Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

Disappointing Findings on Anti-Poverty Cash Transfers in the USA

straighttalkonevidence.org

Post reply on HN
Print view View original

141–149 of 149 posts

Re: Disappointing Findings on Anti-Poverty Cash Transfers in the USA

#141

Post by olympus »

gamblor956 wrote: post #136
olympus wrote: post #39

Earlier quoted context omitted.

While it sounds low, this is almost the same efficiency that a temporary work program for homeless people in Denver had (which initially caused me to look into the low sounding efficiency number). Think of it like this: if all of the remaining 52% went to paying a NYC social worker's average salary of $73.9k (source below), then each social worker has to manage 10.5 families to pay their salary. That's managing ten f…

Employees do not actually cost 2x their salary in benefits, payroll taxes, etc. The 2x includes office space and the associated costs of running the office, and a Silicon Valley markup for other benefits like free food and luxury shuttle service. For most companies, an employee's total cost 1.25-1.5x of salary. Most government agencies generally own the land and buildings in which the agencies are located (excepting…

Sure, but that was just a thought experiment to show that ~50% efficiency isn't some insanely low number. If you wanted you could ask for their records and dig in, and I bet most people would find each individual expense is reasonable. Then you add it all up and are somehow still surprised at how much running a large scale operation really costs.

And sure, we can delete the cost of office rent for government workers, but you have to add in the nice benefits package and you have to add in all the mileage they get reimbursed for driving their cars, and the fact that a considerable portion of their time is lost in training and meetings that don't directly support accomplishing the CCT's core mission. None of those things were included in my thought experiment, and erode the efficiency a little bit at a time.

Source: Am statistician employed by the government.

Re: Disappointing Findings on Anti-Poverty Cash Transfers in the USA

#142

Post by olympus »

sharemywin wrote: post #70
olympus wrote: post #39

Earlier quoted context omitted.

While it sounds low, this is almost the same efficiency that a temporary work program for homeless people in Denver had (which initially caused me to look into the low sounding efficiency number). Think of it like this: if all of the remaining 52% went to paying a NYC social worker's average salary of $73.9k (source below), then each social worker has to manage 10.5 families to pay their salary. That's managing ten f…

but the "cash" assistance was all about lowering costs. I get you've got costs of measuring the success rates etc, but I think the whole concepts needs to be rethought if they can't get that number much lower.

Most people underestimate the inefficiency of just about every government organization. Probably the only way to significantly improve the program's efficiency would be to turn it over to a for-profit company that had an incentive structure to distribute the cash properly but as efficiently as possible. So far I don't think that company exists yet. Pay for performance contracts get twisted by the contractors into a game to maximize the performance metrics while spending as little as possible. The problem is that it's really hard (some say impossible) to set up performance metrics that can't be gamed.

Re: Disappointing Findings on Anti-Poverty Cash Transfers in the USA

#143

Post by classics2 »

btilly wrote: post #4

I think that the problem is simple. People respond poorly to long-term rewards. You need to set up a short-term reward payoff cycle. https://buildingpharmabrands.com/2013/05/27/the-ad-that-crea... is inspiring that way. It shows how people who could not be motivated to solve a well-known long-term problem, tooth decay, were successfully motivated to solve a short-term problem that they previously ignored - the fact t…

The actual problem here is giving someone $180/mo in one of the most expensive cities in the country is worthless. It failed to change anything simply because $180 isn’t ever going to change anything for someone who’s improverished and living in city where basic rent is 10x that amount.

Re: Disappointing Findings on Anti-Poverty Cash Transfers in the USA

#144

Post by ctrl-j »

Para2016 wrote: post #138
ctrl-j wrote: post #135

Earlier quoted context omitted.

So, just to be clear, it sounds you're arguing that we should put barriers on how good people who are homeless/broke should be able to spend their financial assistance - in order to prevent addicts from buying drugs. That sounds like conflating the issue of poverty with the issue of drug abuse. There is overlap, but I don't believe that the harm to addicts outweighs the help to the people struggling to survive.

I said for addicts. It's pretty fuckin clear.

Ok, if you can outline a plan that gives non-addicts free reign over their assistance money, but restricts addicts, then we're cool.

Re: Disappointing Findings on Anti-Poverty Cash Transfers in the USA

#145

Post by setgree »

r00fus wrote: post #94
setgree wrote: post #68

Earlier quoted context omitted.

compliance and monitoring, most likely. The cash was conditioned on various attendance metrics, and the data collection process with such things is never as smooth as you'd like (I did a stint at a non-profit that ran experiments like this on behalf of economists).

I wonder if it'd be more successful if that overhead were dropped to say 10-20% (via automation and simplification of qualifications)

Regarding whether the effect would have been larger if the subjects got more cash -- in the original Progresa study, the ` size of the cash transfer is large, corresponding to about a 25 percent average increase in income of households living in extreme poverty' [0] (p.4). (Though in dollar amounts, if they are using the 'extreme poverty' metric of $1.25/day/person, we're talking about a lot less money in absolute terms). So there's something to this.

But this is hard to square away with this newest study's findings of modest adverse impacts on labor participation rates. This, combined with null health effects, suggests that the CCTs are performing an income substitution effect, and that what people are doing with their newfound free-time is probably leisure rather than other forms of 'work' (i.e. exercising, cooking healthier meals, etc.) that policymakers might wish the subjects undertake.

[0] https://www.povertyactionlab.org/sites/default/files/publica...

Re: Disappointing Findings on Anti-Poverty Cash Transfers in the USA

#146

Post by wutbrodo »

brudgers wrote: post #137
wutbrodo wrote: post #24

Earlier quoted context omitted.

Opportunity NYC was the previous program. The current one is called Family Rewards 2.0 > On average, the program cost $13,459 per family, 48 percent of which was paid directly to families as cash rewards. This was over a three year period, which means that each family received 13459*.48/3 = $2150, or about $180/mo. So they provided between 2-5x as much money. I share your confusion though, as this is still categorica…

$2k a year without strings attached is non-trivial. Tying the money to activities that may come with direct or indirect costs may make earning the money non-rational. At the granularity of paychecks, it's less than $45 a week. That's about four hours overtime at minimum wage. Or four hours of part time work at $11/hour. The value proposition of extra work is pretty straight forward. There's no bureaucratic process th…

> $2k a year without strings attached is non-trivial. Tying the money to activities that may come with direct or indirect costs may make earning the money non-rational. At the granularity of paychecks, it's less than $45 a week. That's about four hours overtime at minimum wage. Or four hours of part time work at $11/hour.

Sure, but we're talking about things that most people would agree is good anyway: kids doing well in school, being healthy, etc. The goal of the CCTs isn't to make people want to do things that they vehemently disagree with: it's to shift the incentive boundary so that they're on the right side of it for things they would like to do anyway.

Re: Disappointing Findings on Anti-Poverty Cash Transfers in the USA

#147

Post by evolve2k »

brudgers wrote: post #137
wutbrodo wrote: post #24

Earlier quoted context omitted.

Opportunity NYC was the previous program. The current one is called Family Rewards 2.0 > On average, the program cost $13,459 per family, 48 percent of which was paid directly to families as cash rewards. This was over a three year period, which means that each family received 13459*.48/3 = $2150, or about $180/mo. So they provided between 2-5x as much money. I share your confusion though, as this is still categorica…

$2k a year without strings attached is non-trivial. Tying the money to activities that may come with direct or indirect costs may make earning the money non-rational. At the granularity of paychecks, it's less than $45 a week. That's about four hours overtime at minimum wage. Or four hours of part time work at $11/hour. The value proposition of extra work is pretty straight forward. There's no bureaucratic process th…

I'm really surprised now that consideration of the 'quantum of funds' was not mentioned at all in the report, even as a footnote.

On face value, I would argue that $45/week extra although appreciated is not significant.

As a thought experiment it's in the scale of, 1hr week spend more time help kids with homework, 1hr spend additional time job searching, 1hr deal with program administrative/reporting requirements.

If a family say rewards themselves with a night out once a week for a dinner or movies etc, these funds are quickly gone.

I think the scale here implies an incentive based around addressing a lack of will e.g. "Well if they just had a little more encouragement they'd spend more time helping their kids with education and seeking employment and good health".

At this scale it does not address what I'll call 'incentive based on providing significant additional means for positive risk taking behaviour and structural optimization'.

I define structural behavior as when a family makes significant changes to their own setup towards have an effect; like moving closer to the school, hiring a tutor, pursuing course further studies part time (2-3 days a week).

Although the research process seems robust I agree the low quantum of funds discounts the greater significance of the findings.

TLDR; Summary of findings: 'When we gave poor familes enough weekly funds to go out for dinner once a week alongside a bunch of criteria to seek improvement in their life, it made little difference'

Re: Disappointing Findings on Anti-Poverty Cash Transfers in the USA

#148

Post by danieltillett »

valar_m wrote: post #102
danieltillett wrote: post #87

Earlier quoted context omitted.

Science and technology. Has been the only thing that has gotten the world out of the default historical state of 99% poverty.

Sounds like you're answering "What strategy may work to solve poverty?" or something close to that. OP said there are "known good solutions for poverty that work reliably and consistently well". I'm interested to hear what is known to work, not what might work.

No science and technology are the only things that has been shown to work. Since the industrial revolution the same playbook has been applied successfully in country after country. Look at China. It was not until they allowed science and technology to be used to its full capacity did they drag the vast majority of people out of grinding poverty.

The parts of the world where poverty is still high are those where science and technology have not yet arrived.

Re: Disappointing Findings on Anti-Poverty Cash Transfers in the USA

#149

Post by dang »

yasp wrote: post #127
tootie wrote: post #121

Earlier quoted context omitted.

That's not really how it works, but I also didn't comment on it's effectiveness, just that it's not anti-poverty. If you're rich and a minority you get the same advantage. The test for eligibility isn't means-based. Hence, not anti-poverty. The intent is to get more minorities into colleges and into jobs. You can argue whether or not that's fair, but it probably works pretty well as intended.

> Even With Affirmative Action, Blacks and Hispanics Are More Underrepresented at Top Colleges Than 35 Years Ago https://www.nytimes.com/interactive/2017/08/24/us/affirmativ... Maybe it works as intended as long as you're trying to maximize enrollment in "any colleges", not "best colleges".

It looks like this account is using HN primarily for ideological battle. We ban accounts that do that, as I've explained at length here:

https://news.ycombinator.com/item?id=16402648

https://news.ycombinator.com/item?id=16402618

https://news.ycombinator.com/item?id=16185062

... and plenty more at https://hn.algolia.com/?sort=byDate&dateRange=all&type=comme... for anyone who wants it.

Would you please (re-)read the site rules at https://news.ycombinator.com/newsguidelines.html and use HN as intended—that means for intellectual curiosity—from now on?

Post reply on HN