Hacker News Reader

You make a good point -- if you're in a place without basic infrastructure, the government and the people really want that new infrastructure and few will try to stop you. Everyone benefits from the new drastic efficiency improvements. But if all the basic infrastructure needs are met, do you really want someone building a noisy interstate near you?

My theory: in a growing economy, the easiest way to make money is to build something new (e.g. in a city with tons of new real estate being built). In a developed economy, though, there's more incentive for rent-seeking behavior (e.g., there's not a ton of capacity for new development and it's easier to stop your competitor).

Finally, labor is just cheaper in developing economies.

I think your point is correct in general, but I’d like to nitpick in a specific part here:

> it's hard for Australia to build, for Germany to build, for Turkey to build, for Japan to build.

Just looking at the cost of rail infrastructure, it looks like the US has a significantly harder time generating value for money with what we build. https://www.bloomberg.com/news/articles/2011-11-09/why-1-bil...

The cost difference per km of rail comparing Tokyo or Berlin to NYC is insane. It’s literally an order of magnitude difference in some cases. And this issue arises in many parts of our society. IT modernization projects in the federal government cost boatloads of money, yet never seem to make any real headway. Even in the private sector, real value adding things like transitioning to electronic medical records never really took until insurance companies and the government started forcing it. And once they did the focus on their design was entirely around billing and coding and not really improving the physician’s workflow or enhancing care.

So while the point may be true everywhere, I think there really is a baked in tendency to veto and strike down proposals for change in the US that doesn’t seem to get quite so rampant in other places.

I think you're both right.

To the meta-point made by the OP... We are now sitting on 200 years of social "science" failure to gain an understanding of this stuff. Theories fall in and out of favour, but they never seem to be affected by reality very much either way.

I think science, as an approach, just isn't a useful way of getting an understanding of why US rail construction costs are so outrageous. It isn't that kind of question. Aspiring to an abstract philosophical or scientific Theory of Government Procurement doesn't take you anywhere useful.

Was the "Theory of Government Procurement" ever been the goal of social science?

The goalposts of social science always seems to be moved to where the ball goes. But, I think if we're being honest, than yes.

Not the only goal obviously. But, there are countless economists & policy researchers publishing on similar topics all the time.

To me, it seems that actual scientific articles that would fall into social science umbrella have much smaller scope and goals.

I'm not sure what you mean by "actual scientific articles," but I mean stuff that's generally published in economics, policy, psychology or other "social science" related journals.

I'm not commenting on which is "scientific" in a Popperian sense. I'm not sure any of these publications clear that bar.

Questions of scope and/or magnitude is one of the moving goalposts. Scope is implied to be very broad in most contexts, limited in more critical contexts. This theory only applies to preschools in this neighborhood.

The 2018 Nobel Laureate for Economics was about "Endogenous growth theory," a theory of how human capital, innovation, and knowledge contribute to economic growth. Is that a small scope?

The most influential social science book of the last few years was "Capital in the 21st century" which is a theory of income and wealth gaps. Is that small in scope?

Ultimately, a theory that can't be generalized iss not really a theory, in a scientific sense. It's punditry.

It is a consideration in the field of political economy or development economics. Development economists rarely get invited to talk about domestic politics in OECD countries though so the focus is all on developing ones.

>Theories fall in and out of favour, but they never seem to be affected by reality very much either way.

It's the opposite. Theories happen, but reality isn't affected by them in any way. Part of the reason is because there is no longer a pipeline for public intellectuals or academics to be in any sort of advisory position in government. Those roles are all taken up by career bureaucrats, industry-linked experts from the private sector, or well-connected political flunkies now.

The trouble with that is that while these people are usually established with PhDs or whatever in the relevant field, they're usually a decade or more out of practicing in that field by the time people reach the level in their careers where they can end up on a Presidential committee. So they end up with ideas and preconceptions about their own field that are frozen in amber from a long time ago.

What's more, they've spent the time since assimilating the values of other people who are even older than them. So all their practice and discipline involves explaining theory that is a decade out of date to get buy in from someone whose understanding of the theory is 2 decades out of date.

I totally disagree.

Theories have enormous effect on reality. Coase Theorem^ dictated how radio spectrum was managed in the US, with disastrous results. Monetarism dictates most decisions made by central banks. Unionisation and later union-busting were very closely tied to social science theories. "Romer theory" had profound influence on world bank and UN policies.

Reality impacting theory though...

Keynesianism remained totally intact when 1980s stagflation disproved it. Monetarism is taught as-is despite the last 10 years of macroeconomic reality. Marxist theories of ancient history and the origins of high civilisation are virtually indisputed despite the discovery of gobleki tepeh, for an example outside of economics.

In the 30s, almost all economists predicted a much reduced workweek. Keynes, the most famous, predicted a 15 hour workweek by the 1970s. Most economists agreed: marxists, liberals, keynsians, corporatists....

The prediction was based on some fundamental theories in economics, combined with GDP & productivity assumptions. The assumptions were right. The theories have not changed. The way economists explain the result is by tweaking theory in a non fundamental way. They mostly argue that people (surprisingly) prefered working 40 hrs, earning more and spending more... The in-theory implication is that people place nearly no value on "leisure."

No need to rethink theories just because they predict falsities. No problem if the updated parameter values imply a reality that is obviously false.

^Technically, a bastardised form of the theory that reverses the authors' intended conclusions.

> Theories have enormous effect on reality. Coase Theorem^ dictated how radio spectrum was managed in the US, with disastrous results. Monetarism dictates most decisions made by central banks. Unionisation and later union-busting were very closely tied to social science theories. "Romer theory" had profound influence on world bank and UN policies.

All of these rose to prominence in the 70s, 80s, and 90s, which is my point. Our institutions are stuck there and haven't been able to break out of them even though most of the theory has refined or invalidated many of these claims based on the empirical evidence and natural experiments we now have.

The most recent one you cited, the Endogenous Growth theory, is actually the one that has been most open to adaptation and criticism. And this is largely because the organizations to whom it is most germane--like the World Bank and the IMF--are largely staffed by economists and social scientists rather than the career bureaucrats and political/industry flacks that the rest of the government is. But even then, it has lots of trouble being reflected in US government policy (agencies like State/USAID) because of that same issue I mentioned.

> No need to rethink theories just because they predict falsities. No problem if the updated parameter values imply a reality that is obviously false.

I don't know where you're getting this from. Basically every 101 level Econ textbook mentions this exact thing and how it led to having to seriously reconsider how much agency workers have in actually deciding their labor/leisure mix. Behavioral and Welfare Economists are particularly interested in this question.

Maybe I'm just old, but my 101 econ texbooks made no mention of behavioral economics and very little of welfare economics. Those were in the wishy washy electives bucket.

I'd actually like to try this experiment, irl. My guess is that the majority of micro teachers today would deal with "What was Keynes' mistake" in the "tweak and defend" way. At most traditional economics schools, LSE or whatnot, I'd put money on it.

Look... I think what you're describing is an effect, not a cause. Social science theories, when faithfully implemented, have had some terrible results.

Formal, orthodox economics still holds a lot of power in the most formal bureaucracies like the IMF, central banks, UN agencies, etc. They like legible, citable orthodoxy. They like to have a beyond reproach "scientific basis" or expert consensus for policies.

There's a noteworthy principle-agent-ish aspect to this. Social science strongholds tend to exist outside of both market and electoral discipline. A municipality or HR department isn't, so social science doesn't thrive there.

All this is a matter of degrees though. Social sciences still do have influence, certainly at the national-political level.

Slight tangent: I'm not actually very familiar with Romer, beyond the basics. Was Endogenous Growth Theory the basis for higher education subsidies, or is that a different flavour of human capital theory?

Behavioral Econ is where all the energy is right now. Advances in computing, data collection, and statistical methods have enabled levels of quantitative analysis around natural experiments that nobody thought possible even a few years ago.

>Social science theories, when faithfully implemented, have had some terrible results.

I think the challenge is that there isn't such a thing as "faithfully" implementing a social science theory. There is too much unaccounted for to just blindly apply. What you're describing sounds more like ideology, which is generally a step or two removed from the underlying theory. Social scientists are themselves aware of the limitations around generalizability of any specific truth claim. It's mostly the pop-version that get a mind of their own and start getting applied uncritically. And the shift of expert opinion in the policy world from actual, practicing social scientists to political flacks and industry types has really ramped up this tendency.

Part of the challenge is that there is also lots of ideologically motivated funding in economics and social science that is attempting to put its fingers on the scales of academic inquiry.

>Was Endogenous Growth Theory the basis for higher education subsidies, or is that a different flavour of human capital theory?

Partly. Depends on whether you're talking about in the US context or a developing country context. In the latter, from the perspective of financial institutions like the World Bank it mostly guided where the loans were disbursed. Previously the focus was very much on infrastructure. Endogenous growth theory started encouraging them to pivot to investments in education and healthcare instead. One of the challenges is that infrastructure is fairly easy and politically popular (powerful people get to bid on the contracts and it creates a rich vein for cronyism and kickbacks). Education and healthcare take more mature organizational capacity and, once you start promising it, people start getting mad at you when you suck at it and vote you out.

In developed countries education subsidies are very old. Even high education was mostly viewed as a public responsibility until fairly recently when the risk was offloaded to students instead. American Land Grant universities, particularly the University of California school system, were an early case study to prove the value of education as a growth driver.

>> Advances in computing, data collection, and statistical methods have enabled levels of quantitative analysis around natural experiments that nobody thought possible even a few years ago.

I have to say I'm skeptical. What's the best study to change my mind?

>> What you're describing sounds more like ideology, which is generally a step or two removed from the underlying theory.

The relationship between ideology and social science theory is pretty much intractable, IMO. I think you could equally argue that theory is a step or two removed from the underlying ideology. This has pretty much always been the case, certainly true of the fields' founders. Maybe Khaldun was non ideological, but that's going back a long ways.

The point about social scientists being aware of the limitations around generalizability.... Again, I totally disagree. A non generalizable theory is a just so story, or punditry. Social scientists claim (or at least imply) generalizability, usually. Non generalizability is a retreat in the face of criticism.

IRL, an education in social science, at graduate level, at highly regarded schools, across social science disciplines... is training on how to string together an argument/analysis which generalizes theory, typologies, and whatnot. That is what social science is, in practice.

(I hope we're both taking these disagreements as academic and friendly. No mean spirit intended)

> What's the best study to change my mind?

Freakonomics is a pop-Econ book that’s all about behavioral economics and really mainstreamed much of the field when it came out. There’s lots of methodological issues with individual studies but it’s a good primer on how to think like an economist.

> The relationship between ideology and social science theory is pretty much intractable, IMO.

It’s really not. But I’m not really sure that’s a productive avenue to go down so I’m going to leave it.

> The point about social scientists being aware of the limitations around generalizability.... Again, I totally disagree. A non generalizable theory is a just so story, or punditry. Social scientists claim (or at least imply) generalizability, usually. Non generalizability is a retreat in the face of criticism.

No. Caveating the limitations of any particular study or finding is simply an exercise in humility about what is known and what is uncertain. “Limits of generalizability” doesn’t mean it’s non-generalizable. It just means it’s not infinitely generalizable time all circumstances. Punditry is the opposite of that.

I’m starting to have trouble following what you believe “theory” means if you’re interpreting it in these ways.

> IRL, an education in social science, at graduate level, at highly regarded schools, across social science disciplines... is training on how to string together an argument/analysis which generalizes theory, typologies, and whatnot. That is what social science is, in practice.

Ah, training on how make an argument and do analysis is the core of what academics is. I have no idea what “generalizes theory and whatnot” is supposed to mean. Most of the work people actually do is looking into case studies and comparing them against each other, doing metaanalysis on a variety of studies, or doing empirical research or hypothesis testing on a particular situation or an interesting data set. There are some groundbreaking works that manage to piece these together into grand theories, but that’s a small minority to what social scientists actually do. The lions share is just looking into “what happens to Y when X changes?”

I’m actually going to turn this back around on you. Human beings are complex and society is even moreso. I would argue that an education in hard STEM fields leaves their practitioners incapable of coping with uncertainty, evaluating the quality of evidence, or understanding how to make reasonable inferences in the face of ambiguity. The implication that theories, to be useful, must be infinitely generalizable regardless of extenuating circumstances or that outcomes much be 100% certain to be valid shows this tendency. But these assumptions aren’t actually congruent with how the real world works. The real world is messy, and collected data in it even moreso. Being able to speak about it requires a certain level of judgement and capacity for discrimination that lets you cut through the fog of ambiguity. Rather than weaknesses, these are the skills that social science equips people with.

Like you say, I think we've reached an end point on most of this.

FWIW, I think we're using the term "theory" in exactly the same way, given that we've discussed half a dozen examples.

I'll just respond to the last point. Human minds and societies are complex. The job is hard. So, what to do? The modern conception of the scientific method (eg a scientific theory must be falsifiable) was laid out most famously by Karl Popper.

His work was adopted as a framework by hard sciences. But, it was intended as a criticism of social sciences of his day... notably marxist & freudian fields. The methods of reasoning in these fields. Social science was, I believe, deeply affected by Popper's work. But, I think the direction was entirely wrong.

The conclusion should been to abandon scientific pretext when it is not possible. If what you are doing is philosophy, punditry or what Popper called "justificationism," treat it as a nonscientific avenue. Non scientific fields are valid. Non scientific theories presented as scientific are not. They're pseudoscience.

The replication crisis in psychology should have been a wake up call. I fear that the loss of position social sciences are experiencing (as you say, they've lost a lot of pull in policy formation) will make things worse, not better.

The way social sciences influence the world these days is mostly undergraduate education. That's a lot of influence still, but it's also a refuge from criticism. The only remaining threat to a social science theory is prevailing biases within its own field, because it's unlikely to leave the classroom.

Look, leaving economics aside... what are some tangible achievements of social sciences in recent generations? What has behavioural economics or psychology taught us, that we use to some benefit?

> The conclusion should been to abandon scientific pretext when it is not possible.

It’s not “not possible.” It’s only applicable under constrained context. Again I point back to certain styles of education leaving people underequipped to cope with ambiguity. That doesn’t mean the methods aren’t useful for understanding systems or processes.

> Non scientific theories presented as scientific are not.

This is just descending into self-referential semantic hair splitting. You’ve decided on a definition of “science” and you’ve decided that this definition is what social scientists must be trying to hew to based on no direct experience of your own on what social scientists actually do. It sounds like you’re conflating “wonkbloggers,” like 538 or Vox, with what actual social scientists do. (And these bloggers shoddy methods are something social scientists have been complaining about for years). But this is like taking Michio Kaku’s pop-sci writing as representative of how physics works.

It’s also not clear why it’s a hill to die on whether something is “scientific.” It’s not social scientists who worry about this, most of them are comfortable with ambiguity and understand what contexts conclusions can be drawn and when they can’t. Nobody actually cares what’s scientific or not except people who are deeply invested in the primacy of positivist scientism as a source of truth.

But that fixation is almost exclusively from non-social scientists. There is reason to doubt your sincerity when saying that non-scientific methods are valid, considering that you have, in the same breath, referred to these methods of analysis with pejorative terms like “punditry” or “just so stories.”

> The replication crisis in psychology should have been a wake up call.

The “replication crisis” isn’t even that germane in most social sciences. It’s more of a thing in clinical trials so I don’t know why you would expect it to be a wake up call to sociologists or political scientists.

> Look, leaving economics aside... what are some tangible achievements of social sciences in recent generations? What has behavioural economics or psychology taught us, that we use to some benefit?

All of public policy and policy analysis is predicated on various social science frameworks. This is a bizarre question. Where do you think explanations of things like the effects of women’s education of fertility or the most efficacious criminal justice approaches come from?

The idea that the primary area of influence is in teaching undergraduates is downright bizarre. Who do you think staffs the State Department or sits on advisory committees at think tanks?

> Marxist theories of ancient history and the origins of high civilisation are virtually indisputed

This is not true.

OK. Undisputed is an overstatement, unfair even. There's actually quite a lot of action in this space, and the publications/excavations are recent.

You do see my point though, no? I think many/most history teachers don't really know much about alternative "how it all began" history, and those old narratives are still certainly the history 101 interpretation. If high school teaches neolithic history, they probably teach a historical materialism version.

As the evidence stands currently, that's no longer evidence based.

First, there is massive difference between "theory", especially academic theory and what high school teacher teaches. That is completely different claim.

Second, no I don't really think so. The kind of worst about high school neolithic all history I can say is that what I seen of it was terminally boring.

I think we may be talking past each other somewhat. By "marxist theory" I meant the timeline and cause-effect chains that are generally in the textbooks, history books what your likely to be taught in college, etc. Marxist interpretation would have been a better choice of words.

I agree about high school (and younger) history being boring. I hated it. Don't even remember what periods they taught.

We are not talking past each other. Neither colleges nor high schools are teaching marxist interpretation of history in their history courses. The exception are courses like "marxist interpretation of x" which indeed teaches what marxist interpretation of event x is.

Also, I suspect that you dont know what Marxist interpretation of historical events is.

I don't if this is relevant but at least one of these things is not like the others. In Tokyo rail is privately owned. There are 10+ train companies in Tokyo alone and many more all over. Osaka/Kyoto probably have a similar number.

They get to own the land the trains are on and the stations so many stations become shopping centers run by the train company. Some stores they run, For example Tokyu Trains runs Tokyu groceries, Tokyu department stores. The famous 109 buildings in Shibuya are called 109 because 10 can be pronounced "tou" and 9 can be "kyu". The famous Tobu department store in Ikebukuro with 64 restaurants is run by the Tobu train company. A lot of stations have an "Atre" shopping center above them. Those are Japan Rail owned. Etc...

Some stores they rent out. Many of the new 20-30 story buildings and the new 50 story building above Shibuya station is owned by Tokyu.

That may or may not make it cheaper to build but it gives them more money to support trains and there are also arguably incentives to make your lines the most attractive so people will want to live on that line and patronize your line and all your stores and businesses.

I don't know where it will end but it's a long way from the no self interested incentive to do a good job publicly run trains. Of course many publicly run trains do great as well. I'd be curious to know if there are other incentives that help. For example Singapore trains seem have a shopping center over many stations. Maybe the government runs that and can use the rent to help balance the train books. No idea.

> Building is hardest and costliest when population density is high, when 'land' (or equivalent resource) is in small parcels, and property rights are strong - but those are exactly the conditions that lots of people love to live in.

I don't think this holds up. Throughout history the greatest examples of "building" in all domains have occurred in these conditions. I don't think the impediments to building are related to scarcity. The author is making this point as well, it doesn't matter where a person is located, a city, a frontier, a farm, or a village, what matters are the attitudes and beliefs of the people in those settings. It's the old adage "where there's a will, there's a way". I think this is a better explanation of why it is so difficult to build in some places _at the moment_ because it also explains why it was not difficult to build in the same places at different times in the past. The question here should be how do we effectively teach people to think this way en masse despite their current circumstances or past training?

Attitudes are shaped by external circumstances. By seeing someone attempt to build and seeing either failure or success or neutral result.

> where there's a will, there's a way

That is motivational, but it is not factual fact.

> why it was not difficult to build in the same places at different times in the past

None of those places have same laws and conditions as it used to.

The article is kicking off with the implicit belief that America is the greatest at building. So if America is no longer the greatest at building, we have to investigate where America went wrong. If New Zealand is bad at building but was always at building, there isn’t much to say. But if America has fallen, we have to look at the reasons, specific to America, to find out why.

> ...we have to investigate where America went wrong.

There's a very strong implicit assumption in using here the word wrong.

> A 'code smell' for social science articles

I'd watch out for this. If a historian came to you and said "I have a heuristic, a code smell for programs" you'd be incredibly skeptical. It is easy to develop strong feelings about social science that seem right and may even be right, but probably a better bet is to reach out to experts and verify these approaches.

The question is whether the writer has any clue what they are talking about either.

Also known as the "Gell-Mann Amnesia" effect, where you realize how terribly inaccurate reporting is on topics you know well, but assume that surely they do better for other topics.

They used to have a Wikipedia article about it, but I guess the censors decided it didn't warrant mention anymore. Here is a random article I found on the topic [1].

[1] https://www.epsilontheory.com/gell-mann-amnesia/

It is surely the case that a large amount of secondary writing on academic subjects is bad. This is why I suggest getting feedback from experts instead of building heuristics about topics outside ones field of expertise.

I remember this article on Wikipedia. Look at what it is now: https://en.wikipedia.org/w/index.php?title=Gell-Mann_amnesia...

A redirect to the article for Michael Crichton of Jurassic Park fame; unfortunately it appears to have been removed from that page as well. It looks like Crichton is the one who actually came up with the effect, and named it after Gell-Man after discussing it with him.

Here is the page on the edit immediately before being removed from Gell-Mann's article: https://en.wikipedia.org/w/index.php?title=Murray_Gell-Mann&...

And here is the best mention of the effect currently in a Wikipedia article: https://en.wikipedia.org/wiki/Speeches_by_Michael_Crichton#G...

Doing some more detective work, the history for the actual article is here [1].

Looks like someone named Wugapodes decided it wasn't an article worth having back in August 2019.

Here is a version from August 2018 that looks about like I remember [2].

[1] https://en.wikipedia.org/w/index.php?title=Gell-Mann_amnesia...

[2] https://en.wikipedia.org/w/index.php?title=Gell-Mann_amnesia...

This detracts from the point of the article. The author is not writing a program, nor is he aiming for "perfect correlation". He is noting a series of observations and proposing some connections between them, as is often the case with writings about society. If we take a level of rigor that works for sub-100KLOC algorithms running on dozens of cores and try to apply it to millions of interacting neural networks, each with 100 billion neurons, there will hardly be any writings about society.

I thought the article was pretty explicit in comparing US against european countries that are still able to build?