iiinova — where your digital self keeps evolving

The transfer doesn’t seem to have affected births in either linked admin records or in our survey measures. So much for kids being “normal”!

130

Linking to admin records on college going for the kids, the point estimates suggest positive effects but most of the kids are just too young for this to be conclusive. We’ll just have to wait..

150

Statewide standardized math test scores may have somewhat worsened for the treated kids, although the effect is not significant after adjusting for multiple hypothesis testing and we see no similar effect for English.

271

We linked kids to administrative records from schools. We don’t find much of a change in these outcomes like attendance or repeating a grade.

250

We don’t find changes in many other outcomes we might have thought would have improved—like the home or neighborhood environment, children’s food security, or parental satisfaction.

161

But! Parents in the treatment group also reported that their kids had more stress and developmental difficulties. Possibly because they were monitoring them more closely and picked up on problems they might not have noticed.

171

Parents also spent more $ on their kids, especially those with low baseline incomes. Among these parents, the use of and quality of non-parental care (like daycare) also increased.

170

Treated parents reported better parenting behaviors, with the biggest improvements among the parents with the lowest baseline incomes; monitoring and supervision improved and reports of using corporal punishment fell.

190

J-PAL North America @j-palna.bsky.social · 03/03/2025

1000 low income adults were randomly selected to receive $1000/month for 3 years, with a control group receiving $50/month over that same period. Many of them had children in the household. How did it affect how they parented and their kids’ outcomes? www.nber.org/papers/w34040

Abstract for NBER WP 34040, “The Impact of Unconditional Cash Transfers on Parenting and Children”

210138

Reposted by Sarah Miller

1/4 🗓️Save the date! On March 18, from 3:30-5pm ET we are hosting a webinar featuring 3 #RCTs that evaluated the impact of cash transfers on health. Read this 🧵 to learn more about the research and register for the event (www.povertyactionlab.org/event/unders...

Alt text: Featured Event: Understanding the Impacts of Cash Transfers on Health Outcomes | March 18, 2025 | 3:30-5:00pm ET | Live over Zoom. Left hand side: Photo of a woman holding a credit card and phone while smiling.

122

Reposted by Sarah Miller

Grant McDermott @gmcd.bsky.social · 28/02/2025

What can you say, really? Just an endless series of disgraceful lows.

0272

Sarah Miller @smiller.bsky.social · 24/01/2025

I’m sorry to hear this. Unfortunately this is not the first example I’ve heard of…

051

Reposted by Sarah Miller

Eva Vivalt @evavivalt.bsky.social · 21/01/2025

We recently added several analyses to our paper on the employment effects of a $1000/month guaranteed income. 🚨 These new analyses include, for the first time, administrative data on income and employment and results for marriage/ divorce and benefits, among other outcomes. Read on for results! 1/

26923

Sarah Miller @smiller.bsky.social · 13/01/2025

Amazing work—congratulations! Wow!!

140

Sarah Miller @smiller.bsky.social · 25/12/2024

Is there a more millennial Christmas album? Merry Christmas!

1140

Sarah Miller @smiller.bsky.social · 19/12/2024

Wow, huge get!! Massive! Congrats to Brown!

040

public.websites.umich.edu

Link to paper: public.websites.umich.edu/~mille/OHIE_...

020

In the population we study, it doesn’t seem to reduce criminal justice involvement, which of course is the product of many complicated social, economic, and personal factors.

120

We know Medicaid does a lot of great things like reducing mortality and generating long term health and economic benefits among those enrolled as kids. No program can do everything for everyone.

100

One implication is that benefits of Medicaid on reduced criminal justice involvement may be concentrated on the highest risk groups (such as those returning from prison) where other work has found effects but we don’t have precision due to small N.

100

However, contrary to what we might expect, we do not find that those randomly assigned to Medicaid have lower rates of charges and convictions than those who did not receive Medicaid. Our estimates are fairly precise and can rule out most effects estimated in other settings/populations.

100

In 2008, Oregon allocated Medicaid eligibility to low-income adult applicants via a lottery, giving the chance to estimate the causal effect of Medicaid on outcomes. We link applicants to the lottery to admin data on criminal charges and convictions from the state of Oregon.

100

It also improves financial outcomes which could in turn reduce the need to turn to income generating crimes out of economic necessity.

100

Why might Medicaid have an effect? Other work shows Medicaid improves access to treatment for substance use disorder and mental health—both are massively over represented among those involved with the CJ system so we might hope improvement on these dimensions could reduce crime.

100

Does Medicaid reduce involvement with the criminal justice system? In this new paper we find no effect of Medicaid on criminal charges and convictions using experimental variation in Medicaid coverage from the Oregon Health Insurance Experiment.

1194

Possibly, but once you start incorporating benefit phase outs etc marginal tax rates can be really quite high at the lower end of the income distribution!

120

Congratulations to the team on getting these results out! I’m looking forward to reading this

010

I’ve been looking forward to seeing these results! The evidence base on this topic is rapidly growing.

1121

1. IRB would find it coercive; 2. Conditioning the money on surveys means you are paying participants for a service. In this case, the money would be taxable. Not so if it’s given unconditionally.

230

Thanks for your interest in the study!

020

Hi! We are looking at how effects evolve after the program ends but mostly for labor market and consumption outcomes not the political outcomes. I think those we will only have info about from during the intervention period.

140

They were not aware they were in the control group during the study (at the end we provided an optional de-brief)

190