Why we cosleep with our infant and you (perhaps) should too

The feeling of a soft little one gradually melting into my arms is lovely, and I wouldn't soon give up rocking my baby to sleep. That said, shifting from foot to foot in the dark several hundred times night after night can get repetitive. So tonight, as I was rocking little Peregrine, I set myself an intellectual challenge. I was going to simultaneously count how many times I shifted from foot to foot and plan out this blog post. It turned out that after only 564 rocks he woke up and demanded to be nursed, but I will write what I planned anyway.

Peregrine is now almost two months old, and we've slept with him in our bed with us from the very beginning, as we did with both his older sisters. I can hear the million voices crying out in horror, but hold on and let me explain why. The benefits, I hope are obvious (snuggles, not needing to wake up to nurse the baby, baby sleeping better, family bonding, etc.) but most people (in the US anyway) don't sleep in the same bed as their baby, don't feel allowed to, because the public health advice is that it can cause the infant to strangle or suffocate.

Zero day old Peregrine cosleeping (don't tell the nurses)

Our three children were born in Germany, Denmark and Wisconsin, respectively, and we have learned to be quite skeptical of official advice and cultural mandates that vary wildly from place to place. Advice regarding infant suffocation risk certainly depends on where one lives. When we told our Japanese friends how strongly Americans are cautioned against cosleeping, they were surprised and amused. In Japan, apparently, the official advice and common practice is for the baby to sleep between the mother and father, like Lancelot's sword (misplaced cultural reference, I know.) Our German friend warned us strongly against letting our cat near our baby, as a smothering would surely ensue.

Tigerlily and Flopper, dressed for Halloween

As an American scientist with a professional interest in early mortality, and with kids, I of course looked up the science upon which the American advice is based. The most common reference regarding the risks of cosleeping is:
Blair et al. (1999). Babies sleeping with parents: case-control study of factors influencing the risk of the sudden infant death syndrome. British Medical Journal. 319, 1457-1462.
There are more recent papers on this conducted in several countries, and as far as I can see none of them have basically contradicted Blair et al.'s clearly stated "Key messsages" (sic):

Key messsages

• Cosleeping with an infant on a sofa was associated with a particularly high risk of sudden infant death syndrome
• Sharing a room with the parents was associated with a lower risk
• There was no increased risk associated with bed sharing when the infant was placed back in his or her cot
• Among parents who do not smoke or infants older than 14 weeks there was no association between infants being found in the parental bed and an increased risk of sudden infant death syndrome
• The risk linked with bed sharing among younger infants seems to be associated with recent parental consumption of alcohol, overcrowded housing conditions, extreme parental tiredness, and the infant being under a duvet

Now, my wife and I do not smoke, do not drink, are not sleeping on a sofa, do not put the baby under a duvet, do not have overcrowded housing conditions and are merely very, rather than extremely, tired. As this 2016 study makes clear, cosleeping is often associated with overcrowding because people in poverty don't have the money for an extra room or an extra bed. Poverty increases the risk of almost all causes of death, especially infant deaths, and much of the risk from cosleeping may actually be risk from poverty. Our baby is under 14 weeks, but according to Blair et al.'s findings, without these other risk factors (particularly smoking), there is no increased risk associated with cosleeping, even for neonates.

Not child endangerment (with Tigerlily)

Other studies have added parental obesity and extreme youth as risk factors. I am certainly overweight by the standard definition, but not obese. We are not particularly young parents. But still, if there is any risk at all of rolling over onto the baby, wouldn't it be safer to have him in a crib in another room? Emphatically, disastrously, not. Blair et al. write, "There was an increased risk for ... infants who slept in a separate room from their parents." Their estimate, that having the baby sleep in another room increases risk by about ten times, has been revised by more recent research to increasing risk by about half. Long story short, having the baby in a separate room is dangerous, while having him in our bed with us presents no documented risk compared to a good modern crib.

This presents two obvious questions: How should parents who aren't trained in interpreting regression tables make this decision? And, why is the official US advice (which resembles that in several other countries) what it is?
My interpretation, based on Blair et al., and the studies that have followed from their work: Always sleep in the same room as the baby. (Shockingly, this has only now become the advice from the American Academy of Pediatrics)  Never sleep on the couch with the baby (they tend to roll into the cracks, and this is truly dangerous). Don't cosleep with a baby under 14 weeks old if you are a smoker, are under the influence, are obese, are overcrowded (which is associated with poverty and all its ills), or are otherwise difficult to wake. If those risk factors don't apply to you, there are hundreds of hours of baby snuggling available to you, and you can decide to take them or leave the baby in a proper crib in your room. Quitting smoking is astoundingly good for your children's survival, even if you don't smoke inside the home.

Now why is the official advice a simple, "Never cosleep," or starker versions thereof? Because official advice has to be short and simple to be effective. My paragraph of advice, above, is 150 words. "Never cosleep" is two. "Don't live in poverty," would be great, if people were given the opportunity of escaping. People are, on the whole, really bad at following complex advice, and really good at finding reasons why things don't apply to them. Have my wife and I really never used a duvet with a baby? Questionable. Am I really overweight rather than obese? I haven't weighed myself in over a year. How tired is "extremely?" Have I ever fallen asleep on the couch with a baby? Certainly. You see, it gets messy, and it is easier for officialdom to just say, "No."

Baby Kestrel sleeping on the couch with Big Sister. Closely supervised.

So my message to you, should you be in the position of choosing where to sleep your baby is to make an informed decision. Good slogans are rarely good advice, and extensive snuggles are one of the things babies need most.

This baby (Kestrel) is not asleep. A desk drawer is not a proper crib.

Breaking a truly ancient tradition

"Modal age at death" means the age at which the largest number of individual deaths occur. The modal age at death is determined by the interaction of two factors, the increasing risk to each individual with advancing age, and the decreasing number of individuals still alive with advancing age. As a population of adults gets older, each person is at ever increasing risk of death, so at first the number of deaths at each age goes up. But eventually, even though individual risk is still increasing, the number of people experiencing that risk is going down so fast that the number of deaths at each age (what demographers call dx) goes down. So modal age at death is neither at the age with the most individuals or the age at the highest risk, but somewhere in between. And studying the modal age at death, how it varies between populations and over time can tell demographers many useful things that I'm not going to go into because I am more interested in subverting the paradigm.

The paradigm in this case is that modal age at death is some time in early old-age, when death rate is going up but not too many people have died yet. Except that often it hasn't been. If one looks at life-tables for historical populations of humans, the highest dx (by far) is often d0, the number of individuals who die before their first birthday. Put another way, far more people died between birth and their 1st birthday than during any other year of age. The Human Mortality Database (the world's premier source for data on this sort of thing) estimates that of the baby girls born in Sweden between 1751 and 1759, slightly over 20% died in their first year. To put in perspective how the resulting distribution of deaths over age looks, please glance at this graph:

If I asked you to guess at what exact age a random female had died, by far your best guess would be age 0. And surely the people living in such a population must have been affected in all sorts of ways by the frequency that babies die. To ask how this population would have adjusted to such a thing is to ask a misleading question, because adjustment would imply that this is something new. In fact, the modal age at death is almost always 0, and modern humans are highly unusual in having it be much later in life. In fact, even in the cohort of Swedes born in the 1920s, modal age at death was still 0. In some countries it may still be 0, although such places are generally harder to get good demographic data from.

How far back does this go? Well it is 0 for hunter gather populations. It is zero for wild primates. It is zero for other mammals. It is zero for.... As far as I can tell, modal age at death has been zero for almost every population of almost every kind of organism for the entire history of life on earth. Contemporary wealthy humans still suffer much higher mortality in our first year than at any other pre-adult age, but infant mortality has been gradually brought down over the last centuries, so far that somewhere in the 1930s and '40s, many of the world's nations unknowingly broke with hundreds of millions of years of tradition by having a modal age of death that wasn't zero. So the right question to ask may well be not, "how did they adjust to zero being the mode?" but rather, "how are we adjusting to zero not being the mode?" And my impression is that one important way that we have adjusted is by very gladly forgetting that things were ever different than they are now.

Seen another way though, we haven't changed the modal age at death at all. If one is willing to classify the loss of an embryo or fetus as a death, modal age at death has never been zero. It has always been, and remains, -1. Minus one because far more individuals are lost in the year prior to birth (even if only nine months) than the year after it, or any other age. Where before many cultures avoided naming newborns until a week or month or even a year had passed, we still often avoid it for those not yet born, for the same set of emotional (some say superstitious) reasons.

Will science progress to the point that we feel safe naming that recently implanted embryo, knowing that she will almost surely make it? Perhaps. If so, modal age at death will finally, at long last, be in the age range that my demographer friends like to consider.

Article in The Scientist

http://www.the-scientist.com/?articles.view/articleNo/37154/title/Why-Women-Lose-Fertility/

Field specific meanings

In demography, the phrase "historical demography" means studying the population processes of human populations of which we have historical records. In population genetics and other subfields of biology, it means trying to estimate past changes in population sizes based on the genetic patterns of current populations.

In demography, fecundity means capacity to produce babies and fertility means realized production of babies. In biology, these meanings are reversed.

In demography, EPC means European Population Conference, the largest yearly demography meeting in Europe. In biology, EPC means Extra-Pair Copulation.

National Academy of the Extremely Vigorous

People of higher educational attainment live longer. This is widely known. Somewhat less widely known is just how far up the attainment ladder this pattern goes. People with Master's Degrees tend to live longer than those who stop with a Bachelor's. Even more longevous, on the average, are those with doctorate. But people who get a doctorate and then go get some random job don't tend to live as long as those who become tenured professors. And sitting atop this hierarchy of attainment and longevity are the members of the elite scientific academies, such as the National Academies and the Royal Society. Many people's tendency when thinking about this correlation, between attainment and lifespan, is to assume that being better educated helps one live longer. To some extent this is certainly true, and at the level of primary education, and even college education, there is experimental evidence (both true experiments and accidental experiments through policy changes) to prove this. But in thinking about the differences between groups of people with graduate degrees, I rather suspect that the causal relationship is rather different.

I'm thinking about this at present because I have had a very productive evening. Since coming home from a full day at work I've made dinner, done the dishes, played with my daughter, rocked her to sleep, folded the laundry, done more laundry and folded that also, cleaned the cat's box, organized things around the house, rinsed the drop-cloth we put under the highchair while my daughter learns to eat, cleaned the broccoli and potato bits out of the bathtub, written work emails, taken down the garbage and the recycling, climbed the 18 flights of stairs to come back up and written half a blog post. This is extraordinary for me, especially this time of year. I am almost always either coming down with something or trying to recover from it, or coddling an inflamed joint, or just feeling low energy. I lose a disgusting amount of potential productivity to being sickly. The elite academies members I know, and those who are not yet in those academies but seem likely to be in them some day, are all people who are this energetic all the time. If they do get sick, they seem to almost always be back at in after a day or two. It is rare for me to recover from a cold in less than a week, and not rare for me to be out for two or three weeks at a stretch. This is not to say that many of these people are not also smarter than me in important respects, but the trait that most unifies the really successful academics I know is their extraordinary energy and vigor. My boss, nearing his 70th birthday, and a National Academy member, hardly seems to know what it is to feel tired. He'll attend meetings on four continents in the course of a week, say how exhausted he is, and still spring from his chair to scribble equations on his whiteboard. So my belief is that people of the highest academic attainment live longest not because they are of high attainment, but because they are remarkable in their health and energy, which also allows them to produce the torrent of great science necessary to be elected to one of these societies. Alright, enough writing, I'm exhausted.

Emissions of the aged

Projections of future carbon-dioxide emissions are complicated. How will energy consumption habits change as societies get richer or more urban? What mix of sources will we be getting our energy from?

My friend and colleague Emilo Zagheni decided we should make the calculus that much more complicated (and informative) by also asking how the aging of the population will influence carbon outputs. A demographer's demographer, which Emilio surely is, is never happy with any calculation that does not include age-structure in one way or another.

This article in the Economist summarizes what he did and what he found. As people get older, they tend to consume more and more, emitting more and more carbon, until 65ish, at which point consumption tends to start declining. See the graph, and the analysis, in the Economist, or the original in the journal Demography (2011) pp 371-399.  The punchline for the carbon-watcher is that the changing age-structure will tend to increase carbon emissions until about 2050, after which point such a large portion of the population will be above 65 (I'll be 73) that the age effect will begin to marginally decrease emissions.

I should finish with a quote from Ron Lee, Professor of both Demography and Economics at UC Berkeley, who both Emilio and I studied under. Ron was one of the inventors of the widely used Lee-Carter method* (1992) for forecasting future mortality patterns. Seventeen years later, I asked Ron how well his forecasts for the first 17 years matched what had actually happened in those years. He cocked his head slightly to the left, sighed sagely and said, "Well, demographers are well aware that our projections don't always fair so well in a complex world. But we console ourselves with the knowledge that we do much better than the economists."


* The original article has been cited more than 1000 times in the peer-reviewed literature, and modifications are used by the US Census Bureau, the UN, and so forth.

Undermining the Wall of Death

Different fields of science often don't talk to each other, even when coming at the same problem from different angles. A stark example of this can be found in the literature on aging. I'm in the field of evolutionary demography, and aging is one of our central focuses. We ask why and how it happens by studying the demographic patterns of different species under different circumstances. The evolutionary demographic theory of aging is built around the idea that there are alleles that have effects at different ages and natural selection acts on these genes to sculpt the age-specific mortality at different ages. Because dying young (before you've had a chance to reproduce) is more disadvantageous than dying old (after you've already passed on some genes) natural selection acts more strongly to minimize mortality in early adulthood than later adulthood, resulting in a chance of dying that increases in age. There are decades of theory built up around this idea, and the idea is not without merit, but it does assume these age-specific gene effects, generally without bothering to say what the actual genes are or how they influence mortality.

Another field within biology that focuses heavily on understanding aging is biogerontology. Biogerontology focuses on understanding the mechanistic basis of aging at the cellular and molecular level. They describe aging as a process of narrowing of the homeodynamic space, often due to accumulation of damage. Homeodynamic space is a concept related to homeostasis (the tendency of organisms to push their physiological state back to some optimum), but with the recognition that the goal that the individual is pushing towards, and its options for pushing, change over time. For example, as the cells in an organism accumulate mutations, it becomes more dangerous to allow them to continue replicating, because this could spawn a cancer. So the cells are forced to turn down expression of genes that allow for cell replication. But if your cells are replicating less, then you should be more reluctant to allow apoptosis, programmed cell death, because cells that die can't as easily be replaced. But if you've down-regulated the genes involved in apoptosis, this means infected cells will be less likely to kill themselves, so you need to have a stronger inflammation response, so that white blood cells will be brought to areas of infection and kill the infected cells from the outside. But increased inflammation has all sorts of nasty side effects, which themselves need to be compensated for. Note that I am just making this chain up as an example. The point being that the organism, in order to deal with the accumulation of damage, has to adjust various aspects of its physiology, which can cause damage or challenges to the system, which requires further adjustments. The organism gradually loses wiggle room, paints itself into a corner as it were. When this homeodynamic space gets too small, the organism can't respond to whatever insults (internal or external) come along and gets killed.

Reading papers in biogerontology, I am struck by two things. The first is how naive and outdated their evolutionary assumptions tend to be. For example, they still will state that aging is not observed in the wild because no individual lives long enough to grow old in the wild, an opinion that evolutionary biologists began to reject in the 1960s and have now disproved with data from numerous species from plankton to humans and birds to aphids. But I am also struck by how naive they would think our assumptions about age-specific genes are. They state as one of the basic principles of biogerontology that are no genes whose roll it is to cause aging, or which act at a particular age to regulate the chance of death. You will remember I said that such age-specific gene effects, from unspecified genes, are at the center of much of the theory behind evolutionary demography. Yet biogerontologists know such genes not to exist. So our assumptions about the mechanisms are as naive and simplistic as their assumptions regarding the demography.

This lack of communication, with each field basing its thinking on ideas the other has long since rejected, is common in science. There are simply too many journals, papers, conferences, etc., too many fields that may produce important information, for anyone to keep a useful fraction of an eye on most of them. So the lack of communication between fields is to some extent inevitable, but it does have significant consequences.

This is obvious when we introduce the gerontological observation that gene expression is not highly age specific (at least not late in life) to the evolutionary literature on post-reproductive lifespan (PRLS). Much of the study of PRLS has been motivated by the idea that PRLS shouldn't exist unless post-reproductive individuals do something useful for their younger kin. This idea arises from the evolutionary demographic theory of aging I described above. If an individual has reached the age where it can no longer reproduce, the genes it is expressing at that age should be genes that selection doesn't care about at all, because whether she dies at that age has no effect on how many offspring she has. So mutations that kill post-reproductive individuals should accumulate rapidly, unopposed by natural selection. W.D. Hamilton, a preeminent evolutionary theorist of the mid-20th century, wrote in 1966 that “In the absence of complications due to parental care or other altruistic contributions due to post-reproductives, the [mortality] curve should be roughly asymptotic to the age of the ending of reproduction.” By this he means that as the individual approaches the end of her reproductive period, her chance of dying at each instant should approach 100%. This has been dubbed "Hamilton's Wall of Death." Hamilton's work is influential enough, and his basic logic sound enough, that many of my colleagues still believe we should find the Wall of Death. But in fact we can find PRLS in a huge range of organisms where there is no parental care or anything comparable, and the Wall of Death is nowhere to be found. Hamilton's prediction fails because his model is built around high age-specificity of gene expression, which we now know not to exist. Genes which are being expressed at and after the age of reproductive cessation are the same genes being expressed prior to that age, doing the same things they did prior to that age (except of course reproduction) and so they can't just suddenly cause all sorts of lethal effects. This represents a major constraint on the ways selection can shape the pattern of mortality over age, and we evolutionary demographers are just starting to come to terms with the ramifications of this. When I have time to write another longish post, I'll explain how this leads to a major question in evolutionary demography that I have been thinking about but don't yet have any plausible answer to.

Constraints

The field of life-history evolution, into which the stuff I do roughly fits, spends a lot of time thinking about optimality. What is the optimal age to start reproducing? What is the optimal amount of time to spend foraging each day? What is the optimal body size for a creature in a particular niche? The factor being optimized here is fitness, usually measured by the rate at which a sub-population with a particular trait would increase in size. The trait value that leads to the fastest increase is the optimum. Quite often when we calculate an optimum value and then measure the actual values in the population, most individuals are fairly close to the optimum. But fairly often this does not happen, and then we start talking about "optimality with constraints." By this we mean that there are certain conditions that must be met, and the optimum we are looking for is the best value that is consistent with those conditions. The most commonly considered constraint in my field is the constraint of limited resources. If you have only 100 calories a day of energy to expend, you can't expend it all on growth and maintenance and all on reproduction, your reproduction is constrained by the need for maintenance, and vice versa. So we can calculate, given a set of biologically informed assumptions, what the optimal investment in reproduction is at each age. This kind of constraint makes sense to people, and yields many useful insights. That said, it is often the only type of constraint considered in situations where many other constraints potentially come into play.

One type of constraint that is particularly hard to build theory around is that natural selection can only favor those traits that exist. That is, a trait may be drastically suboptimal, but if all individuals in the population have that trait, and the genes which determine it cannot easily be altered by mutation such that they allow a higher fitness solution, the population will continue being far from optimal.

A classic example of this type of suboptimality is known as the 'obstetric dilemma.' This is the problem that humans have narrow pelvises and big heads, and the head has to pass through the pelvis during birth. In a (now somewhat out of date but still sound for our purposes) summary of one hypothesis of how humans diverged from our chimply relatives, Kristen Hawkes (the anthropologist behind the Grandmother Hypothesis) described (in 2003) the central role this obstetric dilemma played in human evolution thusly:

* Drying environments in the late Tertiary constricted African forests, making capacities to use alternative foods more advantageous among ancestral apes.
* Bipedalism was then favored because it freed hands for tool use, which
increased success at hunting big animals, and this put a premium
on larger brains.
* But the mechanics of bipedal locomotion limited pelvic width, so brain expansion created an ‘‘obstetrical dilemma’’ requiring most brain growth to be postnatal.
Consequently, children with developing brains were immature longer and were more dependent, for a longer time, on maternal care.
* The care requirements interfered with maternal hunting, so mothers relied on
provisioning from hunting mates. This help from fathers allowed mothers to produce more surviving offspring.
* Thus, parents formed lasting bonds and nuclear families became the fundamental
units of cooperation in which a sexual division of labor served familial goals of production and reproduction.

Now according to this story, variations of which are still supported by the scientific evidence,much of the distinctness of human life-history comes through:
1. The need for large brains and small pelvises
2. Which explains why our babies are so undeveloped
3. Which explains we take so long to mature
4. Which is an important part in explaining why we end up with our social system.
5. Which explains why we live so long.

So the optimality of a narrow pelvis, the optimality of a large brain and the need for
that brain to pass through that pelvis ends up being a central fact of human evolution. And why, we may ask, is it optimal for the baby's skull to pass through the mother's pelvis? The apparent answer is that if there is only one possible trait, that trait is the best of all possible traits.

The pattern of vertebrates expelling their young through their pelvis dates back to
before vertebrates actually had pelvises.

Note that this fish has its gonads above and in front of its pelvic fin. That is a common trait among fish, including the lobe-finned fish from which all terrestiral vertebrates are descended. The lobe-finned fishes had bony feet with which they could support themselves on the sea floor, and the bones in their pelvic fins would eventually be modified by evolution into the legs and pelvis.



Now the first terrestrial vertebrates were amphibians, and like most frogs and salamanders, laid small soft eggs, so it was probably no problem for them to continue having the gonads in front and running a tube through the pelvis to the cloaca. This system only became problematic when the eggs got large and hard, as they are in reptiles like turtles. Turtle people like to talk about "pelvic consraint" when they discuss why turtles don't make bigger eggs.

The only non-fish vertebrates to escape the need to run the babies through the pelvis are those that no longer have ana full pelvis, like whales and most snakes. To my knowledge nobody has managed to invent an alternative outlet, so everybody, including us, has to find one way or another to get through the pelvis. In fact, the only alternative is a human invention, the cesarian section.

This obstetric dillema is a very obvious contraint of the 'no alternative' type. Whenever I get a chance to write another longish post, I'll give an example of a constraint where the lack of alternatives is less obvious because it is genetic rather than anatomical.

Who's a demographer?

I tend to find religious fundamentalists, whatever the religion, hard to listen to. One of the habits I find objectionable is the tendency for fundamentalists of a religion to argue that others who practice the same religion differently aren't really practitioners of that religion at all. This tendency has recently been popping up in the news as Christian fundamentalists in the US argue that because Mitt Romney is a Mormon, he isn't really a Christian. I have no fondness for Mitt Romney, and frankly suspect he would claim to be a Pastafarian if he thought it would get him elected. That said, the Mormons declare themselves to be Christians (they do talk about Christ a lot), and I can't see as anyone has the authority to tell them they aren't. It is just hubris to go around telling people that you can describe their religious beliefs better than they can.

This is all a roundabout way of getting to the question of who is a demographer. Ask a demographer what is the largest annual scientific meeting for demographers, and she will probably say The Population Association of America (PAA). I consider myself a demographer as well as a biologist, but I think most PAA members would say the stuff I do isn't demography. This is for the simple reason that I mostly study non-human populations, and the PAA defines demography as the study of human population processes. Studying the same processes in non-humans is, by this definition, not demography. Last year they had a session on evolutionary demography, but all the accepted papers were on humans. This year they don't even have such a session. I think that inserting the word 'human' into the definition of demography is roughly akin to saying that anyone who doesn't follow the teachings of a particular Rabii isn't really Jewish, so I call myself a demographer.

I was recently surprised to find myself in a conversation in which the tables were turned. A colleague was arguing that most PAA members are not really demographers, but sociologists. His argument was that many human-focused hard-core demographers feel out of place at the PAA. After their meetings this spring several colleagues complained that most talks at PAA meetings are really quantitative sociology rather than demography. The distinction is a fine one, but basically classical demography has a core set of questions and methods, and these have certainly been supplanted to a considerable degree by questions coming out of sociology, mostly approached with methods that don't require the quantitative machinery of formal demography. My colleague told me, "All the talks are full of regression tables, and most of the regressions aren't even done well."

So is it fair to say that my colleagues and I, who apply classical demographic methods to non-humans are more demographery than the quantitative sociologists at the PAA? I'm afraid not. We can no more revoke their demographer label than they can revoke ours. However, since the social demographers who control the PAA aren't interested in evolutionary demography, and most of their presentations frankly aren't that interesting to us (there really are a ridiculous number of regression tables, mostly demonstrating the relationship between fertility and female education for yet another population), I'm going to let my membership lapse. I'm thinking I'll join the British Ecological Society instead. They have lots of evolutionary demography at their meetings. I don't even feel the need to call myself an ecologist.

Decreasing population, increasing density

The Institute where I work is built on land that used to be part of huge East German ship building facility. The facility folded shortly after the German Democratic Republic did. The buildings were, for the most part, left derelict. One of the former factory buildings has been converted into an indoor shopping center. Each store is its own single story box built into the bottom of the cavernous factory building, with walkways and cranes still hanging overhead. Most of the other buildings have been torn down, and ~15 waterfront apartment buildings have been built (or are being built) on the land, most just within the two years I have been here. Apartments in these new buildings are on the expensive side for Rostock, but are filling up fast. One or two huge windowless cement monstrosities (one holds a dance club called The Bunker) remain, as well as a couple of older brick warehouses that have been refurbished. One is being used as a temporary home for the Rostock Volkstheater (community theater) while their main hall is closed for fire-code violations. It feels as though a new, fairly fancy, neighborhood is simply sprouting from the root-system of the old shipyards, among the remaining shells, foundations and train tracks to nowhere.

The demand for housing that fuels all this building is a story of migration. Rostock is Mecklenberg-Vorpomern, the least densely populated state in Germany, and a state that has steadily lost population since reunification. Many of the outylying villages are dominated by abandoned buildings. Apartment complexes on the outer edges of Rostock, plunked down in the middle of fields by the communist planners, now offer multiple months of free rent to anyone who will move in and still are emptying out. Rostock is full of college students, who don’t want to be on the outskirts in half empty buildings, and an aging population of long-term residents, who don’t either. The more abandoned the outskirts get, the strong the incentive to move toward the city center. So the center of Rostock is becoming denser even as the state loses population.

The demographics of evolutionary demographers

A group of us considering organizing a new scientific society have been compiling a list of those we would like to invite to the initial meeting of the society, to be held here in Rostock, probably next year. We are up to 125 people (excluding people who work here, as everyone who works here will be invited). The group making this list is about half male, half female. The list is 100 males and 25 females.

Why this skew? We each listed people whose names came to mind, and in some of the subfields we are drawing from (e.g., mathematical ecology) almost all of the well known people are male. Higher level academics in general still skew strongly male, and the higher the level the stronger the skew, in most cases. This is both a cohort effect (older cohorts of scientists are both more well known and more male) and a selection effect (males find it easier to advance up the ladder). Being demographers, we are very much aware of this, and are very much interested in having a diverse society, but it is not clear what we can do about it. There is also a preponderance of Europeans, North Americans and East Asians; again this is unintentional and difficult to reasonably address.

Despite these skews, it is a wonderful list of researchers, and I hope we can get most of them to attend.

Finally out

This is the first of my dissertation chapters to be published. It is good to have it out. It is in a new Journal called Methods in Ecology and Evolution. It seems like it will be a very good journal, but it is too soon to know for sure.

A measure for describing and comparing postreproductive life span as a population trait

Summary:
While classical life-history theory does not predict postreproductive life span (PRLS), it has been detected in a great number of taxa, leading to the view that it is a broadly conserved trait and attempts to reconcile theory with these observations. We suggest an alternative: the apparently wide distribution of significant PRLS is an artefact of insufficient methods.

2. PRLS is traditionally measured in units of time between each individual’s last parturition and death, after excluding those individuals for whom this interval is short. A mean of this measure is then calculated as a population value. We show this traditional population measure (which we denote PrT) to be inconsistently calculated, inherently biased, strongly correlated with overall longevity, uninformative on the importance of PRLS in a population’s life history, unable to use the most commonly available form of relevant data and without a realistic null hypothesis. Using data altered to ensure that the null hypothesis is true, we find a false-positive rate of 0·47 for PrT.

3.  We propose an alternative population measure, using life-table methods. Postreproductive representation (PrR) is the proportion of adult years lived which are postreproductive. We briefly derive PrR and discuss its properties. We employ a demographic simulation, based on the null hypothesis of simultaneous and proportional decline in survivorship and fecundity, to produce a null distribution for PrR based on the age-specific rates of a population.

4.  In an example analysis, using data on 84 populations of human and nonhuman primates, we demonstrate the ability of PrR to represent the effects of artificial protection from mortality and of humanness on PRLS. PrR is found to be higher for all human populations under a wide range of conditions than for any nonhuman primate in our sample. A strong effect of artificial protection is found, but humans under the most adverse conditions still achieve PrR of >0·3.

5.  PrT should not be used as a population measure and should be used as an individual measure only with great caution. The use of PrR as an intuitive, statistically valid and intercomparable population life-history measure is encouraged.

One of my goals in this paper was to show how badly some evolutionary questions need demographic methods. I think we accomplished that.

Competing by not competing

I am an expert in evolutionary demography. I am not an expert demographer, in that my knowledge is primarily on topics relevant to, but not central to, demography as it is currently practiced. I have friends who have detailed well informed opinions on the relative merits of the various hypotheses purporting to explain the Second Demographic Transition. I by contrast have no idea what the Second Demographic Transition is. I can't fit a Gamma-Gompertz model of senescence to age-specific mortality data without the help of a statistical demographer. I don't know the literature on pretty much any topic in human demography well enough to write a demography paper without having to first do a great deal of slow background reading.

Despite this, I am setting out to write a human demography paper, with little if any evolution in it. I can do this with some confidence because, as far as I know, I am working in an area that has been almost completely overlooked, and therefore I have no competition. Most anyone with a solid demography background could do a better job of what I want to do than I can, and I really couldn't compete. But because they haven't bothered, I don't have to compete. I can do a decent job and hopefully publish in a good journal without worrying about the competition, because I believe there to be none.

In my recent grant application, I wrote, "My work falls between demography and evolution, outside the well explored territory of either. Work within evolutionary demography tends to focus on senescence and reproduction; I have intentionally eschewed these to seek the question others have avoided. This is a high risk strategy; my work does not fit neatly into any one topic-specific journal or discipline. However this unconventional approach has the opportunity to found a new direction of study and investigate the most important questions therein." This is a polite way of saying that I intentionally avoid competition by seeking out the questions others have ignored, or deemed less interesting. Most successful scientists are successful because they look for opportunities to do what others aren't doing. I don't know to what extent other seek out whole areas that others haven't bothered with. I also don't know how successful or sustainable a strategy this is likely to be in the long run. After all, the success of the work will ultimately be measured by how many other people get interested in it, and try to improve upon it. Successful work, by definition, must therefore attract competition. So if I want to be successful, but continue not having competitors, I’ll need to move on to some other under-appreciated topic fairly quickly. As I rather like the topic I’m currently seeding, I may just have to put up with some competition, and try to stay ahead of them. having my own research group would be a great help in this. Not being an expert demographer is less of a problem if you have an expert demographer on staff.

neologistic challenge

In demography, we spend a lot of time thinking about how the risk of dying increases with age. Economics calls itself the dismal science, demography is the morbid science. Anyway, we call the increase in mortality risk with age in adults "senescence", from the Latin root senex (old). I prefer to think and write about the fact that mortality rate decreases with age in those who are not yet adults. My problem is that there is no word equivalent to senescence that mean this, so I have to invent one.

So I want a word that can be defined as "the decline in mortality risk with age from conception to maturity." A root that means improvement, growth, blossoming, development, growing up, or something along those lines would be best. Twelve points for whoever comes up with the best term.

To survive and reproduce in good times and bad

jte asks:

Is there reason to believe, or evidence to support, that the forms of evolution occurring among species during a period of abundant resources is different from the forms of evolution occurring among species during a period of deficient resources?

It seems that a lot of the argument in evolutionary theory is that it takes a lot of energy to grow extra and useless appendages or what have you, so if they really are useless, you'd expect them to evolve away. But if resources are abundant--energy is not a particularly limiting factor--do you then get a scenario in which all kinds of wacky and useless appendages appear and are not attritioned away? Which gives those appendages time to hang around enough to be available when the environment changes and all of a sudden they are useful and confer an advantage?

Or something like that?

I wouldn't go so far as to say the "forms of evolution" are different. In good times and bad evolution acts through natural selection, genetic drift, mutation and all the same basic mechanisms. Rather I would say that selection acts of different traits, or favors different forms of those traits, depending on if times are good or bad. One excellent example of this has been documented by Peter and Rosemary Grant in long term studies of Darwin's Finches on the Galapagos Islands. The climate in the Galapagos is impacted hugely by the El Nino/La Nina climate cycles. In some parts of the cycle, the islands are cool and damp, vegetation grows lush, and there are lots of big seeds to be had. In other parts of the cycle, it is very hot and very dry and only the desert plants with their tiny little seeds are producing. In the good years, the finches with the big bills can eat lots of big seeds, and reproduce like mad. In only a few years the population of one finch species is dominated by big-billed finches. Then when the rains stop, the population starts to crash, and the finches with the little bills good for extracting and opening small seeds are much more likely to survive. After a few years of that, the population of that same species is again dominated by small billed finches. This isn't individuals developing differently depending on the food supply, this is just massive, cyclical natural selection driving the population's genetic make-up around in circles.

On a much larger time scale, generalists are much more likely to survive large extinction events, while specialists often dominate in habitats that have been very stable for millions of years. Consider which of each of these pairs of species is in greater danger of extinction?

German Cockroaches or Lord Howe Island Woodeating Cockroaches
The Black Rat or the Salt-Marsh Harvest Mouse
The Common Pigeon or the Mariana Fruit Dove
The Common Raccoon or the Cozumel Raccoon
Goats or Alpine Ibex
Humans or Sumatran Orangutans

In each case the generalist are doing fine, while their specialized relatives can't cope with change. The fossil record shows multiple examples of large groups going extinct when the coprolites hit the fan, but one or two very generalized species in those groups making it through and giving rise to many new species. The amazing thing is that over and over most of those new species are specialists, evolving to be increasingly good at dominating increasingly narrow sets of resources. Give Rattus rattus a few tens of millions of years and no other mammals on the planet, and they would evolve into many thousands of separate species, filling a vast array of niches, and most of those species would be specialists. If another great collapse came, the ones most likely to make it through would again be the super generalized rat.

As far as the "useless appendages" argument goes, remember that even when resources are abundant, there is still the race to see who can convert those resources into the most offspring the fastest. Plus, the ideal situation rarely lasts very long. Usually within a few generations the population of predators has increased, the food supply has diminished, or population density has gotten so high that pathogens are spread easily. Exponential growth is not to be underestimated. So with the possible exception of humans over the last couple of hundred years, it is almost never the case that a population goes on growing for many generations without selection knocking back those who spend their energy recklessly.

That said, there are traits that are advantageous in bad time and costly in good times, or the opposite. Sometimes species evolve plasticity, such as the ability to grow a thicker coat when the winter is colder, but not waste the protein in mild winters. And sometimes, like Darwin's finches, they just evolve back and forth. The camel's hump is probably something of a hindrance when water and food are plentiful, but it bears that cost because more often than not things will get dry again, and that hump will save its life. If camels lived in an environment where it didn't get dry for some thousands of years, they might end up sans hump, looking more like big llamas. Or they might just die out, vanquished by cows and goats. Camels, after all, are specialists.

Will do math for chocolate

I don’t usually pay too much attention to the conversations in my office that are in German. My office mates both have many people coming in to talk to them and are both German, so speak in German with their German visitors. This particular conversation caught my attention primarily because it repeatedly used the work ‘Schokolade” one of the few German words I have no trouble with. Also there were four people involved. Eventually one of my office mates said, “Ask Dan, he is very interested in chocolate!”

So the conversation was about a simple demography puzzle, and my boss had offered a chocolate bar to whoever solved it first.
The problem is this: 100% of a population were alive at the beginning of a year. 60% were alive at the end of the year. Assuming that the mortality rate is constant throughout the year, what percentage of individuals were alive at the exact middle of the year?

Note that a constant mortality rate does not mean that the same absolute number of individuals die each day, but rather than the same proportion of those starting the day alive end the day dead.
Motivated by chocolate, and using one simple bit of algebra, I secured the Intense Orange Lindt dark chocolate bar.

So I can’t hang a chocolate bar in the comments section, but I can promise kudos if you can tell me what portion of individuals were alive after six months, and how you got that answer.

The right's next big scare tactic

My office mate occasionally organizes a movie night, showing something or other that people at the Institute might want to watch in the seminar room. When he heard that there were two movies, called Demographic Winter and Demography Bomb, he figured they would be of interest to an Institute full of demographers, and he got a copy of the DVDs from somewhere.

Twenty or thirty of us came to watch. Now Demographers are a pretty staid crowd, but by ten minutes in people were chuckling, and within a half hour there was full-bellied, breathing-difficulty, rolled over in pain laughing going on. Clearly, people who actually think about demography were not the target audience.

These two documentaries, are based on the premise, which they take intensely seriously, that humanity (or at least the white section thereof) is at risk of economic ruin, political chaos, pain, suffering, homosexuality, the death of the traditional family and ultimate and total extinction because of a population control conspiracy by the UN, the EU, the Ford Foundation, Democrats, Liberals, Communists, UC Berkeley, Charles Darwin, Malthus, W.E.B DuBois, Gloria Steinem, Big Gay, Nazis, Feminazis, and most of all Paul Ehrlich. Dr. Erhlich, a professor at Stanford is a favorite punching bag of the right because he made many and varied dire predictions back in the sixties and seventies, and like most prolific prognosticators, many of his forecasts were wrong. His 1968 book, The Population Bomb, is identified as the driving source behind the idea that overpopulation might be a problem. In it he made various overly pessimistic predictions about the rate of population growth and the rate of food productivity growth, and concluded that we would see mass starvation of hundreds of millions of people by the 1970s or '80s. That this didn't happen is taken as proof that there are no possible ill effects of overpopulation, and that in fact humans are now heading toward population decline, which will by the end of the century see an under populated world dominated by geriatric patients, Muslims and Latinos. The same three or four animated graphics are shown incessantly, demonstrating that "westerners" (i.e., white people) are going extinct. "By the end of the century, there may not be any actual French people in France." They string together statements from representatives of various Teabagger interest groups and wildly out of context statements by academics (often repeating the same clip of the same academic to make it seem like they are agreeing with two very different ideas) with alternatively faux-reasonable and openly snarky narrators. The failure of the US auto industry, the housing bubble, immigration, homosexuality, terrorism, crime, feminism and the excesses of Wall Street are all shown to be symptoms of insufficiently rapid population growth. Demographers from all over the world hooted and guffawed and thought it was just the funniest damn thing they had seen.

Now like most vehemently presented lies, there is a grain of truth to the hysteria behind the movie. That grain is this: some, but not all, population forecasts predict that world population will cease growing some time around 2050, and may gradually decline for some decades thereafter, falling from maybe 9 billion to maybe 8.5 billion by 2100, and in the mean time the ratio of older people to younger people will temporarily increase. (Forecasts any further out than that are beyond the realm of speculation into pure fantasy, but one thing we can say with high confidence is that no population is likely to just stop breeding and go extinct.) This will pose some serious issues we need to think about over the next few decades. (I've written about this before, here.) Our markets, as currently structured, assume continuous population growth, and we don't yet have very clear ideas about how we need to adapt to population decline if it happens. Having lots of old people per working adult is a problem if you assume that it takes the same number of working adults to take care of each old person, and ignore that their will simultaneously be fewer children for those adults to take care of. The problem with the argument, even if you take out the hysteria, conspiracy theorizing, snarkiness, smear tactics and brain-washing techniques, is that overpopulation has known, current and disastrous consequences (see "Collapse" by Jared Diamond for several hundred pages on that, also a significant portion of the articles in Population and Development Review or Conservation Biology). I have never met a demographer who argues the problems of population decline are likely to be worse than those associated with population growth, or with trying to maintain a planet with more than 9,000,000,000 people on it. Second, I've never heard any demographer suggest that the cause of the eventual decline will be anything resulting from any policy of population control. Rather, educated, urban, mobile populations (and especially educated women) have fewer kids later in life, and that slows population growth. As the educational level and economic mobility of the world's women improves, and as people continue to move to cities, they will tend to reproduce less, no matter what the UN or Professor Ehrlich tells them. The clergy also have relatively little influence (see Italy).

But now that the lie is out there, no matter how laughable, it is a convenient tool for anyone who wants to argue against population control measures, family planning, contraception, feminism, etc. A friend pointed out to me a post on The Weekly Standard's blog stating that, "the discussion in demography circles isn't 'How do we cope with two extra China's?' Rather, it's "'How do we manage one of those extra China's disappearing?'" Living in a "demography circle," I can report that the Weekly Standard's unnamed source for that statement is a made-for-Fox-News propaganda special called "Demographic Winter" and its sequel (which borrows numerous lengthy sections from the first part) called "Demography Bomb." Type "population decline" into Google blog search and up come numerous posts on conservative blogs mumbling the same point.

So while my international colleagues were laughing their lungs out, I was exchanging dark glances with the only other American in the room. To those not familiar with the propaganda machine of the U.S. far right, the movie was pure, bizarre, hilarious fluff. Man-eating purple platypus stuff. One of my colleagues later asked me, "Republicans aren't idiots right? So they do not take that [compound expletive] seriously. Maybe a few nutballs? This is a joke?" But to me, it was clear this was yet one more battle being opened in the American right's war on science. As long as one is denying evolution, climate change and the moon landings, may as well claim that demographers don't see any possible drawbacks to overpopulation, and in fact that population collapse is just around the corner.

Expect to hear more of this particular lie in the years ahead. As the right touts the four biologists willing to deny the possibility of evolution, expect them to repeatedly trot out the few demographers willing to pretend that humanity's very existence is threatened by population control.

Return on Investment in Higher Education

One of the major debates in US politics is usually framed like this: Should we spend more on things that are worth doing, or should we maintain balanced budgets by spending less. The underlying assumption of this type of argument is that spending more necessarily worsens the budget outcome. If the budget outcome income minus spending, doesn't it make sense that spending more worsens the outcome? Only if we can assume that the spending does not spur more income. In other words, only if we assume that investment is impossible. Investment is spending intended to increase income.

After reading Mike Hout's recent article on the importance of public higher education, I was left wondering how public expenditure on higher education compared to other investments in purely budgetary terms. Mike points out that education leads to higher personal incomes, higher rates of entrepreneurship and other good things that increase tax revenues. At the same time education leads to lower rates of things that cost the state money, like imprisonment and needing public assistance through social security/unemployment/health expenditure. So I wondered, how does the expenditure function as a investment. If the state of California could either spend an extra Billion on education, or invest that money in the stock market, how fast would the market have to rise to give better returns than the education (ignoring the non-fiscal benefits of education)?

I wrote to Mike to ask him, and it turns out he is scheduled to give a talk on that very subject here in Rostock some time soon. He sent me a copy of a 129 page report he and colleagues wrote on the subject in 2005. A pdf of that report (or a draft of it) is here.


The concluding summary of the report says:

The state devotes a substantial portion of its budget to supporting education in
California. That support is not wasted: the costs of neglecting education are
high, and the return this investment brings to the state is equally high. Laudable
though it may be, California’s investment in higher education is insufficient. If
things stay as they are now, that is, if future students progress through their
educational careers at the same rates as their ethnic counterparts did in 2000, the
state will suffer a net loss, and that loss will increase as years pass. With no other
changes, the state will forgo revenues from the increased earnings that education
encourages, and pay more to support a population in a situation of increased
poverty and incarceration. If, rather than maintaining the per-person level of
educational support and access, the state were to limit capacity, the situation
would become even more dire.
However, based on existing trends in educational demand, we expect that high
school graduation rates and college going rates will increase, and demands on
state support for education will climb commensurately. California will have to
invest in community colleges and universities in the short run, but both the state
and its residents will benefit handsomely from this additional support in the long
run. Our calculations suggest net savings to the state will exceed the additional cost by three-fold or four-fold, while its population will enjoy lower levels of poverty, crime, and dependency, and higher levels of average income and political participation.

(Emphasis mine)

The report also specifies in great detail the time scale on which these returns occur, and the majority of returns are "realized in the first ten years after investment because the benefits it buys -- lower welfare, less crime, and healthier children -- are problems/ benefits that disproportionately affect 18-34 year olds."

Now this suggests that we are getting 300% return within the first ten years. So let's just say our $1Billion turns into $3Billion by the end of ten years. That is about 11.5% annual return on investment (1.115^10= 2.97), better than the 11% average rise in the Dow Jones from 1926 through to its pre-dot com bust peak in 1999.

I am a strong believer in the value of education, and would gladly quadruple expenditure on education, but I find this purely monetary claim incredible. This is such an extraordinary claim that we are forced to consider the background of the person making it. The only one of the authors I know is Mike, and what I know about him is that he is the Chair of the best Demography graduate program in the Western Hemisphere, a member of the National Academy of Sciences and a careful scholar. Mike does not go around spouting grand pronouncements. If he makes a claim like this it is because his analysis tells him it is true. There really is a strong argument to be made that in purely fiscal terms, investment in higher education pays off better and more consistently than investment in stock.

In that light, two things surprise me. First, that I haven't heard that argument made before. I hang around with a lot of people distraught about the direction American (and particularly Californian) higher education is going, and no one has ever mentioned this. Second, I'm amazed that Wall Street hasn't tried to securitize this yet.

I'm eager to hear what new work on the subject Mike will be presenting at his talk.

Finally, and sadly, I've just heard that Berkeley's demography department won't be able to accept any new graduate students this year. The lack of funding makes it impossible for them to hire faculty to replace those who have retired or left for better funded institutions. If this does not improve in the next few years, I would guess the department will have to close its doors.

Education, hippies and commies

Iris: This guy is writing about a pipe dream:

Me: True, but it is a good pipe-dream. A well reasoned and researched pipe-dream with a firm empirical basis and deep social wisdom.

Iris: You mean he's a hippy.

Me: Actually, he's the Demography Department Chair at Berkeley.

Iris: Difference?


Prof. Mike Hout's recent piece on Rationing College Opportunity is timely and worth a read. One question he doesn't address, but I will likely write to him and ask: How long does it take for government investment in college level education to budgetarily* repay the government through taxes and avoided expenditures due to all the economic benefits he discusses? I would guess someone has made such an estimate, and if they haven't they certainly should.

The piece was published in The American Prospect. The subtitle of The American Prospect is "Liberal Intelligence" which Iris takes as further proof that only hippies or possibly communists** write for it.

Anyway, Mike writes well and I suggest you give his pipe-dream a read.


* Yes I know that budgetarily isn't a word, but my Grandma Esther, one of the finest kindergarten/1st grade teachers New York City has ever known, told me that because the English Language does so much for us, we have to do what we can for it. Therefore, to her thinking, we have the responsibility not only to use it well, but to give it new words that it should have but somehow doesn't have yet. 'Budgetarily' for example.

** Note that Iris was raised among hippies, and is the only person I know who waxes poetical about Soviet architecture, so she means the terms "communists" and "hippy" in the fondest possible senses. Further note that my Grandma Esther was a socialist with strong communist leanings, and would have agreed with Iris that Prof. Hout sounds like some kind of commie.

More musings on individuality and Hydra

It was a trick question. I admit. Well not a trick exactly, but a question to which science does not have a right answer. Even when the question is defined fairly exactly, it isn't clear what unit we should be looking at.

I've already talked a little bit about the hyrda, but I want to give you more detail, because they are such an interesting and bizarre case. There are at least four levels at which we could define the individual in hydra. The smallest is the individual cell. Most cells in a hydra are capable of turning into any kind of hydra cell, producing a whole new hydra, and moving on their own. People have turned a whole hydra body inside out, and the cells that were on the outside just become inside digestive cells, and the cells that were on the outside become skin cells, and the community of cells goes on about its business. Second, the polyp, that thing with the tentacles and digestive system we classically think of as the individual animal. It looks like a little animal. It acts like a little animal (in most ways). It hunts, it reproduces itself, it has different cells doing different jobs. Third, there is the physically attached cluster of hydra. Through budding (growing a new hydra-shaped organism off the side of the old one) hydra reproduce asexually, but the buds get to be a fair portion of the size of the parent before separating, and are generally of almost full complexity while still physically and physiologically attached. One or two or occasionally more buds can be growing off the main polyp at the same time, and one could easily see this mass of genetically identical connected cells as one individual, despite the fact that it has multiple sets of tentacles feeding multiple digestive systems. Finally, one could consider that these genetically identical groups of cells remain part of the same individual even after physical separation. The genetic individual could after a short time encompass many thousands of polyps.

At which of these four levels does senescence occur? We know from experimental evidence that the risk of death by individual cells increases with age, so we have senescence in level one. In level two, the polyp, the experimental evidence points to no senescence, and the same goes for level three. At level four we don't have experimental evidence, but Mueller's Ratchet implies that there would be slow senescence of the genetic individual. Without going into details, Mueller's Ratchet is a line of genetic reasoning which makes clear that the number of harmful mutations in an asexually producing population almost always increases with time, where the number could decrease with sexual reproduction. So as the genetic individual of the hydra keeps producing more polyps, the newer polyps on the average will always have more harmful mutations than those of earlier generations. And remember, just because the polpys don't age doesn't mean they are immortal. They still die, in large numbers, from causes such as being eaten. As the polyps are reproducing and dying, we end up with more polyps from more recent generations and fewer from older generations. The mutational load of the genetic individual increases, and over time this should lead to increased risk of the extinction of the genetic individual. So the genetic individuals, like the cells, senescence, but the two layers of organization in between, the polyps and the clusters, don't.

This is a real problem without a clear solution. Do hydra tell us something important about the evolution of aging, because unlike almost all other animals, they don't age, or are we just looking at the wrong scale?

A more useful way to phrase the question may be to ask why the cell and genetic individual age, but the polp and the cluster don't. The first answer that comes to mind is that both cells and genetic individuals accumulate damage in ways that they can't fully repair, while the polyp and the cluster can easily repair any damage that comes along because any one piece can completely rebuild the whole. In this context, aging occurs when organism are built in a way that doesn't allow for easy repair. At some point this idea will combine with some other idea to form something useful. Or it won't.