Chapter 1: Contemporary Models of Development and ...remi.bazillier.free.fr/L3_DevEcon_english_LeCaire.pdf · I A review of the most inﬂuential ofnew models of economic development

Chapter 1:Contemporary

Models

Rémi Bazillier

Underdevelopmentas a coordinationfailureEndogeneous Growth andComplementarities

Coordination failures

Multiple Equilibria:a diagrammaticapproach

Starting EconomicDevelopment: TheBig PushA graphical model

O-Ring Theory

Growthdiagnostics(Rodrik)

Chapter 1: Contemporary Models ofDevelopment and Underdeveloppment

Development EconomicsFESP - Panthéon Sorbonne Master in Development

Economics

Rémi Bazillier 1

1 [email protected]://remi.bazillier.free.frUniversité Paris 1 Panthéon-Sorbonne

Ecole d’Economie de la Sorbonne

[email protected]

http://remi.bazillier.free.fr


Models

Rémi Bazillier





O-Ring Theory


Introduction

I Traditional theories:I Development as a linear phenomenom

I Rostow’s stages of growthI Harrod-Domar ModelI Structural Change Models (Lewis)

I Problem: Development is possible (East-Asia) butmuch more difficult that what we tought(Sub-Saharian Africa)

I Development does not happen automatically; itrequires systematic effort→What makes the studyof economic development so important


Models

Rémi Bazillier





O-Ring Theory


Introduction

I A review of the most influential of new models ofeconomic developmentI These models show that development is harder to

achieve, it faces more barriers than had previouslybeen recognized

I These new models have already influenceddevelopment policies and modes of internationalassistance


Models

Rémi Bazillier





O-Ring Theory


Introduction

I One major theme: incorporating problems ofcoordination among economics agents (firms,workers)

I Other theme: formal exploration of situations inwhich increasing returns to scale, a finer division oflabor, the availability of new economic ideas orknowledge, learning by doing, informationexternalities, and monopolistic competitions

I It often incorporates insights of new institutionaleconomics (Douglas North)

I It departs to some degree from conventionalneoclassical economics (perfect imperfection, noexternalities and the uniqueness and optimality ofequilibria)


Models

Rémi Bazillier





O-Ring Theory


Outline

Underdevelopment as a coordination failureEndogeneous Growth and ComplementaritiesCoordination failures

Multiple Equilibria: a diagrammatic approach

Starting Economic Development: The Big PushA graphical model

O-Ring Theory

Growth diagnostics (Rodrik)


Models

Rémi Bazillier





O-Ring Theory


Complementarities

I Complementarities between several conditions arenecessary for succesful developmenntI Investments must be undertaken by many agents in

order for the results to be profitable for any individualagent

I An action taken by one firm, worker, or organizationincreases the incentives for other agents to takesimilar actions

I Endogeneous Growth ModelsI Increasing returns due to these complementarities

I Coordination Failure Approach


Models

Rémi Bazillier





O-Ring Theory


Endogeneous Growth theories

I Limitations of the Solow model:I Similar access to technology (A) for all countriesI Technological progress is exogeneous (growth of A)I Decreasing returns of capital

I How to sustain growth: positive externalitiesI More benefits from inputs in order to compensate

decreasing returns


Models

Rémi Bazillier





O-Ring Theory


Endogeneous growth (Romer)

I Externalities: Learning effects (learning by doing)I Capital accumation in one firm has positive spillover

effects (externality) on productivity of other firms

Yi = AKαi L1−α

i Kβ

(1)

I If all sectors are identical, the agregate productionfunction is

Y = AKα+βL1−α (2)


Models

Rémi Bazillier





O-Ring Theory


Endogenous Growth (Romer)

I Implications: Constant returns at the firm level andincreasing return at the macroeconomic levelI Positive growth in the long-run (if β > 0)

g − n =βn

1− α− β(3)


Models

Rémi Bazillier





O-Ring Theory



I A state of affairs in which agents’ inability tocoordinate their behavior (choices) leads to anoutcome (equilibrium) that leaves all agents worseoff than in an alternative situation that is also anequilibrium

I This may occur even when all agents are fullyinformed about the preferred alternative equilibriumI Because people hold different expectationsI or because everyone is better off waiting for

someone else to make the first move


Models

Rémi Bazillier





O-Ring Theory


Coordination failuresI When complementarities are present, an action

taken by one firm increases the incentives for otheragents to take similar actions

I These complementarities often involves investmentswhose returns depends on other investments beingmade by others

I Examples:I Big Push, in which production decisions by

modern-sector firms are mutually reinforcingI O-Ring model, in which the value of upgrading skills

or quality depends on similar upgrading by otheragents

I Analyses of frontier technologies in developedcountries (information technologies)

I Analyses of middle-income trap in which countriesdevelop to a degree but chronically fail to reachhigh-income status, often due to lack of innovationcapacity


Models

Rémi Bazillier





O-Ring Theory


Examples

I Skill workersI Firms will not enter a market or locate in an area if

workers do not possess the skills the firms needI But workers will not acquire the skills if there are no

firms to employ themI This coordination problem can leave an economy

stuck in a bad equilibrium (low average income with aclass of citizens trapped in extreme poverty)


Models

Rémi Bazillier





O-Ring Theory


Examples

I Commercialization of agricultureI Specialization is one of the sources of high

productivity but we can specialize only if we cantrade for the other goods and services we need

I In the development of agricultural markets,middlemen play a key role by effectively vouching forthe quality of the products they sell

I There needs to be a sufficient number ofconcentrated producers with whom a middleman canwork effectively

I But without available middlemen to whom thefarmers can sell, they will have little incentive tospecialize in the first place and will prefer to continueproducing a range of goods primarily for personalconsumption or sales within the village

I Underdevelopment trap in which a region remainsstuck in subsistence agriculture


Models

Rémi Bazillier





O-Ring Theory


I It expands the scope for potentially valuablegovernment policy interventionsI Pushing the economy toward a self-sustaining, better

equilibrium, in which tehre is no incentive to go backto the behavior associated with the bad equilibrium

I Organizing complementarities / investments andfostering new markets

I Deep intervention in the first steps of development (inorder to reach the high equilibrium), then governmentcan focuses on other goals (education / health, etc)

I Complementarities imply a certain level ofaglomerationI It can turn to congestionI ... Which is also an externality (but negative)


Models

Rémi Bazillier





O-Ring Theory


Multiple Equilibria

Equilibrium: All participants are doing what is best for them, given what theyexpect others to do, which in turn matches what others are actually doing(when the function crosses the 45-degree line)


Models

Rémi Bazillier





O-Ring Theory


I S-shaped “privately rational decision function”I it first increases at an increasing rate, and then at a

decreasing rateI Some agents may take the complementary actionI If only few agents take the action, spillovers are

minimalI Then snowball effect, in which many agents begin to

provide spillover benefits to neighboring agents (thecurve increases at a faster rate)

I After most potential investissors have been positivelyaffected, the rate of increase starts to slow down

I Other curves are possible, depending on the natureand scale of spillovers


Models

Rémi Bazillier





O-Ring Theory


I Equilibrium: when agents observe what theyexpected

I Stable and unstable equilibriumI Here D1 and D3 stableI D2 not stable


Models

Rémi Bazillier





O-Ring Theory


I Coordinating investments decisionsI when the value (rate of return) of one investment

depends on he presence or extent of otherinvestments

I Coordinating investments and technologicalexternalities (as in Romer)

I Possibility to be stuck in a low equilibrium ifpessimistic expectations of investors

I Strategies / policies to change expectations (exampleof mañana mañana in Latin America)

I But it might be insufficient if it is more profitable forfirms to wait for others to invest rather than to be a“pioneer” investor

I In that case, government policy is generally neededin addition to a change of expectations


Models

Rémi Bazillier





O-Ring Theory


Big Push TheoryI Why should it be so difficult to start modern growth?

I In classic theories of development (such as the Lewismodel), perfectly competitive conditions in theindustrial sector

I Under perfect competition, it is not clear why startingeconomic development would be so difficult(provided at least human capital / technology / goodbasic services provided by the gvt)

I Development seems hard to initiate even when bettertechnologies are available (they often go unused)→people do have the incentives to use these newtechnologies

I Perfect competition does not hold under conditions ofincreasing returns to scale

I Looking at the industrial revolution, taking advantageof increasing returns has been key

I The role of market failures, notably pecuniaryexternalitiesI Spillovers effects on costs and revenues


Models

Rémi Bazillier





O-Ring Theory


Big Push, Rosenstein-Rodan (EconomicJournal, 1943)

I Basic coordination issues in a closed economy (noexports)

I Who will buy the goods produced by the first firm toindustrialize?I Starting from a subsistence economy, no workers

have the money to buy the new goodsI The first factory can sell some of its goods to its own

workers (but no one spends spends all of one’sincome on a single good

I Profitability of one factory depends on whetheranother one opensI Which it turn depends on its own potential probability,

that it turns to the profitability of other factoriesI → circular causationI Coordination failure problem


Models

Rémi Bazillier





O-Ring Theory



I Need to train workers, who are accustomed to asubsistence way of lifeI The cost training puts a limit on how high a wage the

factory can payI Once the first firm trains its workers, other

entrepreneurs can offer a slightly higher wage toattract the trained workers (without paying the cost oftraining)

I The first entrepreneur, anticipating this likelihood,does not pay for training in the first place

I No one is trained and industrialization never getsunder way


Models

Rémi Bazillier





O-Ring Theory



I The Big Push is a model of how the presence ofmarket failures can lead to a need for a concertedeconomywide and probably public-policy-led effort toget the long process of economic development underway (or to accelerate it)

I Coordination failure problems work againstsuccessful industrialization

I Formalized by Murphy, Shleifer and Vishny in 1989,simplified and popularized by Krugman (1995) with agraphical approach

I It explains the success of East-Asian countries(South Korea)


Models

Rémi Bazillier





O-Ring Theory


A graphical model

I AssumptionsI Increasing returns in the modern sectorI Factors. Only one factor of production - labour. Fixed

total supply, LI Factor payments.. Two sectors

I Traditional sector: Wage of 1I Modern sector: W > 1 (it may be a compensation for

disutility of modern factory type of work)I In equilibrium, no net utility for workers from switching

sectors during industrialization, but it will represent aPareto improvement of profits are generated


Models

Rémi Bazillier





O-Ring Theory


A graphical modelI Assumptions

I Technology.I N types of products (large number)I Traditional sector: each worker produces one unit

(constant returns to scale)I Modern sector: increasing returns and fixed cost (F ,

here minimum number of workers to start production)L = F + cQwitg c < 1 marginal labor required for an extra unit ofoutput

I Domestic Demand.I Each good receives a constant and equal share of

consumption out of national income (Y/N)I International supply and demand.

I To simplify: closed economyI Most important conclusions will remain in an open

economyI Realistic hypothesis: in South Korea, the dynamic

export sector have benefited from the presence of asubstantial domestic market to which early sales aredirected


Models

Rémi Bazillier





O-Ring Theory


A graphical model

I AssumptionsI Market structure.

I Perfect competition in the traditional sector with freeentry and no profit. Price=1=marginal cost of labour

I At most, one modern-firm can enter each market.They will charge a price of 1

I When a modern-firm enters one market, she gains amonopoly position

I A modern firm enters one market if she is able toproduce at least as much output as the traditionalproducers for the same level of labour


Models

Rémi Bazillier





O-Ring Theory


Conditions for multiple equilibria

I Starting point: subsistence economy without modernproduction

I A potential producer with modern economyconsiders whether it is profitable to enter the marketI For a given level of F , it will depends on:

1. How much more efficient the modern sector is thanthe traditional one

2. How much higher wages are in the modern sectorthan in the traditional sector


Models

Rémi Bazillier





O-Ring Theory


Production function traditional sector


Models

Rémi Bazillier





O-Ring Theory


Production function modern sector


Models

Rémi Bazillier





O-Ring Theory



Models

Rémi Bazillier





O-Ring Theory


I Two cases:I Limited modernization (one good):

I Rise in wage is too limited to change total income (Y)and thus sales (Y/N): point A

I Large modernization (all goods)I Wage rise for all workers → rise of incomeI Level of sales for all producers: Q2 (point B)

I Modernization of one good is a positive externalityfor all other goods because of the induced rise insales and production


Models

Rémi Bazillier





O-Ring Theory


I Modern production (example point A with wage W1)I more likely when F (fixed cost) is lowI more likely whenb c (marginal labour) is low

I If wage=W2: wage is too high for modern productionat point A

I Production should rise up to Q2 (production turns profitable)I It is possible if a sufficient number of modern producers enters the

market (positive externalities)

I At W2, multiple equilibrium depending on theproduction of the modern sector


Models

Rémi Bazillier





O-Ring Theory


I W3I Q2: insufficient to get profitable modern productionI Wage curve beyond the modern production function

(and point B)I It will depend on F and c

I The more efficient is the modern sector (low c) andthe lower are the fixed costs, the less likely is thiscase


Models

Rémi Bazillier





O-Ring Theory


To sum upI For a given level of F and c

I A policy aiming at reducing F (innovation policy) or c(training, investment in productivity) may ease theeconomic take-off

I Wage W1: modern producers are profitable whateveris the number of modern producers. No need forintervention and endogeneous transition towards amodern economu

I Wage W2: Multiple equilibriumI If A(Q1)→ No modern production (too few modern

producers for profitable investment in the moderntechnology)

I If B(Q2)→ Modern production (size of the market islarge enough)

I Possible coordination failure to reach BI Wage W3: no possible industrialization (not

profitable)


Models

Rémi Bazillier





O-Ring Theory


Some remarks

I No need that all sectors switch to the modernproduction to get a sufficient big pushI If enough sectors switch, positive spillovers on

demand for other sectors might be sufficientI Each firm’s failure to take into account the impact of

its investments on demand for other firms’ goodsrepresent a very small distorsion

I But the agregate distorsion (failure to industrialize) isvery large

I Semi-industrializationI Fall of fixed cost in some sectors onlyI Increased dualization of the economy: enclave

economies in wich a modern sector exists side byside with traditional cottage industries in otherproduct sectors


Models

Rémi Bazillier





O-Ring Theory


Other cases in which a big push may benecessary

1. Intertemporal effectsI If W = 1, possible multiple equilibria if investment returns are not

immediateI Although industrialization is Pareto-improving, possible trap if

expected demand in second period is too low (case of pecuniaryexternalities where the firm’s external contribution to overalldemand is not captured)

2. Urbanization effectI Rural (traditional) - Urban (modern) dualismI → Urban demand should be sufficientI → Big Push = urbanization


Models

Rémi Bazillier





O-Ring Theory


Other cases in which a big push may benecessary

3. Infrastructure effectsI Infrastructures have strong impact on firms’ cost / productivityI Infrastructure use depends on industrialization levelI Industrialization increase the scale of the market for other sectors

→ Providing infrastructures more profitableI But infrastructures are not enough by themselves, if other

coordination problems remain

4. Training effectsI Typical problem of coordination: entrepreneurs know that the

workers they train may be enticed away with higher wages offeredby rival firms that do not have to pay these training cost

I Training might be sub-optimal if demand for high-skilled workers istoo low (skill returns)

I Also true in an open economy (barriers to free mobility of workers)


Models

Rémi Bazillier





O-Ring Theory


O-Ring Theory (Kremer, QJE, 1993)

I Modern production: it requires many activities bedone well together in order for any of them to amountto high values→ Strong complementarityI Specialization and division of labor, along with

economies of scaleI O-Ring: taken from the 1986 Challenger disaster, in

which the failure of one small, inexpensive partcaused the space shuttle to explode

I O-Ring theory: explains the existence of povertytraps and reasons that countries cauhgt in such traps


Models

Rémi Bazillier





O-Ring Theory


O-Ring Theory (Kremer, QJE, 1993)

I Production with strong complementarities amonginputsI n tasks, ordererd by level of skill q (0 ≤ q ≤ 1)I The higher the skill, the higher the probability that the

task will be successfully completed (the part createdin this task will not fail)

I Other interpretation: quality index for characteristicsof the good

I The probability of mistakes by different workers isindependent

I O-ring production function (with 2 workers):

BF (qiQj) = qiqj (4)

I With B: productivity (that can be normalized to 1)


Models

Rémi Bazillier





O-Ring Theory


Other hypotheses

1. Firms are risk-neutral2. Labour markets are competitive3. Labour supply is inelastic4. Capital markets are competitive


Models

Rémi Bazillier





O-Ring Theory


I Production function: (positive assortative matching)I Workers will work with workers with similar skillsI Implications: High-value products will be concentrated in

countries with high-value skillsI Each worker wants to be matched with a positive worker →

positive effect on productivity and thus wageI Most productive workers will tend to work together (to

increase their wages)

I Case of an economy with 4 individuals (2 skilled and2 unskilled)

q2H + q2

L > 2qHqL (5)


Models

Rémi Bazillier





O-Ring Theory


Example with 4 individuals (qL = 0.5, qH = 1)

→ Increasing returns


Models

Rémi Bazillier





O-Ring Theory


Implications of the O-Ring model

I Firms tend to employ workers with similar skills fortheir various tasks

I Workers performing the same task earn higherwages in a high-skill firm than in a low-skill firm

I Because wages increase in q at an increasing rate,wages will be more than proportionally higher indeveloped countries than would be predicted fromstandard measures of skills

I In their human capital investment decisions, workersshould consider investments made by other workersas a component of their own decisionI When those around have higher skill; it increase the

incentive to acquire more skills


Models

Rémi Bazillier





O-Ring Theory


Multiple Equilibria and investment in humancapital

→ This graph may explain investments in human capital


Models

Rémi Bazillier





O-Ring Theory


Implications (II)

I A country can be trapped in a low equilibrium withlow production and low quality→ O-Ring effectacross firms and within firmsI → The case for an industrial policy to encourage

quality upgrading (East Asian countries)I O-ring effects magnify the impact of local production

bottlenecks because of their multiplicative effect onother production

I Bottlenecks also reduce the incentive for workers toinvest in skills by lowering the expected return tothese skills


Models

Rémi Bazillier





O-Ring Theory


Implications (III)

I Implications for the choice of technologyI When skill-workers availability is low: a firm is less

likely to adopt a technology exploiting complex tasks(with high added value)

I It may explain why developed countries are morelikely to have large-scale firms, with high-wages

I Migration implications: Brain Drain:I Skilled-workers are more likely to emigrate to find

better job opportunities (higher wages due tocomplementarities)

I A skilled-workers will be more productive (↑ wages)in a developed countries


Models

Rémi Bazillier





O-Ring Theory


To summarize...

If strategic complementarity is sufficiently strong,microeconomically identical nations or groups withinnations could settle into equilibria with different levels ofhuman capital.


Models

Rémi Bazillier





O-Ring Theory


Convergence in policy mindsets

I “What is probably the best argument for theexperimental approach is that it spurs innovation bymaking it easy to see what works” (Banerjee 2007)

I In contrast to the clearcut list of policyrecommendations made by macro-developmenteconomists in the previous years..

I But new development policies since the 2000sI Growth Report: Strategies for Sustained Growth and

Inclusive Development (Spencer Report)I World Bank 2005: “there is no unique universal set of

rules. We need to get away from formulae and thesearch for elusive best practices”


Models

Rémi Bazillier





O-Ring Theory


The example of Growth diagnostics


Models

Rémi Bazillier





O-Ring Theory


The new policy approach

I It starts with relative agnosticism on what works andwhat does not. It is explicitely diagnostic in itsstrategy to identify constraints

I It emphasizes experimentation as a strategy fordiscovery of what works. Monitoring and evaluationsare essential in order to learn which experimentswork and which fail

I It tends to look for selective, relatively narrowlytargeted reforms

I It is supicious of “best practices” or universalremedies. It searches for policy innovations


Models

Rémi Bazillier





O-Ring Theory


Enders, K. (2007). Egypt-searching forbinding constraints on growth, IMF WP 07/57

Documents

Chapter 1: Contemporary Models of Development and ...remi.bazillier.free.fr/L3_DevEcon_english_LeCaire.pdf · I A review of the most inﬂuential ofnew models of economic development