Transcript Lecture 6 - Nottingham

L11200 Introduction to Macroeconomics 2009/10
Lecture 6:
Conditional Convergence and Growth
Reading: Barro Ch.4 : p83-94
4 February 2010
Introduction
• Last time:
– Solow model with no variation in s, n, δ, A
between nations implies all countries (eventually)
move to same GDP per capita and low GDP per
capita nations grow faster: ‘absolute convergence’
– Data appears to reject this
• Today
– Allow these factors to vary and introduce idea of
‘conditional convergence’
Conditional Convergence
• How do the model’s predictions for growth
change when we allow the factors to vary
– E.g. economies have different saving rates
– The economy with the lower saving rate will have
lower steady state k*, y* compared to an
economy with a higher saving rate
– At any level of K(0), the economy with a higher
saving rate will be growing faster
Other factors
• The same is true for n, δ and A
– Higher n implies lower k*, y*
– Higher δ implies lower k*, y*
– Higher A implies higher k*, y*
• And at any K(0), the economy with higher
technology or lower depreciation / population
growth will grow faster.
Implications for growth rates
• This gives to implications
– For a given K(0), the economy with the higher k*
will have a faster growth rate
– For a given k*, a decrease in K(0) raises the growth
rate
• We can write this as:


K / K  f  k (0), k *
 () () 


Y / Y  f  y (0), y *
 () () 
Implications for Convergence
• This may explain the lack of absolute
convergence
– Economies don’t converge to the same GDP per
capita levels, so growth rate doesn’t depend on
level of GDP per capita
– Maybe the economies with lower growth rates
also have lower k*, y* steady states, so they are
on a growth path to a different steady state.
Conditional Convergence
• This is the idea of conditional convergence:
each economy is converging to it’s own steady
state k*, y* determined by it own s, n, δ, A
– This can be tested if we have data on each of
these factors
– Data is available on each: so can plot relationship
between per capita GDP and per capita GDP
growth conditional on these covariates
Conditioning Variables
• Graph actually hold more than just s, n, δ and
A constant. It also controls for other factors
which affect k*, y* not in our model:
– Measures of extent of rule of law and democracy
– Extent of openness to trade
– Investment in health and education
– Measure of inflation
Example I
• Europe after World War II:
– Previously strong characteristics, but capital and
labour had been destroyed by war
– So steady state k*, y* are high, current k low due
to effects of war
– Post WWII fast growth in European economies –
consistent with conditional convergence
Example II
• Sub-Saharan African nations are very poor
– Absolute convergence predicts they should grow
rapidly
– But they don’t: because they have poor levels of
saving and technological growth
– Also (maybe more importantly) they have poor
rule of law, governments, education programmes
and health systems. All factors which influence k*
and y*.
Summary of Progress
• We began with some questions:
– Why are some economies more developed than
other?
– Why do GDP growth rates vary across nations?
– What is the relationship between the level of GDP
and the growth rate of GDP
Explaining the patterns
• Absolute convergence: all economies have the
same steady state. Smaller economies should
grow faster, all should converge to same per
capita GDP.
– Limited evidence for this
• Conditional Convergence: economies
converge to own steady-state, conditional on
structural factors
– Much stronger empirical evidence
Long-Run Growth
• Question still remains: why do we observe
long-run persistent growth rates for U.K. and
U.S.?
– Conditional convergence predict economy moves
towards steady state
– So expect growth rate would slow over time
– But growth rate is steady over time: continual, or
long-run growth.
Summary
• Conditional convergence more plausible
model than absolute convergence
– Better supported by the data
– Explains lack of growth in poorly developed
nations through structural factors
• How to explain long-run growth?
– Need a model in which economy can maintain
high growth rate continually.