The Triumph of Humanity Chart

One of the greatest successes of mankind over the last few centuries has been the enormous amount of wealth that has been created. Once upon a time virtually everyone lived in grinding poverty; now, thanks to the forces of science, capitalism and total factor productivity, we produce enough to support a much larger population at a much higher standard of living.

EAs being a highly intellectual lot, our preferred form of ritual celebration is charts. The ordained chart for celebrating this triumph of our people is the Declining Share of People Living in Extreme Poverty Chart.

Share in Poverty

(Source)

However, as a heretic, I think this chart is a mistake. What is so great about reducing the share? We could achieve that by killing all the poor people, but that would not be a good thing! Life is good, and poverty is not death; it is simply better for it to be rich.

As such, I think this is a much better chart. Here we show the world population. Those in extreme poverty are in purple – not red, for their existence is not bad. Those who the wheels of progress have lifted into wealth unbeknownst to our ancestors, on the other hand, are depicted in blue, rising triumphantly.

Triumph of Humanity2

Long may their rise continue.

 

Advertisements

The Importance of GWWC Cohort Data

There are a few pieces of information that are required to properly analyze the value of Giving What We Can‘s membership.

They’re necessary for GWWC’s managers to evaluate different strategies. If GWWC was an object-level charity, we wouldn’t donate to it without knowing these numbers. And if GWWC were a public company, investors would not provide funding without such disclosure. As such, hopefully these metrics are already being collected internally, and publicly sharing them should not be very difficult, though very valuable. If not, GWWC should start collecting them!

GWWC already publishes the number of members it has at any given point and the total amount pledged. From this it’s easy to derive how many joined in any given year. However, it’s hard to judge what these people did later – how many fulfilled the pledge, and how much did they donate? Worse, this makes it hard to forecast the value of a new member, so we can’t tell how much effort we should put into extensive growth. As far as I can see (sorry if I just couldn’t find the data), we do not currently release the data required to make this analysis.

As part of it’s annual report, GWWC should release data on each cohort: how many of that cohort fulfilled the pledge by donating 10%; how many were ‘excused’ from donating 10% ( e.g. by being students); how many failed to abide by the pledge, donating less than 10% despite having an income; and how many did not respond.

Example Disclosure

In case it’s confusing what exactly I’m suggesting GWWC release, here’s an example (with totally made-up numbers). As part of it’s 2014 annual report, GWWC could report:

  • 2011 cohort:
    • Of the 107 who joined in 2011…
    • 75 donated over 10% in 2014
    • 15 were students and did not donate 10% in 2014
    • 10 had incomes but did not donate 10% in 2014
    • 7 could not be contacted in 2014
    • Total of $450,000 donated in 2014
  • 2012 cohort:
    • Of the 107 who joined in 2012…
    • 50 donated over 10% in 2014
    • 53 were students and did not donate 10% in 2014
    • 2 had incomes but did not donate 10% in 2014
    • 2 could not be contacted in 2014
    • Total of $300,000 donated in 2014
  • 2013 cohort:
    • etc.

While in the 2013 annual report, GWWC would have reported

  • 2011 cohort:
    • Of the 107 who joined in 2011…
    • 45 donated over 10% in 2013
    • 56 were students and did not donate 10% in 2013
    • 3 had incomes but did not donate 10% in 2013
    • 3 could not be contacted in 2013
    • Total of $250,000 donated in 2013
  • 2012 cohort:
    • Of the 107 who joined in 2012…
    • 16 donated over 10% in 2013
    • 89 were students and did not donate 10% in 2013
    • 1 had incomes but did not donate 10% in 2013
    • 1 could not be contacted in 2013
    • Total of $100,000 donated in 2013

This would allow us to see how each cohort matures of time, answering some very important questions:

  • How much is a member worth, after taking into account the risk of non-fulfillment?
  • How much does the value of a member differ with the discount rate we use?
  • How does the donation profile of a member change over time – does it rise as they progress in their career or fall as members drop out?
  • Are the cohorts improving or deteriorating in quality? Are the members who joined in 2012 more likely to still be a member in good standing in 2014 than they 2010 cohort were in 2012? Do they donate more or less?

There are some other numbers that might be nice to know, for example breaking the data down by age, sex, nationality, or even CEA employee vs non-employee, but it’s important not to impose too high a reporting burden.

Why this is not idle speculation

This might seem a bit ambitious. Yes, it would be nice if GWWC released this data. But is it really a pressing issue?

I think it is.

Bank problems: Extend and Pretend

Sometimes banks will make a series of bad loans – loans which are repaid at a significantly lower than expected rate, perhaps because the bank was trying to grow aggressively. When the first signs of this emerge, like people being late on payments, banks have two alternatives. The honest one is to admit there is a problem and ‘write down’ the loan – take a loss to profits. The perhaps less honest one is to extend and pretend – give the borrowers more time to repay and pretend to yourself/auditors/investors that they will come good in the end. This doesn’t actually create any value; it just delays the day of reckoning. Worse, it propagates bad information in the meantime, causing people to make bad decisions.

Unfortunately they neglected the Litany of Gendlin:

What is true is already so.
Owning up to it doesn’t make it worse.
Not being open about it doesn’t make it go away.
And because it’s true, it is what is there to be interacted with.
Anything untrue isn’t there to be lived.
People can stand what is true,
for they are already enduring it.
—Eugene Gendlin

GWWC: Dilutive Growth?

About a year ago, people were concerned that GWWC’s growth was slowing – only growing linearly, rather than exponentially. This would be pretty bad, and people were justifiably concerned. However, GWWC made a few changes with the aim of promoting growth. Most pertinently:

  • Allowing people to sign up online, rather than having to mail in a hand-signed paper form. This happened between April and June 2013.
  • Adjusting the pledge to become more cause-neutral, rather than just about global poverty. This happened late 2014.

GWWC signups labeled

source

The latter change was somewhat controversial, but I didn’t see much discussion of the former at the time.

The concern is that, though these measures have increased the number of members, they may have done so by reducing the average quality of members. Making it easier to join means more marginal people, with less attachment to the idea, can join. This is still good if their membership adds value, but they dilute the membership, which means we shouldn’t account for the average new member being signed up now as being equally valuable as the members who joined up in 2010. Additionally, the reduction in pomp and circumstance might reduce the gravitas of the pledge, making people take it less seriously and increase drop-out rates. If so, moving to paperless pledges might have reduced the value of sub-marginal members as well as diluting them.

The comparison with banks should be pretty clear – a bank that’s struggling to grow starts accepting less creditworthy applicants so it can keep putting up good short term numbers, but at the cost of reducing the long-run profitability. Similarly GWWC, struggling to grow, starts accepting lower quality members so it can keep putting up good short term numbers, but at the cost of reducing the long-run donations. This makes it harder to forecast the value of members, and might lead to over-investment in acquiring new ones.

This seems potentially a big risk, and it’s the sort of issue that this data would allow us to address. Of course, there are many other applications of the data as well.

And GWWC in fact has even stronger reasons than banks to report this data. The bank might be wary of giving information to its competitors, but GWWC has no such concerns. Indeed, if releasing more data makes it easier for someone else to launch a competing, better version of GWWC, all the better!

If you liked this you might also like: Happy 5th Birthday, Giving What We Can and GiveWell is not an Index Fund

Betting the House, not the Economy

Òscar Jordà, Moritz Schularick and Alan Taylor recently put out an interesting paper on the relationship between interest rates, the housing market and financial crises. It’s been reviewed very positively, for example here and here. They take advantage of the macroeconomic trilemma – countries with fixed exchange rates and open capital markets ‘import’ the monetary policy of other countries –  by treating these imported interest rates as exogenous, providing them with an independent random variable.

People seem to be interpreting it as saying that loose monetary policy causes financial instability through the transmission mechanism of a housing bubble.

To be fair, the opening line of the abstract does give that impression:

Is there a link between loose monetary conditions, credit growth, house price booms, and financial instability?

as does the first paragraph’s use of ‘thus’:

Do low interest rates cause households to lever up on mortgages and bid up house prices, thus increasing the risk of financial crisis?

So one could be forgiven for thinking their conclusion was basically

  • Loose Monetary Policy -> Mortgage & Housing Bubble -> Financial Instability

… Well, one could be forgiven for thinking that unless one was going to write about the article. If you were going to do that, you should actually read the article. Then you would realize they in fact show two separate things:

  • Loose Monetary Policy -> Lower Long-term interest rates + Mortgage & Housing Increases
    • (with reasonable p-values in general)

InterestRatesAndHousingMarket

and

  • Mortgage & Housing Market Bubble -> Financial Instability
    • (with ok p-values, and some suspicion about how canonical their independent variable was)

HousingMarketAndFinancialCrisis

Despite having a clever way of treating monetary policy as an independent variable, they never directly test

  • Loose Monetary Policy -> Financial Instability

even though this would be a major victory for the ‘low interest rates caused the bubble and crisis’ crowd.

Why not test this directly? The authors don’t say, but I suspect it’s because the test would fail to yield significant results. Absence of evidence of such a connection is evidence of its absence. And looking at the significance levels of the two results they did provide, I suspect that combining them would cease to be significant (unless their is another, parallel causal mechanism).

Which is a shame! Their independent variable looked really cool, as did their data set.

I think there’s an underlying theoretical reason to not expect it to work, however (quite apart from nothing ever working in macroeconomics. They rightly make much of their finding that exogenous low interest rates cause increases in housing prices. But this is not necessarily caused by increased demand ‘bidding up’ the value of housing in a bubble.

Rather, consider what sort of an asset housing is. Houses allow you to avoid paying rent in the future; their value is the capitalized value of avoided future rent. When interest rates are low, those future rent payments are discounted at a lower rate, so are more valuable: low interest rates increase the inherent value of housing. House prices rising when rates are low isn’t a bubble unless the interest rates themselves are a bubble; it’s rational cross-asset pricing. So we should expect exogenous falls in interest rates to increase house prices.

But wait there’s more!

Exogenous falls in interest rates probably mean rates are now too low (from a Taylor Rule perspective, or similar), or at least not-as-excessively-high as before. This will tend to increase inflation. And as home ownership represents a hedge against rent inflation, higher inflation yet again increases the value of home-ownership. So once again we have a non-bubble based reason to expect exogenous falls in interest rates to increase house prices.

So we have two reasons to think that low interest rates should cause non-bubble increases in house prices, and journal article that is mildly supportive of this thesis.

Let he who is without Science Denial cast the first stone

The Washington Post recently ran an article on how political affiliation and level of religious belief affect support for, or suspicion of, the scientific consensus on various subjects. In it they refer to research by Dale Kahan to argue/imply that opposition to science is primarily driven by conservative ideology.

For example, they have these three very attractive charts, showing that the difference between people of high and low religiosity is small compared to the difference between conservatives and liberals when it comes to global warming,

GlobalWarming

evolution,

.
Evolution

and Stem Cell research:
StemCell

However, as so often happens, their article on causes of political bias ends up displaying some pretty impressive political bias. Unsurprisingly, this bias tends to be flattering towards those who share their political beliefs, and damning of those who don’t.

Firstly, look at those charts again. When looking at on the left-right axis, your eye is naturally drawn to compare the two extremes – to compare the most right wing to the most left wing (especially as the line is monotonic). You note the large difference in height between the leftmost data points and the rightmost, compare it to the relatively small difference between the high and low religiosity lines. The former difference is bigger than the latter difference, so political opinions must be more important than religious ones.

… or so the chart leads us to believe. However, this is hugely deceptive. As you can see, there are 5 tick marks on the horizontal axis, the measure was created from questions using 5 and 7 options, and there are a very large number of little vertical lines. This means they’re using a relatively fine measure of political ideology: they differentiate moderate conservatives from ordinary conservatives from highly conservative people. By doing this, they increase how extreme the extremes are, which increases that vertical difference our eye is naturally drawn to. With religion, however, they only admit of two categories, high and low. Perhaps if they had disambiguated more, so the categories ranged from “More religious than the Hasidim” to “More atheist than Dawkins”, we would have seen more spread between those two lines. As it is, the charts suppress these differences, reducing the apparent effect of religiosity.

That’s not the only problem with the article. The climate change and evolution questions seem pretty good, but the stem cell question does not show what they think it does.

“All in all, do you favor or oppose federal funding for embryonic stem cell research”

Now, in general opposing research for science does seem like prima facie evidence that you’re in some sense anti-science. But not here! There are two other factors at play which conflate the issue.

The first is that this is as much a moral issue as a scientific one. Thinking that stem cell research is immoral doesn’t necessarily mean you disagree with any of the scientific findings, due to the is-ought gap. In the same way that opposing nazi research on cancer (which used a variety of immoral techniques) doesn’t mean you think their conclusions were factually wrong, you can think stem cell research is morally wrong but the conclusions factually correct. Or, to use a clarifying contemporary example, suppose the question instead asked,

“All in all, do you favor or oppose federal funding for methods of treating homosexuality”

My intuition, which I suspect you share, is that the line would slope in the opposite direction – lefties would be more opposed than righties. This isn’t necessarily be because they are anti-science – maybe they simply think we are better off not knowing how to treat homosexuality, or better off not even thinking about the possibility. This moral belief doesn’t, however, mean they disagree with conservatives and scientists on any factual issue.

But there is another, even bigger, problem with this question. It doesn’t just ask about the morality of stem cell research – it asks about federal funding for that research. Conservatives are well known for opposing federal funding of things in general. Yet this research suggests that consistently applying the conservative rule “oppose federal funding of things in general” is suddenly evidence of being anti-science. You would be branded anti-science by this question even if your thought process was

“I think the federal government is very bad at research – it will be inefficiently run, overly politicized, and poorly directed – so I don’t want it to mess up stem cell research. Stem cell research is far too useful and exciting to trust to the government.”

Yet surely such a person should be considered pro-science, not anti-science!

Indeed, it seems that overlooking this issue, and conflating opposition to the state with opposition to science, is a clear sign of political bias on the part of the author. They choose a question which almost by design proved conservatives were anti-science, not by actually measuring the truth, but by simply re-defining opposition to science to include the political opinions they oppose. David Friedman once wrote about something similar – a study which, while claiming to prove that right-wing people were authoritarians, really just defined authoritarianism as ‘respects right-wing authorities’.

Ok, so their choice of data visualization technique was perhaps misleading, and the stem cell funding question was awful. But the other two questions look pretty solid, right?

Perhaps not. It’s well known – or at least widely believed – that conservatives disproportionately disbelieve in evolution and global warming. So if you wanted to prove that conservatives were anti-science, you’d pick those two questions, confident that your prejudices would be confirmed.

Yet there is much more to science than evolution and global warming. There many issues where there’s a scientific consensus at least as strong as that on global warming, yet some people still disagree. For example,

  1. Astrology is nonsense
  2. Lasers are **not** condensed sound waves
  3. The earth orbits the sun

In fact, I would say that science is far more unequivocal on these issues than on global warming – probably around as certain as that evolution is true.

Yet on all these issues, Republicans are more likely to hold the scientific view that Democrats. And there are many more similar examples. If I wanted to make the same charts, but make Democrats look bad, I could easily “prove” that Democrats are morons who believe the sun orbits the earth.

The Washington Post article does contains a homage to data:

But why opine on all this an un-grounded way — we need data.

Unfortunately we need more than data – we also need rigorous statistical techniques.

It would be unfair to blame the original researcher. In his article, he also includes a chart on nuclear power, where conservatives have the more scientific view. Mysteriously, the chart that was flattering to conservatives doesn’t make it into the Washington Post article. Ironically, it turns out the Washington Post article was right – politics really is the mindkiller. It’s just hard to spot when you’re the one getting killed.

Happy 5th Birthday, Giving What We Can!

Giving What We Can recently celebrated its 5th birthday. It’s not much of a party if no-one external congratulates you, so here we go: Happy Birthday, GWWC!

It’s pretty impressive how much GWWC has grown since those early days. Here’s a chart of total membership, which I’ve put together from GWWC emails and liberal use of the internet archive. I’m sure they have better data (without gaps!) internally, but I’ve never seen this chart before. Notably, growth seems to have picked up since the fall of 2013. Did GWWC change their strategy at that point? (or their membership-counting-methodology?)

Lines going up are always good

Putting the same chart on a log scale, we can see that GWWC have actually done a reasonably good job of sustaining exponential growth.

Lines that go up on a log scale are even better!

Fitting a line of best fit to the chart, I estimate GWWC’s membership is growing 73.1% a year. Assuming 2% population growth, it will take just 30.25 years before all the world’s population are GWWC members. Taking over the world by the time I’m 58 sounds like pretty good going!

Happy Birthday, Giving What We Can!

RCT as I say, not as I do

Randomized Controlled Trials (RCTs) are the gold standard in policy evaluation.

Say you’re investigating a third world development policy, like building schools, or installing water pumps, or distributing malaria-resistant bednets. A random sample of the villages in an area are selected to receive the policy. The other villages form the control group, and receive no special treatment. Metrics on various desiderata are recorded for each village, like income, lifespan and school attendance. By comparing these outcomes between villages with and without the intervention, we can judge whether it made a statistically significant difference.

RCTs give us strong evidence of a causal link between the intervention and the result – we assume there were no other systematic differences between the treatment and control villages, so we have good grounds for thinking the differences in outcome were due to the intervention.

This is a marked improvement over typical methods of evaluation. One such method is simply to not investigate results at all, because it seems obvious that the intervention is beneficial. But people’s intuitions are not very good at judging which interventions work. When Michael Kremer and Rachel Glennerster did a series of education RCTs in Kenya, all their best ideas turned out to be totally ineffective – plausible ideas like providing textbooks or teachers to schools had little impact. The one thing that did make a difference – deworming the children of intestinal worms – was not something you’d necessarily have expected to have the biggest impact on education. Our intuitions are not magic – there’s no clear reason to expect our to have evolved good intuitions into the effectiveness of developmental policies.

A common alternative is to give everyone the intervention, and see if outcomes improve. This doesn’t work either – outcomes might have improved for other reasons. Or, if outcomes deteriorated, maybe they would have been even worse without the intervention. Without RCTs, it’s very difficult to tell. Another alternative to RCTs is to compare outcomes for villages which had schools in the first place to those which didn’t, before you intervene at all, and see if the former have better outcomes. But then you can’t tell if there was a third factor that causes both schools and outcomes – maybe the richer villages could afford to build more schools.

The other main use of RCTs is in pharmaceuticals – companies that develop a new drug have to go through years of testing where they randomly assign the drug to some patients but not others, so we can be reasonably confident that the drug both achieves its aims and doesn’t cause harmful side effects.

One of the major criticisms of RCTs is that they are unfair, because you’re denying the benefits of the intervention to those in the control group. You could have given vaccinations to everyone, but instead you only gave them to half the people, thereby depriving the second half of the benefits. That’s horrible, so you should give everyone the treatment instead. This is a reasonably intelligent discussion of the issue.

But this is probably a mistake. Leaving aside the issue that it’s more expensive to give everyone the treatment than a subset (though RCTs do cost money to run), it’s a very static analysis. Perhaps in the short term giving everyone the best we have might produce the best expected results. But in the long term, we need to experiment to learn more about what works best. It is far better to apply the scientific method now and invest in knowledge that will be useful later than to cease progress on the issue.

Indeed, without doing so we could have little confidence that our actions were actually doing any good at all! Many interventions received huge amounts of funding, only for us to realize, years later, that they weren’t really achieving much. For example, for a while PlayPumps – children’s roundabouts that pumped drinking water – were all the rage, and millions of dollars raised, before people realized that they were expensive and inefficient. Worse, they didn’t even work as roundabouts, as the energy taken out of the system to pump the water meant they were no fun to play with.

Another excellent example of the importance of RCTs is Diacidem. Founded in 1965 by Lyndon Diacidem, it now spends $415 million a year, largely funded by the US government, on a variety of healthcare projects in the third world, where it deliberately targets the very poorest people. Given that total US foreign aid spending on healthcare is around $1,318 million, this is a very substantial program.

Diacidem have done RCTs. They did one with 3,958 people from 1974 to 1982, where they randomly treated some people but not others. The long time horizon and large sample size makes this an especially good study.

Unfortunately, they failed to find any improvement on nearly all of the metrics they used, and as they used a 5% confidence interval, you’d expect one to appear significant just by chance.

 “for the average participant, any true differences would be clinically and socially negligible… for the five general health measures, we could detect no significant positive effect… among participants who were judged to be at elevated risk [the intervention] has no detectable effect.

Even for those with low income and initial ill health, surely the easiest to help, they didn’t find any improvements in physical functioning, mental health, or their other metrics.

They did a second study in 2008, with 12,229 people, and the results were similar. People in the treatment groups got diagnosed and treated a lot more, but their actual health outcomes didn’t seem to improve at all. Perhaps most damningly,

“We did not detect a significant difference in the quality of life related to physical health or in self-reported levels of pain or happiness.”

Given that these two studies gave such negative results, you would expect there to be a lot more research on the effectiveness of Diacidem – if not simply closing it down. When there are highly cost-effective charities than can save lives with more funding, it is wrong to waste money on charities that don’t seem to really achieve anything instead. But there seems to be no will at all to do any further study. People like to feel like they’re doing good, and don’t like to have their charity criticized. Diacidem is political popular, so it’s probably here to stay.

Sound bad?

Unfortunately, things are far worse than that. Diacidem does not actually cost $415 million a year – in 2012, they spent over $415 billion, over 300 times as much as the US spends on healthcare aid. It wasn’t founded by Lyndon Diacidem, but by Lyndon Johnson (among others) Nor does it target the very poorest people in the third world – it targets people who are much better off than the average person in the third world.

The RCTs mentioned above are the RAND healthcare experiment and the Oregon healthcare experiment, with some good discussion here and here.

Oh, and it’s not actually called Diacidem – it’s called Medicaid.