Institutions Rule

R= odrik, Subramanian, and Trebbi. J= EG, 2004

Introduction identifi= es three ‘strands of thought’ on what leads to differences in income lev= els over time.

= 1) <= /span>Geography: climate, natural resource endowment, disease burden, transport costs= , diffusion of knowledge and technology.

= 2) <= /span>International trade / market integration. Moderate or maximal version, stres= sing the maximal version ‘…that trade / integration is a major determinant of whether poor countries grow or not.’ (p. 132)

= 3) <= /span>Institutions.&n= bsp; Property rights and the rule of law. Social infrastructure, rules of the game…

Growth theory is about accumulation: physical capita= l, human capital, and technological change.&n= bsp; Countries differ in their rates of accumulation. Why? 1,2,3 a= re possible explanations.

Quality of institutio= ns is likely endogenous.

Integration with the = world economy is likely endogenous.

F= igure 1 on 134.

From Wikipedia

“The gravity model of trade in international economics, similar to other gravity models in social science, predicts bilateral trade flows ba= sed on the economic sizes of (often using GDP measurements) and di= stance between two units. The model was first used by Jan Tinbergen in 1962. The basic theoretical model for trade between two countries (i and j) takes= the form of:

$3D"F_{ij}$

Where F is the trade flow, M is the economic mass of each country, D is= the distance and G is a constant.”

This can be used as an instrument for observed flows.=

Frankel and Romer (1999), first regress bil= ateral trade flows as a share of GDP on measures of country mass, distance between trading partners, other geographical variables to construct ‘trade hat’. This constructed = one is used in place of the actual one.

As we have seen, colonial mortality rates as an instrument for institutional quality. To exp= and the data set, they also use as instruments fractions of the population spea= king English and Western European languages as a first language.

“The research agenda to which this paper contributes is one that clarifies the priority of pursuing different objectives – improving t= he quality of domestic institutions, achieving integration into the world econ= omy, or overcoming geographical adversity – but says very little about how each one of these is best achieved.” P 136

G= DP per capita in 1995 (by PPP).

Institutional Quality= using Kaufmann’s (KK, or KKM, or KKZ sometimes)

Integration is the ra= tio of trade to GDP.

Geography is distance= from the equator.

Consider scatter plot= s.

Consider table 2, bas= ic OLS.

Now try two stages.

F= irst stage regression, regress institutional index on settler mortality, the construct= ed trade measure and geography.

Also, regress trade integration (observed trade share) on the same set of regressors.

G= ives INS and INT “hat”.

SM and CONST do not a= ppear in the second stage, thus satisfying the ‘exclusion restrictions’.=

L= arge impact of institutions.

Table four refines the findings of table three.

Regress the observed = trade openness on geography and constructed openness, regress the observed institutional quality variable on settle mortality and geography. Taking out the cross effect in tab= le three, looking at ‘own’ instrument only.

Around table five, th= e issue of determinants of growth returns. <= /span>Accumulation of physical capital, human capital, and technolo= gy. What determines these variables th= at are the ‘inputs’ to generate income?

Before turning to the response to Sachs, let us look at his critique: NBER working paper 9490. (2003).

Sachs states “T= hree papers in particular….purport to show that the role of geography in explaining cross-country patterns of income per capita operates predominant= ly or exclusively through the choice of institutions, with little direct effec= t of geography on income” p 2.

Have we been looking = at ‘choice of institutions’?

Sachs objects to the regressions being in levels. = Growth rates rather than levels are preferred.&nb= sp;

A= lso, very few variables to explain an awful lot.

Distance from the equ= ator is a problematic variable to capture ‘geography’.

Exogenous variable: malaria risk. The proportion= of each country’s population that live with risk of malaria transmission. Also, transform this by multiplyin= g it by the rate that is the fatal kind.

First, we don’t= have a good number of cases, since so few reported. That is why using at risk.

M= alaria a disease of the tropics / sub-tropics. However, conditions on the gr= ound (puddles) can matter. Instrument using malaria ecology.

Malaria ecology is hi= ghly correlated with being in Africa.

So RST get results li= ke Sachs, but then put in regional dummies. The malaria coefficients are no longer significant.

So is it malaria (geo= graphy) exerting a direct influence or an indirect one?

In closing up, they p= oint out that there is a difference between using an instrument for estimation reaso= ns and making a casual argument.

Also, policy can be s= een as a ‘flow’ variable, while institutions are a ‘stock’ variable. <= /p>

They point out that t= he first stage ‘rule’ on ‘settler mortality’ is a worse and worse fit if we take it by decade.