
development ceases; above, innovation continues.
Concepts of polarisation and ‘club convergence’ (Azariadis, 2006): against
assumptions of growth convergence across countries to similar steady-state
income levels, and explanations of variations in income growth by different
initial conditions. History (past events have large and lasting effects); nonlinearity and lock-in constraints on the growth of certain countries.
Global inequality: different growth patterns: ‘hills’, ‘plateaux’ (Pritchett,
1997; 2000), multiple equilibria explaining the income gap between rich and
poor countries (Graham and Temple, 2006).
Figure 1: relative GDP levels vs. growth rates
Source: Fiaschi and Lavezzi (2003), based on the Penn World Table 5.6 for the 1960-1989 period for 120 countries. GDP =
relative income with respect to the (world) average of the period.
Spillovers and coordination failures as causes of
underdevelopment traps
WWII theorists : G Myrdal, A Hirschman, P Rosenstein-Rodan (1943): why some
economies are unable to trigger the virtuous process of catching-up?
=> spillover effects, linkages and complementarities: cf. cumulative causation
and path dependency.
Rosenstein-Rodan: spillovers = increasing returns to an activity proportional to
the number of others who undertake the same activity. Lack of spillovers,
coordination failures =>traps.
Low equilibria, coordination failures and poverty traps are endogenous and
=> markets alone cannot achieve the coordination necessary for triggering
development (Adelman, 2000; 2001). Markets do not necessarily lead from the
lowest equilibrium to the best one (Hoff, 2000).
=> hence at the early stages of development, the state= the entity most able to
reallocate factors and resources across markets=> ‘big push’ (Murphy et al.,
State capacity is endogenous to the level of development (Bardhan and Udry,
1999) => underdevelopment traps are likely at early stages of development,
their determinants being economic, political, institutional.
Barrett and Swallow (2006): standard growth models = a single dynamic equilibrium
=> convergence of growth paths toward a single level of welfare
Figure 2: Welfare dynamics under the convergence hypothesis.
Source: Barrett and Swallow (2006).
Multiple dynamic equilibria => S-shape of the growth function, with stable dynamic
equilibria at high and low levels of welfare (Wh, Wl), => at least one unstable dynamic
equilibrium, a critical threshold (Wc). Only a large positive shock make economies or
households able to escape the basin of attraction of the low-level equilibrium , move
toward a higher equilibrium.
Figure 3: Welfare dynamics under the poverty traps hypothesis.
Source: Barrett and Swallow (2006)
3. A key determinant of poverty traps in low-income
countries: commodity dependence
Low-income countries as commodity-dependent
Most low-income developing countries: dependence on commodities for their
exports and lack of economic diversification.
E.g., in SSA, since the mid-2000s, fuels represent more than half of exports
(IMF, 2007, table 4.1).
Commodity dependent countries: undiversified export structure.
Table 1: Commodity dependence by geographical region, 1995–1998; 2003–2006
(nb. countries where exports of commodities = more than 50% of total exports)
Source: UNCTAD (2008b), table 2.4.
 Vulnerability of export-oriented countries to external demand:
Figure 4: elasticity of world trade to world income by decade
Source: IMF,
World Economic
Outlook, October
The long-term decline in commodity prices
Correlation between stagnation and commodity dependence
R Prebisch, H Singer: secular decline in world real prices of commodities and
the deterioration in the terms of trade/ToT of commodities vis-à-vis
=> industrialisation, as productivity and technical progress = key factors of
growth (Prebisch, 1959).
Maizels (1984, 1987): over the long term, the trend in the commodity terms of
trade deteriorates: 3 key factors:
the low price-and-income-elasticities of demand for commodities vis-à-vis
the technological superiority of developed countries and the economic power of
their transnational corporations, which allows these countries to capture excess
profits in trade with underdeveloped areas;
the asymmetrical impact of labour union power in developed countries and
labour surplus in developing countries on the division of the benefits of increased
Negative relationship growth-primary products found within developed countries,
e.g., across regions within the US (Papyrakis and Gerlagh, 2007).
Composite index of commodity prices of The Economist: continuous decline since
1845: in 1999, the industrial commodities index had fallen to a record low in real
terms: 80% below its level in 1845 (1845-50=100, and 1999=20) (The Economist,
1999). Over 1862-1999, Cashin and McDermott (2002): downward trend in real
commodity prices by 1% per year over that period + no break in this long-run trend.
IMF (2009a): over the long-run, prices for many commodities have declined relative to
those of manufactures and services.
However, this trend is questioned: commodity prices explained by periodic structural
breaks (Grilli and Yang, 1988) .
IMF (2009a): the secular decline stems from productivity gains in the commodity
sectors; many commodities’ share in total consumption declines as income
increases: but rates of decline vary across commodities (available reserves, industry
structure, demand characteristics, etc). Oil: exception in the decline: oligopolistic
supply structure, concentration of reserves.
Figure 5: The Economist composite index (The Economist, 15th April 1999)
Volatility as a key characteristic of commodity prices
Commodity-dependent countries to be caught in poverty traps, because of
commodity price volatility.
The more a country is dependent on commodities for exports, the more
relative prices (between tradable and non-tradable) may become volatile
(Hausmann and Rigobon, 2002).
Volatility demonstrated on an historical scale. Cf Cashin and McDermott
(2002): the Economist’s composite commodity price index over 1862-1999:
“ratcheting-up” in the variability of price movements; increasing amplitude of
price movements in the early 1900s; increasing frequency of large price
movements after the collapse of the Bretton Woods regime of fixed exchange
rates (early 1970s).
the downward trend in real commodity prices “completely dominated by the
variability of prices”.
Helbling et al. (2009): despite the integration of commodity markets, commodity price
fluctuations dominated by the prices of a few commodities: the 2002-08 price boom =
energy and metals price boom, prices tripled between mid-2002 and mid-2008. Metals
prices follow demand fluctuations (global industrial cycle). Second half of 2008: sharp
drop: energy prices declined by 70%, metals prices, 50%; food prices, by 30%.
Figure 6: historically, commodity prices have been volatile and subject to large
swings (real commodity prices, constant U.S. dollars, 1990=100)
Source: Helbling et al.
Figure 7: Confirmed at a secular scale …
Source: Streifel (2006)
Increasing vulnerability due to the linkages between
Integration of commodity markets among themselves, and of commodity
and financial markets.
Maizels (1984, 1987): intrinsic instability of commodity markets.
Maizels (1994): ‘financialisation’ of commodity markets
Increasing role of the financialisation in the 2000s in price volatility:
impact of derivative markets on price volatility (Nissanke, 2009).
‘International poverty traps’: low productivity, debt traps =>combination of
international trade and finance relationships reinforces the cycle of stagnation,
which, in turn, reinforces the negative impact of external relationships.
UNCTAD (2002): closer linkages between energy and agricultural commodity
markets, as well as commodity and financial markets over the 2000s.
=> increase in price volatility and therefore uncertainty, detrimental effect on
investment and governments’ financial management (UNCTAD, 2008b;
Sindzingre, 2009).
4. Is commodity-dependence an effective cause of traps?
The criticisms of commodity-based poverty traps
Many arguments against the concept of poverty trap:
i) if traps exist, they may be generated by many other factors than commoditybased market structures;
ii) commodities do not always generate traps;
iii) the very existence of poverty trap may be questioned.
Lack of correlation between commodity-based export
structures and traps: traps caused by factors unrelated to
Critiques of commodity-generated traps: 2 arguments.
1) the lack of convergence between groups of countries -a group growing more
slowly relatively to other countries- is caused by many other factors.
Poverty traps may result from poor public policies: e.g., protection
Or from initial economic conditions: e.g., low savings rates as the latter
depend on the level of per-capita income, or credit market imperfection and
borrowing constraints (Banerjee and Newman, 1994).
Azariadis and Drazen (1990): ‘low growth traps’ or ‘underdevelopment
traps’, i.e. multiple and stable equilibria for economies exhibiting similar
initial conditions, result from ‘threshold externalities’ created by increasing
returns in the accumulation of human capital.
Azariadis (1996): why similar countries do not converge to the same steady
state? Many causes of poverty traps: subsistence consumption, limited
human capital, demographic transitions when fertility is endogenous (as in
SSA), political economy problems such as coordination failures among
Azariadis (2006): non-ergodic growth theory: ‘misbehaving governments’
and incomplete markets = determinants of poverty traps.
Explaining per capita income at the individual or household level:
microeconomic poverty traps stem from spatial processes (Benabou, 2000).
spatial poverty traps (Durlauf, 2003): dynamic, as the residence restricts future
opportunities: if composed initially of poor members, a group will remain poor
over generations = self-reinforcing processes: low level of education, poor
schooling infrastructure, low levels of taxes, limited supply of public goods.
Decision for an individual to acquire an education depends on the prior existence
of other educated members in a group: interdependence of behaviour =>
‘neighbourhood effects’ generate different types of groups that have different
steady states (with/without educated members) (Durlauf, 1996; 2003).
Intertemporal interdependence: it affects future social interactions: dynamics
=> persistent income inequality=reciprocal feedbacks micro/macro levels,
mutually reinforcing.
Jalan and Ravallion (1997): does residence make the difference between growth
and contraction in living standards for otherwise identical households? Spatial
poverty traps and geographic externalities= neighbourhood endowments of
physical and human capital influence the productivity of a household's own capital
(cf. Hoff, 2000, China’s ‘local underdevelopment traps’).
Lack of correlation: the export of commodities as a basis
for sustained growth
2) 2nd argument against commodity-generated traps: many commodityexporting countries enjoyed an increase in their per capita income, not
caught in a poverty trap + grounded growth on primary products, e.g., Australia
(metals), New Zealand (agricultural products), Canada, Scandinavian
Many developed countries have based their industrialisation on natural
Wright (1990): rise of US manufacturing in the 1890s associated with a rise
in the resource intensity of exports (natural gas, petroleum, copper); natural
resource abundance lowered input prices => fostered industrial production (steel
products,) => increase in manufactured exports.
The irrelevance of the very concept of trap
Critique of the concept of commodity-based traps: problems of commodityexporting countries are well-explained by more powerful theories: e.g.,
Dutch disease; ‘resource curse’.
Moreover, irrelevance of the concept of trap for analysing growth trajectories.
= Price profiles in commodity markets follow both trends and cycles (Cashin
and Mc Dermott 2002), or ‘supercycles’.
A cycle is not a trap: over the long-run, SSA growth has moved closely with
global real GDP growth: with the slowing of global growth, SSA exports are
affected by lower external demand and declines in commodity prices (IMF,
Chang and Helbling (IMF, 2009a): long-term trends in commodity prices are
not relevant to the understanding of medium term price fluctuations: rates of
change are highly variable and the trend component shifts over time, reflecting
changes in longer-run price determinants (e.g., costs of marginal fields or mines)
=> price fluctuations reflect those in the trend component or those in the
cyclical component.
Concept of poverty trap challenged by growth ‘acceleration’ / ‘deceleration’
Hausmann et al. (2005): turning points in growth performance. Rapid acceleration =
growth sustained for at least 8 years; 80 episodes since the 1950s.
Growth accelerations correlated with increases in investment and trade, and real
exchange rate depreciations.
external shocks produce short-lived growth accelerations
growth accelerations are highly unpredictable
growth ‘decelerations’, or ‘growth collapses’ do not imply poverty traps.
Saba Arbache and Page (2007a): 44 SSA countries, 1975 -2005: low and volatile
growth since 1975, but no evidence that growth volatility is associated with
economic performance over the long run. But turning point in SSA growth in the
1990s and the possible formation of clubs.
Saba Arbache and Page (2007b): growth accelerations-decelerations in SSA, 19752005: no evidence that growth volatility has a relationship with long-term economic
performance: many growth accelerations and as many growth collapse episodes,
which offset the previous ones.
Interestingly (confirming traps?): growth accelerations-decelerations have an asymmetric
impact on human development outcomes.
Figure 8: GDP per capita and growth rate (constant international $, PPP and
Source: Saba Arbache and Page (2007a).
Easterly et al. (1993): ToT shocks explain a large part of the variance in growth,
but fluctuations of growth rates do not necessarily build a poverty trap,
and these fluctuations do not predict what long-run performance and per capita
income will be: global technological change determines long-run growth,
while country characteristics determine relative income levels.
Easterly (2005): the concept of poverty trap is irrelevant: in SSA, over the last
50 years, levels of income per capita have increased slowly despite high
fluctuations in terms of growth rates.
Poverty traps in the sense of zero growth for low income countries rejected by
the data in most time periods.
Divergence between rich and poor nations in the long run does not imply zero
growth for poor countries.
Kraay and Raddatz (2005): no evidence of traditional determinants of poverty
traps: low savings, low technology, low productivity in low-income countries
=> no ‘unfavourable initial conditions’: poverty depends on policies.
5. The concept’s explanatory power: recognising its
definitional features in commodity-dependent countries
The relevance of traps’ 3 main definitional features: low equilibria, cumulative
causation, lock-in and threshold effects
1) the critiques of the concept of poverty traps overlook the definitional features
and properties of the concept= poverty traps refer to growth processes that are nonlinear, subject to cumulative causation, increasing returns, multiple equilibria
and threshold effects.
P Arthur, P David: ‘small events’ may induce large effects that may be irreversible.
Key features = path dependence, irreversible processes (weight of history), lock-in
processes - economies being attracted within a low equilibrium, and their dynamic
consequences, e.g. self-reinforcing lock-in, stabilisation, etc., which make the
reaching of a higher-growth path more costly.
This is why a trap cannot simply be assimilated to growth rates movements such
as decelerations, or fluctuations
2) the concept of trap = processes that are dynamic and relative to other
countries’ dynamics: countries caught in low equilibria, trapped in basins of attraction
(in terms of growth, efficiency) that are lower than in other countries.
Do commodity-dependent low-income countries exhibit these definitional
Low equilibria trapping commodity-dependent countries
UNCTAD (2008a) dependency rate = average share of the 4 main commodity
exports value/value of total exports for the period 2003–2005. Dependency rate >
50% => more than 50% of earnings from exports come from the 4 commodities:
more than half of all developing countries rely on 4 commodities for 50% of
their exports earnings; 31% rely on 4 commodities for more than 75% of their
export earnings.
Relationship with low per capita income: among the 45 LDCs, 30 =dependency
rate > 50%; in Africa, 34 of the 52 countries =more than 50% dependent. DR
> 80% = West African countries, Western Asian countries (oil). Also
agricultural products (cotton, cocoa, coffee): some SSA countries: DR > 65%.
Transmission of world prices shocks and volatility to producing countries =>
volatility of macroeconomic aggregates =>low equilibria (Baffes and Gardner,
Volatility prevents diversification, risk-pooling and long term growth strategies.
Oil countries: better prospects? 2005-08 commodity boom =>impressive growth rates.
But fragility (IMF, 2006): negative relationship between macroeconomic volatility and
growth (Loayza et al., 2007; Ramey, 1995).
Figure 9: macroeconomic volatility and economic growth
Source: Loayza et al., 2007, based on the World Development Indicators, cross-country sample, 1960–2000
Negative impact of terms of trade volatility and shocks on growth: Kose and
Reizman (1998): shocks, i.e. fluctuations in the prices of primary commodities
=> significant decrease in growth and aggregate investment in SSA.
The argument that many developed countries started growth with primary
products does not hold for low-income countries, where commodities cannot
be utilised as inputs in industrial processes. E.g., cocoa, coffee, oil.
Leamer et al (1999): commodity-based market structures can increase income
inequality: on Latin America/ East Asia: natural-resource-intensive sectors (e.g.,
agriculture) absorb capital that otherwise flow to manufacturing => reduces skill
accumulation =>impedes industrialisation.
Questionable argument=long-lasting low equilibria do not exist, some
countries got out of them. Commodity-dependent low-income countries differ
from the 1960s Asian ‘developmental states’ (and China), i.e. growth based on
state-led industrial sectors, protection, limited natural resources, education, etc.
Many commodity-exporting countries = stabilisation in a low
equilibrium, in a low basin of attraction, path dependence, weight of past
market structures, remarkable stability of their export structure over
E.g., at the beginning of the 20th century, Senegal produced 141 000 tons of
groundnuts: 68% of its exports in 1929, and 80% in 1960; this commodity
was still Senegal’s principal export at the end of the 20th century;
E.g., in 1990, oil represented 97% of Nigerian exports, in 2002, 100%, and
98% in 2005 (WDI 2004, 2006, 2007).
Persistence of a low industrial base: in 1990, SSA thus represented 0.79%
of world industrial output, and in 2002, 0.74 %; without South Africa, in
1990, 0.24%, and in 2002, 0.25% (UNIDO, 2005).
Nissanke (2009): channels of transmission of the financial crisis to lowincome countries= precipitous fall in commodity prices, escalating cost of
trade finance and severe difficulties in accessing trade credit.
Cumulative causation and increasing gaps between groups
of countries according to their export structure
Concept of poverty traps = a relative concept.
Even if poor countries do grow, this does not refute traps: specific market
structures create traps relatively to other countries’ growth trajectories.
Hausmann and Rodrik (2006): industrialisation requires structural
transformation, i.e. changing the exported products. But market failures: for a
given level of development, countries with more advanced export package will
grow more rapidly, while the other countries are constrained by the low
productivity associated with their export package.
Dynamics of an increasing gap commodity-exporting countries/other groups:
empirical observation of the secular decline in the price of commodities.
Macroeconomic volatility increases likelihood of lower equilibrium, as it
entails asymmetrical processes: busts last longer than booms (Cashin et al.,
Divergence confirmed in a historical perspective: Blattman et al. (2004):
negative consequences of the exporting of commodities, because they have been
more volatile than other products: countries with more volatile prices have
grown slowly relative both to the industrialised countries and to other primary
product exporters (panel of 35 countries, 1870-1939). Volatility was much more
important for accumulation and growth than was secular change.
A channel = foreign capital inflows declined where commodity prices were volatile
+ asymmetry industrialised/developing countries: changes in volatility had a
negative influence on income growth in developing countries, but not in
industrial countries: divergence confirmed by the asymmetry of the impact of
ToT shocks (Hadass and Williamson, 2003 for the 1870-WWI period), confirming
H Singer: the long-run impact of relative price shocks reinforced industrial
comparative advantage in the ‘centre’ and favoured the sector that carried growth,
while it reinforced primary product comparative advantage in the ‘periphery’,
harming the sector that fostered growth.
Booth (2008): West Africa and South East Asia: widening gap in the 20th century
for agricultural development, export growth and the impact of a shock such as the
1930s slump. South East Asian countries benefited from increases in productivity
and public policies, vs. West African countries.
Commodity-producing countries grow because their products are the object of
international demand (e.g., oil, copper): this global demand is fluctuating and
unpredictable, and is boosted by technology intensity.
Due to technological progress, the quantity of commodities used in a unit of GDP has
steadily decreased since 1971 (World Bank, 2009a).
Figure 10: technological progress has reduced the quantity of commodities used per
unit of GDP
Source: World Bank
(2009), Global
Economic Prospects
 This dynamic and relative dimension of the concept of poverty trap: world
distribution of output per worker =‘twin peaked’ shape: low income countries
associated with a specific export structure.
 Since the 1960s, only East Asian countries caught up with industrialised countries;
the less developed countries are not catching up: polarisation of growth rates
(Quah, 1996).
Table 2: Annual growth rates in p.c. GDP, 1870-1994 (std. deviations in parentheses)
Source: Pritchett (1997).
Azariadis (2006): LDCs grow a bit slower and less predictably than the world
average. Outside East Asian countries, less developed countries are not catching up
with OECD countries.
In dynamic terms, even if these countries grow slowly (Easterly), the elevation of their
income per capita is slower than other group of countries: they do not converge:
clubs of countries with differing growth profiles.
The continuous decrease of the share of SSA in world trade is another signal of
the divergence of a ‘club’ of countries.
Figure 11: Sub-Saharan Africa's share of world exports
Source: Subramanian and Matthijs (2007).
Poor commodity-dependent countries caught in endogenous processes: low
productivity, low value-added and the export of commodities reinforce each other.
These factors cumulate and push economies towards lower equilibria.
Threshold effects, bifurcations and lasting impacts of shocks
Commodity-dependent countries: a 3rd feature of poverty traps= small shocks
may generate large effects and make countries fall into lower equilibria:
A fortiori important and recurrent shocks, world business cycles and commodity
prices cycles, which affected international trade after the 1970s.
Volatility, shocks => thresholds, lock-in effects, irreversibilities: prevents
investment and its profitability, and makes fiscal and debt management
+ Commodity markets are integrated => increasing returns, feedbacks.
Commodity-dependent countries = more likely to be exposed to external shocks.
Funke et al. (2008) (159 countries, 1970-2006) on persistent terms of trade shocks:
SSA and the Middle- East more affected than Western Hemisphere and AsiaPacific countries, due to these 2 regions lesser diversification, dependence on a
few natural resources and lower manufacturing base. SSA exhibited in average
more than 2 persistent terms of trade shocks.
Commodity price volatility=devastating shocks on the macroeconomic management
of countries. Very difficult to maintain any fiscal balance and a credible state capacity
with highly volatile and unpredictable revenues in countries where revenue from
commodities may represent more than ¾ of total revenue.
Figure 12: Commodity revenues to total revenue, 2008 (ratio, in percent of total
Source: International Monetary Fund (2009b)
The shock of the financial crisis…
Figure 13 : the fiscal vulnerability of commodity exporters
Source: IMF (2009), The Implications of the Global Financial Crisis for Low-Income Countries—An Update
Hence: a large shock such as the 2008-09 global recession => threshold
The 2003-09 price shock= largest, longest since 1900, after 3 major
commodity booms and slumps in the 20th century (1915–17; 1950–57; 1973–
74 (World Bank, 2009a).
2008: Oil countries: the sharpest price fluctuation within a year.
Natixis chronicle, Nov. 2009: since mid-2008, a major deterioration in global
trade in value terms: unprecedented levels, more than -30% in year-on-year
terms in the beginning of the spring, partly due to the collapse in commodity
prices, first oil. Oil accounts = 10% of global trade, + its price divided by 5
between mid-2008 and end-2008 (though less for trade figures in volume).
Sharp decline in global demand: most severe since WWII: it affected goods
(capital goods, durable consumer goods, cars, etc.) financed by credit before
the crisis, + drying up of the credit to the private sector.
IMF (2009a): magnitude of price changes and volatility rose to unprecedented
levels for many major commodities.
Table 3: Comparison of commodity price volatility (weekly; in percent)
Source: IMF (2009a, April)
Compared with previous cycles since 1950, prices fell much further and faster during
the crisis and have recovered far more quickly. Compared with an average cycle,
commodity prices dropped by 3 times the usual amount (IMF, WEO, April 2010)
The IMF commodity price index had declined by 55% between the July 2008
peak and December 2008.
Figure 14: Commodity and petroleum prices
Source: IMF, World Economic Outlook, April 2010
Figure 15 : the impact of the crisis on SSA: the plunge in commodity prices
Sources: IMF, Regional Outlook, SSA, October 2009; IMF, Commodity Prices, and UN Comtrade. 1: Composite of
cocoa, coffee, sugar, tea, and wood, weighted by SSA exports.
Source: IMF SSA Regional Outlook, April 2010: for oil producers, massive terms
of-trade losses in 2009, 26.8% of GDP, coincided with a reversal in financial flows
of 3.8% of GDP. For the non-oil-exporter group, terms-of-trade gains in 2008–09
offset the financing shock.
Figure 16: Terms of Trade and Financial Shocks,
Sub-Saharan Africa Non-Oil Exporters
Figure 17: Terms of Trade and Financial
Shocks, Sub-Saharan Africa Oil Exporters
Figure 18 : SSA: declining export demand
Source: IMF, Regional
Outlook, SSA, October
2009; IMF, World
Economic Outlook.
 Developing countries with export-based market structure face a fall in demand from rich
countries for their products => end on investment projects, increased unemployment
(IMF, 2009a) - investment and employment being the aggregates that have the largest
impact on future incomes.
 SSA: drop in external demand, falling export prices, lower capital inflows => expected
growth rate of 1% for 2009, from 5.7% over 2006-08 (World Bank, 2009b).
 IMF: WEO, Oct. 2009: SSA: a weaker-than-expected recovery of the global economy
would slow the recovery in commodity markets => lower inflows (FDI).
Commodity prices booms, e.g., 2003-08: these shocks also have a negative
impact, i.e. increased dependency vis-à-vis commodities due to higher prices.
Undiversified export structure, dependence on volatile and declining
earnings=> vulnerability to external shocks = ingredients of a lack of
resilience to shocks: for economies at the tipping point - in export earnings,
fiscal equilibrium, institutional, individual income, etc – these ingredients
precipitate a fall to a lower equilibrium.
Ex. 1979 drop in commodity prices (1986 for oil countries): despite growth rates
in the 2 previous decades, it toppled commodity-exporting countries into
lower equilibrium still enduring =>‘prolonged users’ of IMF financing: 3
decades later, high cost of getting out of these ‘bad’ equilibria – policies,
financing–, beyond the capacities of any ‘big push’.
Commodity-dependence=>countries dependent on imports of the
commodities they do not export. The volatility of the 2005-08 commodity
boom hit food-importing countries, with many are at subsistence level: pushed
entire groups into poverty.
The transmission of shocks and trapping processes
At the micro-level, macroeconomic volatility – shocks on commodity prices
transmitted to producers’ earnings, or creating fiscal deficits reducing publicly
provided social security - triggers irreversible processes for the individuals
close to subsistence income.
Individuals sell the assets necessary to their future income and productivity =>
pushes them in a trapping, lower equilibrium: selling land, reducing spending on
children’s education.
Zimmerman and Carter (2003): poverty dynamics: different households respond
differently to income shocks depending on their assets.
Change in technology enhances productivity, but requires capital and access to
credit, which creates thresholds and traps at the household level
The rich have access to credit, investment, higher productivity, higher returns. The
poor are caught in a poverty trap compounded by indebtedness (limited access
to credit markets, moneylenders’ distorted interest rates).
Rates of return positively correlated with initial wealth, which creates
threshold-based multiple equilibria (Barrett and Carter, 2005).
=> Deep-rooted, persistent structural poverty vs. temporary poverty
bifurcated accumulation strategies;
dynamic asset poverty thresholds (Carter and Barrett, 2006).
Past experience of shocks on producers’ earnings: e.g., in Indonesia, during the
1997-98 Asian crisis, household spending on education declined – e.g., children
were withdrawn from schools -: even more among the poorest households
(Thomas et al., 2004).
The 2008-09 recession =similar effects. World Bank (2009b): the fall in
internationally traded food prices should alleviate the increases in poverty of
the first half of 200: but does not offset the increase in extreme poverty from
the increase in local food prices between January 2005 and mid-2008
domestic food prices may decrease, but with a lag.
Even if the number of people in extreme poverty decreases, ingredients of
irreversible negative effects on the human capital of future generations
=> possible intergenerational poverty traps (Dasgupta, 1997).
6. Reaching higher equilibria? Commodity-based traps
fostered or countered by the combinations of several
Causality does not mean determinism: the feedbacks between
commodity-based traps and other factors
Argument dismissing the view that commodity-based export structures foster the
formation of poverty traps, since some countries grounded their growth on natural
resources: not a valid argument: the impact of this market-structure may be modified,
reoriented, countered or intensified, by other processes:
=these countries’ ‘initial conditions’: ie, the history and credibility of their
economic and political institutions, level of education, demographic and geographic
characteristics, etc.
Barrett and Swallow (2006): ‘fractal’ poverty traps = a trap in which multiple
dynamic equilibria involve simultaneously micro (households, individuals), meso
(communities), macro - scales of analysis”, these 3 levels “self-reinforcing through
feedback effects”. Simultaneous involvement of all levels => an economy stabilised in
such equilibrium has difficulties to get out of it and reach a different one:
governments, markets and communities simultaneously trapped in low-level equilibria.
Institutions as key factors of the countering or reinforcing
commodity-based traps
Countries which caught up harnessed many factors: e.g., human capital,
capacity of innovation (Thorbecke and Wan, 2004). Contrasts with commodityexporting low-income countries: low levels of human capital, lack of industrial
sectors and labour markets absorbing educated workforce => dualistic market
structures, no spillover effects (cf SSA oil countries).
Commodity-based low-income countries not endowed in the factors that
endogenously cause growth: moreover, they are endowed in primary
products, which generate disincentives for these growth-enhancing factors:
typically, disincentives for education + oligarchic and corrupt political
economies that limit education to an elite.
Countries which succeeded in the catching-up process developed specific
institutions and in turn were helped by them
The form of institutions having a positive or negative relationship with growth, is
difficult to assess ex ante (Engerman and Sokoloff, 2003; Sindzingre, 2007b).
Engerman and Sokoloff, (e.g., 2006) on growth paths divergence in the two
Americas: institutions may create poverty traps, as they shape
Unequal economic and political institutions persist though they close to many
individuals opportunities (land, education, capital), combining with
endowments and market structures (climate, labour abundance).
Typically, ‘low’ equilibria include institutions generating the lock-in of
social groups, e.g., kinship norms: Hoff and Sen (2006): “collective
Bowles (2006): why institutions that have implemented “highly unequal
divisions of the social product” have been so widespread; why they persist
even in cases “where they convey no clear efficiency advantages over other
feasible social arrangements”.
Evolutionary perspective: unequal institutions persist because these
arrangements are “self-enforcing conventions”.
Low or high equilibria as outcomes of self-enforcing
combinations of market structures and institutions
Institutions = key element of the feedback processes underlying commoditypoverty traps.
Poor institutions combined with commodity dependence maintain slow growth.
Slow growth combined with commodity dependence maintains poor institutions.
Commodity dependence, volatile commodity prices, volatile growth rates maintain
poor institutions,
which, in turn, reinforces the negative effects of commodity dependence (Mehlum
2002; Robinson, 2002; Auty, 2001; 2006).
Symmetrically, institutions may shape the exploitation of natural resources in a
way that prevents traps and foster industrialisation.
Wright (1990): end-19th century US, specific institutions transformed the endowment
in natural resources (minerals) in engines of industrialisation and increasing returns:
legal system, geological research, public knowledge, education system: mineral
abundance= “an endogenous historical phenomenon driven by collective learning,
increasing returns, and an accommodating legal environment”.
E.g., Norway: the risks of oil countered by institutions centred on equality and
efficiency, rulers’ long time horizon = institutions able to lock-in governments’
commitments, or ‘meta-institutions’ (Acemoglu, 2003; Kydland and Prescott,1977)
preventing a fall into a worse equilibrium (Dutch disease): Petroleum Fund (Mehlum et
al., 2008).
Combination of appropriate policies and existing institutions modified the negative
effects of commodities in Scandinavian countries: when institutions are ‘producerfriendly’, more natural resources may increase income (Mehlum et al., 2006).
Opposite outcome when such institutions are lacking: even accelerate the fall in a
commodity trap when political instability, predatory rulers, high inequality.
Therefore, multiple ‘growth-export structures-institutions’ equilibria - ‘low’ or
‘high’, possible bifurcations towards growth.
Low equilibria not created by commodity-dependence nor locking-in institutions
alone, but by a combination - a low level of income, high inequality, narrow industrial
sector, an export-structure based on a few commodities. Blattman et al. (2004):
combination of a commodity-based export structure, of volatility, and local
institutions =>lower growth performance.
More than elements taken in isolation, ‘combinations matter’.
7. Conclusion
Poor commodity-dependent countries exhibit key properties of the concept of
the poverty trap:
= non-linear growth processes; small events’ irreversible effects; low equilibria;
cumulative causation and increasing gaps with other groups of countries;
thresholds and lasting effects created by external events; and simultaneity of
macro and micro trapping processes.
Against the critiques of ‘traps’: causalities do not constitute determinism:
Commodity-based market structures combine with other determinants of
growth (institutions), which may aggravate the negative impact, OR, on the
contrary, transform the link towards a basis for growth.
These other determinants are often endogenous to growth => unlikely that
institutions have this latter capacity in low-income countries: this
endogeneity is one of the features of poverty traps.
Fly UP