An Economic and Rational Choice Approach to the Autism Spectrum and Human Neurodiversity

That is a new paper of mine, you will find the link here.  Here is the abstract:

This paper considers an economic approach to autistic individuals, as a window for understanding autism, as a new and growing branch of neuroeconomics (how does behavior vary with neurology?), and as a foil for better understanding non-autistics and their cognitive biases. The relevant economic predictions for autistics involve greater specialization in production and consumption, lower price elasticities of supply and demand, a higher return from choosing features of their environment, less effective use of social focal points, and higher relative returns as economic growth and specialization proceed. There is also evidence that autistics are less subject to framing effects and more rational on the receiving end of ultimatum games. Considering autistics modifies some of the standard results from economic theories of the family and the economics of discrimination. Although there are likely more than seventy million autistic individuals worldwide, the topic has been understudied by economists. An economic approach also helps us see shortcomings in the “pure disorder” models of autism.

Some of you have asked me about the recent debates over the forthcoming DSM-V and autism (and here pdf) , here is one bit:

It is still possible to adhere to a DSM approach for practical fieldwork, and “autism identification,” while rejecting it as our best possible understanding of autism.  Under one view, DSM does not “define” autism but rather the DSM standards provide useful information for identifying autistics who require assistance. Alternatively, in the context of both insurance companies and schools, DSM standards allow payments to be triggered if an individual is judged to be autistic according to the specified criteria. For systems of financial transfer to prove workable, perhaps the relevant legal definitions have to cite unfavorable outcomes rather than defining autism in a more fundamental way. We’ll return to this issue when we consider discrimination. For now the point is that the DSM standards don’t have to be applied to every autism-relevant question and should not be viewed as necessarily trumping other approaches.

The DSM standards also evolve. DSM-III defined autism differently than did DSM-IV and DSM-V will differ as well. It’s well known that the DSM process itself is, for better or worse, heavily influenced by various interest groups, including pharmaceutical lobbies. So DSM approaches have to be judged by some external standard and the cognitive profile approach (and a concomitant rational choice approach) can assist in this endeavor. Again, the DSM standard should not be construed as ruling out competing or more fundamental approaches.


When an economist will write a paper about schizoid personality disorder?

That is implicitly what every economics paper is about

You do realize that autism is a spectrum disorder, right? And well-trained clinical psychologists see the DSM as a guide rather than the bible.

Hopefully it will help us see that life is a spectrum disorder. We draw lines places and should realize these are arbitrary but probably never will. Now people are debating whether the 'autism epidemic' is due to changing diagnostic criteria and the whole thing has overtones of blaming the parents. Noone asks the right questions such as what criteria? Right now, wrong before? Why not all other disorders? How wrong would they have to be before it really is a disaster?

Yes, this. We create so many problems in trying to draw lines around things, often arbitrarily.

Reminds me of the Arthur C. Clarke book where the two guys sleep together, and in the morning they joke about whether anything sexual happened, with the one saying "Hey, you know according to my psych profile I'm only 95% heterosexual!"

Labelling deviations from the mean personality as "disorders" implies that only the mean personality is "in order". Rational choice theory offers a much less biased picture. Precisely because of this, behavioural economics and rational choice theory are superior to most psychological classifications in many respects.

Yet I think it is fine to label deviations as disorders when such deviations distress the individual to a degree that he wishes mitigation or correction.


see p. 20: "Second, many autistic people are on record that they do not wish to be 'cured' of their autism and they consider their condition to be a difference rather than a disorder."

Diagnostic criteria typically fail to grasp this difference between "difference" (divergence from the mean) and "disorder" (confusing term, suggesting some predetermined form of "order").

Using the term "correction" in this context implies that the mean behaviour is "correct" and moving towards mean behaviour is the "correction". However, rational choice theory shows that mean (or crowd?) behaviour is "incorrect" in a number of ways.

Deviations largely distress individuals due to the environment being biased in favour of the majority.

I am not advocating intervention in those cases where the autistic individuals do not themselves (or their guardians) desire it; but I am skeptical that this represents the majority. Most parents do desire some intervention.

Additionally, it does raise the question of which option is the economically efficient outcome: modifying the environment to suit the (minority of) divergent individuals or modifying the individual ( insofar as they can be ) to ameliorate distress by obtaining a better fit in existing mean-biased society.

Really? Like Rahul, I doubt this is true. Just because you hear, read or know of people who do well does not mean that the majority can be "like them". Try spending a week in an autism clinic instead of armchair theorizing.

As for diagnostic criteria, again, most clinical psychologists (distinct from psychiatrists) use the DSM as a guide. The DSM does three things: (1) collect and disseminate up-to-date research, (2) facilitate communication among practitioners, and (3) allow insurance claims (distasteful as it is). The concern many have with it is point #3, for reasons that ought to be apparent. Most think therapists just flip to the DSM to diagnose someone or have the DSM criteria in their head that they check off before pronouncing a diagnosis. Hardly. On the contrary, they understand that things are often more complex and context-dependent. They go to the DSM in order to properly word a report that would allow the patient to claim insurance benefits for his or her very real condition (Treatment is determined not by the therapist but the insurance company). How about a research paper on that? How about an economic analysis of claim denials in a hospital specializing in psychological illnesses? Or a paper examining the differential rates of claim successes between drug treatments and evidence-based cognitive-behavioral/behavioral/acceptance-commitment/exposure therapy, and why?

I won't deny the flood of psychological diagnoses as each becomes "in vogue" and consequently over-diagnosed (consider the wave of so-called "multiple personality disorder" that swept the nation a few decades ago; in reality, very rare. Or depression, or ADHD). I also appreciate the attempts to stem this tide, and in the case of autism, reduce the stigma and associated stereotypes. But celebrating it as "neurodiversity" goes a tad too far and runs the risk of trivializing the very real issues that some face, and the hard work that clinical psychologists are doing to help. What's next? OCD, PTSD?

“Second, many autistic people are on record that they do not wish to be ‘cured’ of their autism and they consider their condition to be a difference rather than a disorder.”

Sociopath is sociopathic?

In many cases, the normal state would be a "disorder" if it were less common. Sometimes, everyone has problems, you just get different problems for different people.

FWIW, I am one of the people who wouldn't want to be "cured", but I also recognize that there are penalties for being me; it's just that I like the upsides more than I dislike the downsides. Also I like existing.

I actually found this more enetritainng than James Joyce.

Boy, the paper was really hard to read through; is it just me or is the overall language really complex and abstruse?

My pet belief remains that papers with little to no math, graphs and figures in them have the most difficult to understand language.

I don't see the same thing, but your general notion might be right. Try describing an equation with words and it's probably a lot. That's basically what math-light papers do. And to not blow up the page count it must be dense.

Two thoughts after partially reading the paper.

1. I would have really liked to have seen a citation for institutional forces at play in the evolution of DSM standards. ("It's well known" made me cringe, since everything else in that blockquote stood on its own pretty well.) A quick google scholar search (DSM institutionalism) didn't turn up much for me, but you might have better luck. Best starting point that I came across (an article that cited the top google scholar result) was the last section ("Whither psychology and psychopharmacology?") from Strand, M. (2011) Where do classifications come from? The DSM-III, the transformation of American psychiatry, and the problem of origins in the sociology of knowledge. Theory and Society (40:273-313).

2. In prediction 5 ("The relative pecuniary returns of autistics will rise with economic growth and increasing market size"), I don't see a fleshed-out mechanism that gets from autistics having "higher earning potential, relative to non-autistics, in a wealthier and larger economy" to actual returns. A larger number of highly specialized jobs may exist and be good landing spots for autistics, but search costs and uncertainty about a new position are a friction even for non-autistics. Could these be even higher for autistics, who may require more specialized environments, or at least work tasks and management structures that are compatible with a work-at-home model? Is there a risk of overspecialization in an autistic leading to less labor mobility? (I'm thinking of Williamson and transaction cost econ)

What I liked most was the sidestepping of impairments/symptoms and thinking of autism as an alternative information processing model. It is more inclusive of high-functioning individuals who would still benefit from coping techniques and strategies in areas of weakness.

I may have more to say after reading the section "The happiness of autistics," but I couldn't get any further until I got those two points out of my head.

The working paper provides a useful framework for modelling individuals that have household production functions with above-average variance of preferences. I'd like to add and emphasize three points that follow from points raised in the paper. My terminology varies somewhat but should be compatible.

(1.) Higher consumer surplus due to higher-variance distribution of reservation prices

As for mean-variance individuals (MVI), high-variance individuals (HVI) choose the goods with the highest difference between their reservation price and the market equilibrium price. Due to the different behaviour of the household production function, HVI will have a higher variance in the distribution of reservation prices. In the distribution of the differences between reservation prices and market prices, i.e. consumer surplus, both the positive and the negative tail will be fatter. As a result, consumption bundles among HVI include goods with a much higher difference between reservation prices and market prices.

(2.) Higher price for integration in social structures

The interaction with other individuals, facilitated by integrating in a social structure (e.g. common-interest group or family), yields utility. The price paid is the necessity to imitate consumption choices of other individuals instead of choosing the consumption portfolio that maximizes personal utility. The cost of social interaction is minimized where the difference in preferences is minimized. As the variance of preferences increases, the probability that individual preferences match the preferences of other individuals decreases. The price for social interaction increases and the prevalence of social interaction decreases.

(3.) Higher consumer utility from consumption-production match

A consumption-production match can be observed wherever an individual produces a valuable good while at the same time gaining utility from the production process. It results in consumer utility both from the production side and from the consumption side. Among HVI, a consumption-production match can yield a higher consumer utility than among mean-variance individuals as it is more likely that the individual has chosen an activity that yields particularly high utility.

These are just some three points that I thought of while reading the first pages of your paper. I'll be happy to add more.

Another note on diagnosis:

Associating current types of diagnosis with adverse selection fits very well. Adverse selection causes the diagnostic estimators to be statistically biased.

If there is a need for a better estimator, measuring the statistical variance in utility from consuming a selection of goods may be more appropriate than just a selection of diagnostic criteria. This could be based on the way utility is measured (be it cardinal or ordinal) to approximate utility functions.

I hope this isn't too much of a deviation from the focused discussion of classification, but I was wondering if anyone else saw this 60 Minutes episode about a child prodigy named Jake who attributes his autism as empowering his understanding of astrophysics. His parents mentioned cultivating activities he was interested in as a way to keep him engaged socially. Here is the link:

While the presentation of "neurotypicality" as a disorder is mostly done for effect, to try to express the problems with some of the clinical language used to frame discussion of autism, I feel that this points out the significance of that way of thinking. Seriously, if it weren't such an absolute majority of humanity, would anyone consider the vulnerability to framing errors, etcetera, *not* to be a disorder? :)

(Just been reading Kahneman's book, _Thinking Fast and Slow_, and finding it interesting how many of the examples don't work on me. How the heck do you guys get anything DONE?)

I have a theory that 90% of autism cases can be prevented with omega-3 supplementation. It seems to help those already diagnosed as well.

I take Omega 3 ('cuz a doctor suggested it). No noticeable effect on autistic traits, nor any reason for which there should be one.

Autistic traits =/= autism.

Human brains are evolved for a diet with a lot more omega-3s than we get in a typical American diet.

As for reasons, there are studies:

If that theory holds Japan ought to have a far lower incidence of Autism; does it?

I suspect they have fewer diagnoses anyway due to being a paternalistic shame culture (they also have a notoriously gender-skewed stillbirth ratio), but a mass comparison of brain MRIs would be very interesting.

The RCT's certainly seem quite promising. Although I'd go with "treatment" rather than "cure".

I agree, as with other things it looks like there will always be a genetic component.

Please, please, please, do not refer to people with autism as "autistics". I am the mother of a young man with autism and consider the word extremely offensive. He is a person, not a diagnosis. You may make an excellent point, but the use of that term sets my teeth so on edge that I cannot even read your publication. The correct phrase is PERSON, PEOPLE, OR INDIVIDUAL WITH AUTISM.

I'm sorry but how is it any more offensive than saying diabetics, paraplegics, schizophrenics, haemophillics and a whole lot of other such similar words in common use? In any case he is writing an academic publication; not a support group newsletter.

You know what?

Fuck. You.

You are not the one who gets to decide whether to be offended. The autistics are, and you know what? I know a lot of autistic pople. NONE of us like "person with autism", because that insultingly implies that autism is this horrible thing that happened to us, and that the real person is seperate from it.

It's like referring to someone as a "person with humanity" or a "person with maleness". This is a matter of *identity*.

Only place I've seen "person with autism" advocated is the eugenics nuts at Autism Speaks who want us wiped out.

I don't *have* autism. I *am* autistic. This permeates everything I experience. It defines what it is like to be me. A thing that was otherwise like me, but not autistic, would be *someone else*.

The cycle is thus:

1. A definition for a disorder is created based on a class (let's call it class A) of people with those qualities experiencing massive disabilities and dysfunction, with no clear or insignificant upside
2. Meds recognise "Hey, these more normally or high functioning people are similar to class A".
3. Lumpers then group these relatively more normally functioning people with class A, possibly aided by identity politics by people who want to feel that their deficits are proper and medical and shouldn't be laughed at.
4. Interate 2 and 3 repeatedly
5. The people who now find themselves in class A, which has a vastly different population to class A in 1, complain bitterly about for the definition of class A in 1 (which has not been updated) as a group of people suffering a disorder does not reflect their "unique abilities" and "diversity". Society complains bitterly about how the people who everyone thought were normal who have now been agglomerated into class A are underrated.
6. At the same time, then carers for the original class A in 1 are pissed off that the population they are looking after is becoming increasingly neglected due to identity politics.

Reasonable argumentation isn't going to accomplish any change in the DSM. If autism rights advocates want change, they need to copy the terroristic tactics of the Gay Liberation Front and disrupt the APA as much as possible. The only reason the APA hasn't caved in to MindFreedom (the genuine crazies) is that organization's lack of, uh, organization.

I find those days so ftartrusing! When you're all prepared to be organized and get as much work as you can done, and then something comes and sidetracks you. Even when it's important and you know it's important, it's still a ftartrusing distraction.I hope you have a better day and I hope the right course of action for this event becomes clear.

Comments for this post are closed