Convergence Scenarios: an Overview

1
Convergence Scenariosan Overview

• Jonathan Temple
• (University of Bristol)

2
Classical Convergence

• Approach represented by papers such as Barro and
  Sala-i-Martin (1992) and Mankiw, Romer and Weil
  (1992)
• Based on closed economy growth models in
  Solow/Swan tradition
• Or Ramsey/Cass/Koopmans
• Adds assumptions about world growth

3
Some Basics

• Predicts countries converge in growth rates
• But income levels differ in steady-state
• This is because steady-state determinants vary
  across countries
• Most familiar investment, population growth
• These determinants treated as exogenous

4
A Key Strength

• A key strength of the approach is to recognise
  that, in principle
• there is no reason why most countries cannot
  become rich
• So future may not look like the past
• Viewpoint differs from non-economists

5
Empirical Findings

• Wide variety of convergence estimates
• Recent work suggests rates are slow
• So countries spend a lot of time away from
  steady-state
• Some evidence suggests clubs
• But a lot of uncertainty (small N)
• and asking wrong question?

6
This is the Wrong Approach!

• Bad approach for emissions scenarios
• Estimated convergence rate is (implicitly) an
  average
• Taken across different countries/regions
• Regressions weight countries equally
• Effectively small countries (sub-Saharan Africa)
  carry much more weight

7
The Wrong Approach, Part II

• Approach treats countries as closed
• These economies evolve independently
• For empirical purposes, treats steady-state
  determinants as exogenous
• Rules out feedback from growth
• How to relax these assumptions?

8
Possible Feedback Mechanisms

• Endogenous world growth
• Geography and market access
• Demography
• Endogenous institutions
• Financial development
• Equipment prices
• Energy prices
• Feedback from climate

9
Endogenous World Growth

• Clear externalities in knowledge
• Powerful effects of scale in endogenous growth
  models
• World population of researchers expanding rapidly
  (India, China?)
• Will only be offset if research process is
  becoming more difficult

10
Geography and Market Access

• Recent work in trade emphasizes role of spatial
  location and market access
• Suggests multiple equilibria at level of world
  regions
• Fast growth in specific regions will change
  patterns of market access
• Hence feedback from regional growth

11
Demography

• Crucial for emissions scenarios to build in
  endogenous demography
• Offsetting effects rising longevity versus
  fertility transition
• Quantity/quality trade-off models predict human
  capital accumulation rise
• Importance of policies in China, India

12
Endogenous Institutions

• Work such as AJR (AER 2001) suggests institutions
  crucial to steady-state GDP
• Strong correlations between institution indices
  and levels of GDP
• Endogenous democratization
• might reduce forecast dispersion?
• Less risk of civil war as income rises?

13
Growth and Finance

• Much work links growth to financial development
• Development of banks endogenous to income level
• Then stock markets
• Financial globalization relevant to convergence
  (e.g. FDI flows)

14
Equipment Price Effects

• Why does real investment differ?
• Work such as Hsieh/Klenow suggests relative price
  of equipment crucial
• Evidence suggests this price much higher in poor
  countries
• Can be expected to fall as part of convergence
  process?

15
Energy Price Effects

• Price of energy unlikely to be independent of
  global convergence
• May feedback into world growth
• This effect will also tend to reduce dispersion
  of emissions scenarios, relative to dispersion of
  GDP scenarios

16
Global Warming Feedbacks

• Central point endogenous climate
• So different convergence scenarios may have
  feedbacks via climate
• For example, agriculture, water stress
• Changes in infrastructure, global distribution of
  population, etc.

17
How to Model All These Effects?

• Full structural model very complex
• So best approach may be reduced-form
• Aggregate major regions
• Consider different convergence behaviour in each
• But build in virtuous circle effects as a
  reduced-form