Syndicate content

Difference-in-difference

A Curated List of Our Postings on Technical Topics – Your One-Stop Shop for Methodology

David McKenzie's picture

With William Kerr and Alexander Segura.

Developing countries have become much more globally integrated. On a global dimension, they now trade a lot more with other countries than they did two decades ago. Domestic connectivity at the country level has been helped through investments in road networks as urban and rural areas are becoming better connected not just through roads but phones lines and faster flows of knowledge.

How global and local connectivity influence spatial development and distribution of jobs is a tricky question. A naïve approach might look industry-by-industry at their exposure to trade or highways and ask what happens to the average wage in the industry as highways form.  This approach, however, risks confounding several forces.  It could be, for example, that a decline in the average wage represents a positive outcome if 1) those firms and workers who were already in the industry are experiencing wage growth, and 2) the industry overall expands to pull in workers from subsistence agriculture.  A declining average wage can signify the second factor is dominating the first factor, but this might be cause for celebration by policy makers rather than a source of concern.

Another reason to prefer Ancova: dealing with changes in measurement between baseline and follow-up

David McKenzie's picture
Private capital flows to developing countries eased in October, but remain close to their highest level in more than a year, led by robust bond issuance by emerging market sovereigns and firms. Business sentiment has strengthened in some countries, including the US and several emerging markets, but remains weak in general amid US “fiscal cliff” and Euro Area risks.

Why is Difference-in-Difference Estimation Still so Popular in Experimental Analysis?

Berk Ozler's picture
“I started working from when I was 8. When father could not send me to school, I decided to do something to support my family financially. Today, I am working as a cook in this tea shop.” - Puran Saud, Achham (Photo Credit – Stories of Nepal)

Goma is a girl, born in rural Kalikot. Her parents are illiterate, belong to the Dalit community and are in the bottom 20 percent of Nepal’s wealth distribution. Champa is also a girl born to a household very similar to Goma’s, but her parents are from a village in Siraha. Avidit is a boy born to an upper caste household in urban Kathmandu. Both his parents have a university education and come from affluent backgrounds.

In a society where opportunities are equally available for children of all socio-economic backgrounds, Goma, Avidit and Champa would all have equal odds of becoming doctors, or engineers or successful entrepreneurs. But in Nepal, the life trajectory of these children begins to diverge very early in life.

The often (unspoken) assumptions behind the difference-in-difference estimator in practice

Jed Friedman's picture
Worker in Ghana
For the poor and vulnerable of the world, jobs are key to ending poverty and driving development. But not all jobs are equally transformational.  
Photo: Jonathan Ernst / World Bank

Jobs are what we earn, what we do, and sometimes even who we are. For the poor and vulnerable of the world, jobs are key to ending poverty and driving development. But not all jobs are equally transformational. Good jobs add value to society, taking into account the benefits they have on the people who hold them, and the potential spillover effects on others. For example, inclusive jobs, such as those that employ women, can change the way families spend money and invest in the education and health of children.