=Housekeeping=
  - Next wednesday's readings are a 3-part freakonomics entry
  - Write up to 7, and they will grade the best X (X <= 4)
  - Essays due the monday after a unit ends

=Methods=
Poverty line
  - initially nutrition/food based, but then extended to include quality-of-life
  - even with PPP adjustment, technology and other innovations over time
    increase amount you can buy, and you need to re-evaluate periodically

Datasources to measure consumption
  - surveys (world bank survey): detailed vs. frequent
  - national accounts, such as GDP
    - frequent: released quarterly, etc.
    - crude: don't see income distribution, etc.

The surveys are internally consistent, but happen infrequently.  It's an
  accurate measure every 10 or so years.

==Sampling Solution for Growth==
Dollar & Kray and Sal-e-Martin interpolate between infrequent surveys
  by taking survey to find fraction of income that bottom 20% of people
  represent (s), and multiplying that by GDP each time gdp (y) is released to
  result in the growth in their share.

So x (income share) = sy

Then they plot logx vs logy, to see how poor people's share grows as gdp grows.
  But log x = log s + log y
  So they are plotting log s + log y vs log y.  That's just a shifted line!
  So sure it looks like there's correlation of growth in society to
    growth in poor.

==GINI Coefficient==
Put people in order of income.
x = %ile of ppl ordered by income
frac = % of GDP (y) that x owns

Ploy x vs y, and see a 45-degree line (/)---that's a totally egalitarian society.
  bottom .1 of society owns .1 of GDP.  In a totally unegalitarian society,
  You will see ____|, which says that top person in society has 100% of money.
GINI = 1--> un-egalitarian society
GINI = 0--> egalitarian society

GINI is a relative measure of wealth distribution

Variance (or actually standard deviation) is different---it measures
  how different people are in dollars.

=Duflo's Conclusions (from Dollar & Kray article)=
  - methods matter: internal consistency helps, rather than trying to add
    a time component which has unhelpful data.
  - Growth is good for the poor in absolute terms
    conclusion came from studying household income (in absolute terms) vs gdp.
    same trend as Dollar & Kray, but for maybe more legitimate reasons
  - Inequality is orthogonal to growth
    GINI not correlated with GDP
  - The household income vs gdp data, while showing what dollar & kray also
    said, is a lot more noisy.  The interesting stuff is in the noise.
  - Causality matters.  Oftentimes, it's not _schooling_ that leads to growth,
    it's _anticipation of growth_ that leads to schooling.  This is what
    happened with the famine---people thought it was coming, so they hoarded.

=Essay-writing=
  - not stream of conscience---organize thoughts!
  - you should respond to readings, as questions relate to them
  - feel free to change the topics
  - share an opinion---it's OK to disagree with article