Our new paper on judicial bias in India’s lower courts, using new data from 6 million criminal court cases.

Do Muslims or women get better outcomes when their cases are heard by Muslim / female judges?

Surprise: No evidence on bias! 1/11

http://www.devdatalab.org/judicial-bias 

This is work by @ellliottt @thesamasher @aditibhowmick Daniel Chen Tanaya Devi Christoph Goessman @bilalsiddiqi.
2/11

We use standard approaches in the literature to estimate in-group bias.

Arbitrarily assignment of cases to judges = as-good-as-random judge identity
See how case outcomes change when judge staff changes change the gender / religion balance of the court
3/11

India’s judiciary is decisively male and non-Muslim. If there is ingroup bias in judicial decisions, that will lead to unjust outcomes for women and Muslims. 4/11

The randomized judge assignment results: extremely low and statistically significant levels of bias.

In-group bias effect for religion is 0.002 percentage points, similar for gender. 5/11

Strategy 2: RD. Do Muslim or non-Muslim acquittal rates change when the judges’ bench becomes more non-Muslim?

Answer: not at all.

(Aside, we do find acquittal rates fall briefly after *all* transitions — but it has nothing to do with judge or defendant gender/religion) 6/11

Why is this surprising? Because this kind of in-group bias is found in courts around the world, over and over and over.

We searched for all other studies using our identification strategy and looking at in-group bias among judges. Here’s how they line up: 7/11

But there may be a publication bias concern with these prior studies. In the absence of publication bias, we should see no correlation between effect sizes and standard errors.

Here is a graph of the effect magnitudes and errors in prior studies: 8/11

Note the brown line in that graph is the threshold for statistical significance.

There is no reason that actual effect sizes should be above or below that line. 9/11

For more info, check out the paper description and paper here: http://www.devdatalab.org/judicial-bias 

Or our oped in the Hindustan Times: https://www.hindustantimes.com/opinion/in-india-s-lower-judiciary-the-absence-of-in-group-bias-101610378365396.html

Fin 11/11

An important note: this is *one* test of bias in *one* step of the judicial process!

There is lots of evidence the judicial system is less than fair to Muslims and to women! Don't take this study to mean there is no bias in India!! 12/N

We find in other work that Muslims are the least upwardly mobile group in India: http://paulnovosad.com/pdf/anr-india-mobility.pdf

The 2006 Sachar Report is still the seminal piece on the economic experience of being Muslim in India— it's not great!
https://www.prsindia.org/administrator/uploads/general/1242304423~~Summary%20of%20Sachar%20Committee%20Report.pdf