Open data and open source strategies for reducing poverty
Policy Engine
It's about data driving policy to alleviate poverty, and not exacerbate it
a lot of stats come out of a micro-simulation model.
Step 1: Idea is an individual household and all its characteristics (number of kids, ages, etc0) can computer benefits available to and taxes on the household. Could be eligible for food stamps, child tax credit, etc., use this to arrive at a net income. The house also has a poverty threshhold based on income. With the model, can recompute all this under a different policy or tax regime. Policy Engine is an open sourced rules mechanism.
Step 2: Pair with representative survey on geographic region you're looking at.
Poverty researchers mainly use same mechanism and by that, poverty fell recently during pandemic more than in any other recent years (at least in a long time).
Columbia Universtiy has one of best known programs for analyzing poverty reduction in US.
Current Population Survey is known to underestimate people in poverty, but it's better than an IRS survey that misses people who don't make enough to file taxes in the US.
SNAP is currently the biggest source of a benefit cliff where when you cross an income line you lose 100% of SNAP, leading to no incentive to increase earnings from 100% to 200% of poverty line, but after you hit 200% line there are incentives.
Only 78% of eligible houses use the Earned Income Tax Credit; Child Tax Credit is similar among people in poverty.
Lifeline, a broadband service for low income people, has a 20% take-up rate but its quality is also pretty bad. The micro-simulation model assumes a benefit recipient will continute to use it after its policy changes.
How do you think about policies to supporty a new policy so it doesn't create other problems.
Much of the micro-credit world has moved to unconditional cash transfers because loans really only work for entreprenurial people because you have to pay it back, potentially comes with more tax burden.
A study in Africa of giving $1k to poorest people they can find (like living on $1/day), and others $0 and comparing. Seeing that the people who get the money some end up making more money on their own long-term.
Economic stability can lead to people being willing to challenge the problems of the status quo. Defunding education in the US for 40 years after the turmoil of the '60s is counteracting this.
Early childhood investment is quantifiable in terms of long-term benefits to the children's lives.
Benchmark studies have three arms: one group gets money, one group gets nothing, and third group gets some other kind of intervention of similar value. One that looked at nutrtion found that straight-up cash worked better. Was the nutrition education culturally relevant?
Universal Basic Income (UBI) program in Los Angeles asks about your sexual assault history, making it antagonistic and many people self-select out.
Looking at more intersectionality, like health outcomes and health equity, is desired to be happening more.
For big national policies, Congressional Budget Office and Joint Committee on Taxation, look at 10 years projection. The first year for infants being listed out of poverty has biggest impacts. $1k in first year of life corresponded to 1% revenue increase in life, in US.
Policy Engine developing model to look at state, local, and federal taxes all in one.
Abundant Earths Project looking at maternity outcomes based on policies giving money to black pregnant women.
https://github.com/policyengine
US rules engine repo is policyengine-us. UK model is most complete.
Want to see benefits and tax offices using this tool.
Outreach is a lot of Twitter use. Group for Guaranteed Income using the API to help people understand outcomes of UBI program. Think tanks. Policy analysts are where the group comes from. Mostly a volunteer effort.