***New published paper***

In PLOSOne, @DorianAbbot & I:

- extend the limits to economic growth debate
- show the waste heat limit to growth is perhaps a 1000 years away

Our audience is scientists (to whom we make economic args) & economists (to whom we explain the physics arg)

We focus on waste heat bec a recent Tom Murphy article in @Nature Physics suggested heat from power generation would bake the earth in 400y.

We show this ignores productivity growth (in use of energy) & slowing population.

Critically, we show these projections depend on a number of parameters & slight changes in those generate huge variation in predicted date of heat death.

We make this argument more systematic by connecting a model of waste heat with a model of growth & innovation, & data.


A really neat result is that there are 2 ways to massively delay the waste heat limit:

1/ Malthusian strategy: show population growth
2/ Neo-classical strategy: grow so fast that you innovate around the limit

While our focus is waste heat, we also note that our args apply to other resource limits. In particular, we argue that the massive changes in capital factor shares (ratio of resources used as inputs) demonstrates humankind's ability to shift resources based on price (scarcity).

That said, all is not hunky dory.

Scientists are right that there is a limit to *population* growth. With 1% pop growth, the mass of humans would exceed the mass of Earth in ~3000y, of the universe in ~9000y!

So we probably need to slow population growth or get to space well before the Sun expands! @elonmusk @cremieuxrecueil @UChicago @UChicagoLaw @BeckerFriedman 

You can read the unrolled version of this thread here: https://typefully.com/anup_malani/ctwpBtF

Extending Limits: Exploring the Waste Heat Barrier to Economic Growth

Global fertility is declining. Many say that modern culture is the cause: it lowers the status of family. But good evidence is scarce.

Among the most compelling ev may be from Brazil, where a soap opera that emphasized social mobility led to a 5% decline in the fertility rate.

La Ferrera, Chong & Durya examine the rollout of Globo, an 8pm soap opera. Using the rollout of broadcast antennas as an instrument for exposure to the soap opera, they calculate that exposure to the soap opera reduced fertility by 0.5 percentage pts in the 1980s.

They argue the effect is due not to TV generally, but Globo specifically because the soap opera led parents to name their kids after Globo characters. 

The mechanism is likely story lines that emphasized the importance of upward mobility. They observe the greatest reduction fertility in the year after a novela emphasizes economic advancement and for women similar in age to the leads in such novelas.


@_alice_evans @MoreBirths @Anne_red_head @BirthGauge @lymanstoneky 

The Impact of Soap Operas on Fertility

***New working paper alert***

With Ari Jacob, I present a new measure of the # of surviving children per female over time & across countries.

It shows slower decline globally than total fertility rate (TFR).
Show remarkable stability within countries over time.

@_alice_evans 

Available at https://www.nber.org/papers/w33175 and https://bfi.uchicago.edu/working-papers/effective-fertility-a-new-measure-of-surviving-children-to-analyze-fertility-decline/

TFR measures number of kids each female has. But not all kids survive.

We create a measure called effective fertility which is roughly:
TFR x probability of survival.

Survival is measured either through reproductive years (EFR-R) or theough working years (EFR-L).

@lymanstoneky

EFR-R can, conveniently, be approximated by existing demographic concepts like:

1/ Net reproductive rate (reported by UN)
2/ 2 x (Ratio of TFR to replacement level fertility)

Even though NRR is not designed to measure surviving kids.

Unfortunately, these don't measure EFR-L. 


What does EFR reveal?

Globally: 
- 1/3 of changes in TFR since 1950 simply compensate for decline in mortality rates, i.e., don't change EFR
- Of the change in EFR, 40% or more cannot be explained by economic factors like income


Historical European data since 1750 show:
- Before WW1 EFR is relatively stable, even during demog. transition
- EFR collapses after WW1, spike after WW2
- After baby boom, while EFRs fall below 2, they hold stead there

So pop growth is slowing, but not at an accelerating pace. 

Historical US data show:
- Although Blacks have much higher mortality & fertility, Black and White EFR are remarkably similar over time



Both European and US data show cycling of EFR with 20-30 yr periods. 

Consistent with Easterlin hypothesis that high wages lead to more births, which then expand labor supply and lower wages, leading to lower births in next generation, which elevates wages, leading to more births, and so on. 


Comments welcome!

People who might life this:
@ATabarrok @slatestarcodex @robinhanson @bryan_caplan @GarettJones @elonmusk @DorianAbbot @MoreBirths @Aaronal16 @salonium @lefineder @BirthGauge @kearney_melissa @WorksInProgMag @arpitrage @StatisticUrban @cremieuxrecueil 

In future work, we hope to calculate total human capital produced by females, total number of grandkids and of employed persons per female. 

We'll also explore the wild fluctuation in fertility around calamities like wars & famines, & estimate net impact on future population.  

You can read the unrolled version of this thread here: https://typefully.com/anup_malani/9mvow95

Effective Fertility: A New Measure

The worst thing that can happen to an urban slum is involuntary demolition. Poor people have their entire (meager) wealth wiped out.

I have a new essay on why this occurs. It's not primarily evil govts or developers. Instead, I argue that it's largely changing land use.

When the value of land for different uses changes, e.g., residential land becomes useful for commerce or a poor neighborhood gentrifies, the "normal" process for change is a developer buys the land from existing owners, demolishes existing structures, and build new ones.

But with urban slums--communities of squatters--existing residents have no formal rights to the land. So they can't sell the land be compensated for demolition. Instead, the developers goes to the govt & asks for an involuntary demolition. Slums are not compensated.

The crux of the problem is changing land use combined with lack of property rights.

Without changing land use, developers have no reason to demolish slums. That's why some slums are 50+ yrs old. With property rights, demolition would be voluntary, & not tragic.

I do not deny there are some demolitions that are arbitrary and capricious. By this I mean land is more valuable as a slum than however developers decide to use it. The lack of property rights allows that to occur: prices no longer signal relative value.  

But I suspect that demolitions that reduce value aren't as common as ppl think. Demolitions are costly, causing riots & bad press, esp. in democracies like India. If anything, land use changes probably happen too late, well after the point alternative land uses are higher value.

One should not be distracted by gentrification. That happens whether or not there are slums or poor renters. One can debate gentrification separately from demolition.

Giving slum residents property rights is not a simple solution to demolition. That involves trade-offs between informal owners and renters in slums which are hard problems. I'll write about those another time.

This is not a defense of demolitions, but at an attempt to understand when and why they occur.

Here is the link to my essay: https://open.substack.com/pub/anupmalani/p/why-do-urban-slums-get-demolished?r=i0cs&utm_campaign=post&utm_medium=web

Ppl who might find this take on slum demolition interesting (or would certainly be able to tell me why I'm wrong): @srajagopalan @ATabarrok @CafeEconomics @nebuer42 @profmohanan @adamschilton @adam_m_auerbach @tariqthachil @mattkahn1966

You can read the unrolled version of this thread here: https://typefully.com/anup_malani/Wz20gOz

The Impact of Changing Land Use on Urban Slums

I suspect that excessive public ownership of land is a substantial cause of the large population of slums in lower income countries.

Roughly 1/2 of slums are on (unused) public lands in places like India. If the govt gave up ownership, slum residents would not be squatters.

I don't deny that govt regulation of housing also cause slums. That reg drives down supply, increase price, pushing ppl into informal housing.

But we neglect the role excessive govt ownership of land plays. It reduces supply of land, an input into housing.

My point is a corollary to proposals to give squatters land rights. Note that land right laws mainly operate on govt land, because other land has private owners and is politically harder for the govt to (involuntarily) transfer to squatters.

My (implied) proposal for govts to give up ownership in some lands asks for more than land rights laws! Those laws tend to retain govt ownership, & giving squatters tenant rights on land (with 0 rent). But that stops squatters from selling the land, or borrowing against it.

I do not think govt should give up rights in all land--just land that squatter occupy or are likely to occupy in the vicinity of local cities. The option value of govt ownership there is small, at least relative to value to existing &  potential squatters, who we should think of as poor migrants to cities. 

I am not naive that govt abandonment of land rights would always cause squatters to become owners or eliminate low quality housing. But it is a useful step.

It would increase supply of land, an input into housing price. 

Yes, regs that limit construction would also get in the way. But if we relaxes regs, (artificially) low supply of private land would also raise housing price, increasing the cost of housing.

Some abandoned land would go to uses other than low-income housing. But not all.

Abandoned land won't eliminate low-quality housing. Squatting is driven by unmet demand for *low-cost*, low-quality housing by the poor. Abandonment will simply lower the price of such housing. It won't increase their incomes. They'll still demand low-quality housing.

But, like reducing regulations that limit housing supply, abandonment of public lands will allow the poor to afford more with the limited income they have.

@arpitrage @mattkahn1966 @gandhisahil @adamschilton @adam_m_auerbach @tariqthachil @nebuer42 @CafeEconomics

Excessive Public Ownership of Land and Slums

Question for economists about the Neolithic Demographic Transition (wherein population surged after agriculture, but morbidity & mortality rose, & height shrank):

Is this best understood as a decline in the price of fertility and so a substitution towards quantity of kids?

For data on the trade-off, see e.g. this paper:

https://www.pnas.org/doi/pdf/10.1073/pnas.1524031113

Humans chose quantity over quality.

One puzzle is that "incomes" seemingly fell after the Neolithic Revolution (i.e., calories per person fell). But that is not the right way to think about this.

Potential income rose. Ppl spent it on fertility so had less for non-fertility (personal) consumption.

Natural question: Why did we do the opposite during the modern Demographic Transition, ie choose quality over quantity? Maybe because the return to quality rose over time. That was a function of technology.

@GalorOded 

One might ask, did potential income increase in the Neolithic Revolution. Kinda hard to argue it did not given that population doubled every 55 years for a few hundred years after NR. 

Neolithic Demographic Transition: Quantity Over Quality

@AlainBertaud2 highlights a puzzle about slums:

While slums are VERY DENSE in terms of people living per square foot of FLOORSPACE,

slums are NOT DENSE in terms of people per sqft of LAND

Why is that?

You can see what Alain means in this picture I took as I was flying out from Mumbai to Addis Ababa a few years ago. The areas with blue roofs are slums.

The answer is that the land slums are on is so low quality, it would not be economical for a developers to build a taller building there.

Usually it is low quality b/c it's prone to floods or landslides, or close to nuisance like garbage dumps.

Why aren't tall buildings located on low quality land? Because building taller building are expensive, and people value land quality. So for developers to cover their costs of building up, they need to be on high quality land that people are willing to pay for.

I explain more at why at https://anupmalani.substack.com/p/why-are-slums-not-dense?r=i0cs

Alain was on @srajagopalan's great Ideas of India podcast.

@arpitrage @adamschilton @mattkahn1966 @ATabarrok

You can read the unrolled version of this thread here: https://typefully.com/anup_malani/XwE6LPC

The Puzzle of Slum Density

*** Working Paper Alert ***

Insurance isn't the only way to manage financial shocks (eg medical bills).  You can also borrow & save. Account for credit access lowers the value of health insurance ~60%.

With @JulianReif @spjaffe

https://bfi.uchicago.edu/working-paper/access-to-credit-reduces-the-value-of-insurance/

WHY IS THIS IMPORTANT? 1) It explains low uptake into (health) insurance, even when it is subsidized. This could be because ppl have access to loans. 2) It explain the relationship b/w policies like the ACA & EMTALA. EMTALA gives 0-interest loans, reducing the value of the ACA!

MOTIVATION. Models that value insurance calculate willingness to pay (WTP) for contracts to smooth consumption in the face of financial shocks (eg medical or auto repair bills). But, these models usu assume insurance is the ONLY way to smooth consumption. That's false.

MODEL. We present a lifetime model w/ shocks that allows ppl to choose between insurance, borrowing & saving to smooth consumption. We provide exact formulas (if you have claims data) & approximate formulas (if you have aggregate data). This model permits markups & interest rts. 

APPLICATION. Although the model is generic for all insurance, we use MEPS data to calculate the value of *health* insurance. We consider community & experience rated insurance, by risk, income, age, insurance markup, & interest rate.

RESULT 1: Access to loans reduces value of insurance ~60% on average across decile of predicted medical spending, regardless of pricing. 

RESULT 2: Markups reduce value of insurance a lot.  For experience rated policies, markups + loans makes insurance have non-positive value!

RESULT 3: Access to credit reduces the value of insurance most for the poorest individuals. The reasons is that diminishing MU of income means insurance is most valuable for the poor. But also the value of loans is most valuable for the poor too!


RESULT 4: Higher interest rates *increase* the value of insurance! Because higher rates lower the value of the substitute. 

HOWEVER, even at very high interest rate, loans reduce the value of health by ~25%! 


Comments/feedback appreciated!

@GottliebEcon @nealemahoney @ProfNoto @jacobsgoldin @UChicagoLaw @BeckerFriedman @alz_zyd_ 
@arpitrage @JohnMullahy


You can read the unrolled version of this thread here: https://typefully.com/anup_malani/pMgf2We

Access to Credit Reduces Value of Insurance

***New paper ALERT - Nature Scientific Reports***

We conducted a population-level cellular immunity survey for COVID from Bangalore, India.

We find that humoral immunity--measured with sero-surveys--substantially **underestimates** cellular immunity!

https://rdcu.be/dAEdy

TLDR:

* Only 16% of population had neutralizing antibodies, 30% for IgG antibodies to spike protein, but 46% had cellular immunity!

* Cellular immunity slightly higher in slums

This is only one of 2 population-level cellular immunity surveys done during COVID!

*** Why is this paper useful? ***

* Antibodies (humoral immunity or HI) are not the only protection against COVID. You also have cellular immunity (CI), which can last longer. HI underestimates CI!

* Our study shows how to conduct population-level cellular immunity surveys.

*** (Lay-person's) explainer on immunity ***

When you're first infected, your body cycles thru different antibodies to see which ones "attach" to an invading pathogen. This takes time, allowing the pathogen to multiply. If you don't match fast enough, you fall ill, maybe die.

If you do match & your immune system catches up, you survive. After the pathogen is defeated, the attaching antibodies stay in your blood for some time. Like soldiers after a war. They provide some protection if the pathogen invades again. You're already mobilized.

If you're not re-infected quickly, you stop producing specific antibodies & existing ones are cleared from your blood. Maintaining a large army during peacetime is too costly!

This is why humoral (antibody) immunity declines. Often in just a few months. https://www.nature.com/articles/s41598-023-50338-3

But T-cells remember what antibodies attach in case the pathogen returns.

Next time you're infected, you T-cells both tell B-cells what antibodies will attach & they coordinate attacks on the pathogen.

This is cellular immunity! It can last longer than humoral immunity.

*** Our contribution ***

During COVID, millions took antigen tests. They only tell if you're infected NOW.

Researchers did tons of sero-surveys. They only show humoral immunity.

Some did lab cellular immunity studies. They don't reveal cellular immunity in the community!

This paper is one of just 2 population-level cellular immunity surveys! The other was done among young adults in Sweden.

Our study randomly selected ~2400 people from slums & non-slum communities in Bangalore in Feb 2021.

We conducted 2 humoral immunity tests on each person. We did a cellular immunity assay on 600. Syngene Intl developed the tests, with help from @andreasantlab.

This was fun statistics. We used ROC curves & picked thresholds that set sensitivity = specificity.

***Our key finding (again) ***

* Only 16% of population had neutralizing antibodies, 30% for IgG antibodies to spike protein, but 46% had cellular immunity!

* Cellular immunity slightly higher in slums



This was amazing work by @anuacharya, Syngene Intl (Jayashree Aiyar, Neha Kamran), @andreasantlab. Great help from @taneja_saloni, Wanran Zhao, @profmohanan.

You can read the unrolled version of this thread here: https://typefully.com/anup_malani/RnpUZng

New Findings: Cellular Immunity Surpasses Antibodies

**New paper in Nature Scientific Reports**

State of Tamil Nadu, India, conducted 4 of sero-surveys during COVID, each >25,000 people.

We find:
- Antibodies decline by 30-60% within 6 months.
- Both infection & vaccination contribute to seroprevalence.

https://www.nature.com/articles/s41598-023-50338-3

This study was a repeated cross-section on randomly-selection ppl across each of 30+ districts in Tamil Nadu state.

The survey ran from Oct 2020 - Jan 2022, spanning 2 COVID waves.

This shows it is feasible to do large scale, population representative surveys during pandemic.

Our study reveals much higher seroprevalence than reported infection rates -- 35 times larger!

Moreover, Tamil Nadu showed higher seroprevalence than most of India.

Antibodies wane rapidly!

Seroprevalence fell by 30-60% (depending on district) from Oct 2020 - April 2021, just 6 months.

Does that mean immunity fell?  Humoral immunity certainly did.  I'll report on cellular immunity in a future thread.


There is some evidence that humoral immunity falls after vaccination.  But more research is needed: our time series is not long enough to be definitive.


Both infection & vaccination contribute to humoral immunity!

We show how to attribute changes in seroprevalence to infection v. vaccination with data on individual seroprevalence & vaccination.

Up to 63% of increase in seropositivity from April - Nov 2021 was from infection during the Delta wave.

Up to 88% of increase in seropositivity from Nov 2021 - Jan 2022 was from vaccination.


We find that additional vaccine doses contribute to seroprevalence, but with effects diminishing in the # of doses.

Covishield vaccine is associated with greater population seroprevalence than Covaxin, consistent with prior studies.

Thanks to Directorate of Pub. Health & Preventive Medicine, Govt Tamil Nadu; Inst. of Community Medicine, Madras Med. College; & @Artha__Global for amazing efforts. It was great to work with TS Selvavinayagam, Sudharshini Subramaniam, @r_sabareesh, @VaidehiTandel, @imad_sofia.


You can read the unrolled version of this thread here: https://typefully.com/anup_malani/2AAUJYr