nrr

I'm mostly trying to pick apart the epistemology of thinking about economic systems this way because, hey, maybe there's something that my professional background in the subject is blind to.

It isn't especially interesting to me to try devise frameworks for coercing Caesar to do one thing over another; Caesar's just going to do what Caesar's going to do. The most that I can do is hedge against needing to migrate again because this Caesar has decided, e.g., that staunch pacifism is treasonous.

Hmm. Why should the word monopoly only refer to the legal mechanism by which market exclusivity is enforced rather than referring to the resulting market structure or market power?

If I'm not legally prohibited from competing with a firm, but that firm's position nonetheless also makes it unprofitable for me to enter or compete, what is that market structure called?

> The Crown did not heavily tax the colonials.

Huh, did they not? After the Seven Years' War, the Currency Act of 1764 received assent, and it functionally required that public debts be paid in hard money. Unless I'm remembering incorrectly, this was a little hard to come by in the colonies.

> Bone evidence showed that the colonials worked like dogs and died young.

I'd be interested to see the anthropological chain of inference for this.

> When productivity rose sufficiently, child labor disappeared, as did the long hours.

I'd interrogate the steps that led from rising productivity to both the disappearance of child labor and the disappearance of long hours. The claim that rising productivity alone led to this is perhaps a little hard to defend.

> The economy was not productive enough to enable them to not work. The same for hours worked.

This shows necessity but not sufficiency.

> Colonial America had much of nothing.

I'm unsure that I'd draw that conclusion. Why else would the Crown want to tax the colonials so heavily?

> "Taiwan has a highly developed free-market economy where prices are set by supply and demand with little government control."

Forgive me. Was this in dispute? I was asking what causal work the label "free market" was doing with respect to Taiwan and how the country's economic development was adequately explained by laissez-faire economics.

The patient is healthy now, so the treatment that produced the health must have been "being healthy."

> The high levels of immigration came from poor people in un-free countries coming to the US to make their fortunes. Free markets is what has distinguished the US from other countries.

I accept that free markets is only one possible answer here. A lot of folks moved here because the US had all of that land that I'd mentioned. There were also some unusually high wages here compared to most of Europe in the 1700's[0, 1]. The political institutions here also allowed certain freedoms that weren't available in Europe (of particular note for me personally is the Münster rebellion), and by the time the earliest members of my family moved here in the 1860's, there were already immigrant communities to lean on.

This goes notwithstanding that the US had developed a bonkers huge continental economy with relatively abundant capital.

My argument isn't that markets didn't make this possible. My argument is that drawing a causal inference based on markets alone is intellectually dishonest.

> And every country that has tried free markets since has prospered accordingly, regardless of their natural resources, waterways, etc.

Can we identify the countries that "tried" a free market approach? Can we identify which of those subsequently prospered? Is the relationship between those two things consistent enough to support the consequent of prospering accordingly?

I think the first is perhaps most obvious, but I'm at a loss for how to answer the latter two.

> Like, what natural resources does Taiwan have?

I'm honestly left scratching my head a little at what about Taiwan meets the bar for being a free market or where the comparison is here.

Taiwan has historically had heavy-handed state-directed industrial domestic policy. Like, the favoring of semiconductors wasn't an accident of the market figuring out what to do there. The domestic capital markets at the time of TSMC's founding were too small to finance the construction of a fabrication plant, so the government stepped in to assume the risk[2].

From a risk management perspective, though, Taiwan does tend to eschew a lot of the economic protectionism that we see elsewhere. Tariffs are less of a problem there, and the state will tend to sell its interests in businesses that don't do well abroad, so there isn't a sort of pathological "Weekend at Bernie's" or "too big to fail" kind of situation.

Taiwan's government also tends to fund research and development much more heavily.

--

0: Lindert, Peter H. and Jeffrey G. Williamson. "American Colonial Incomes 1650-1774." NBER Working Paper Series, no. 19861. January 2014.

1: Lindert, Peter H. and Jeffrey G. Williamson. "American Incomes 1774-1860." NBER Working Paper Series, no. 18396. September 2012.

2: This condenses a lot of the history of TSMC. It actually starts with ITRI in 1973 and the edict from above that research should focus on semiconductor development. Morris Chang then built on a lot of expertise developed via RCA and devised the foundry model, which then led to TSMC's founding. In some sense, it could be argued that the state deliberately constructed the market surrounding the semiconductor manufacturing industry.

Is this true? I ask because the metric for determining the distance from the US's construction at any point in time to the libertarian ideal is unclear. Like, given an assertion of closest, it should be quantifiable[0].

It should also be falsifiable and admit a counterfactual.

1. The US was within some epsilon of being libertarian;

2. The US became rich;

3. Therefore, the prosperity of the US was a direct consequence of libertarianism.

It's 2. here that I scratch my head at because, uh, the US has through its entire history, even pre-1776, been rich. It wasn't liquid wealth in the financial sense, but the US has had oodles of arable land, navigable waterways[1], natural resources, and favorable geography. It's only within the last 250 years that we've realized wealth in the form of bonkers huge internal markets and (at least, before recently) high levels of immigration.

In fact, going back to 1., if we consider the laissez-faire dimension alone and compare it against the rate of wealth growth the US has sustained at least over my lifetime, the relationship is inverted[2]: we've somehow managed to accumulate more wealth as we've become less laissez-faire. This is a confounding variable that needs to be controlled for before we can draw a causal inference either way.

I'd also argue that broader land policy, education, formal banking institutions, infrastructure, immigration policy, patents and other forms of economic protectionism, and research and development all belong in that causal model since they also tend to be things I look at when analyzing risk and uncertainty.

--

0: I'm still unsure if this quantification is even appropriate. Libertarian-ness admits a whole-ass vector space of institutional characteristics, and it isn't obvious to me how the norm would fall down to aggregate all of them as a single scalar for the purposes of making this argument.

1: The fact that pre-Columbian civilizations didn't have the wheel was often a justification for how backwards they were. As it turns out, when you have as many rivers as North America has, getting around without the wheel is no big deal. It's just pre-Columbian YAGNI.

2: Well, kinda. Both "wealth" and "laissez-faire-ness" minimally need operational definitions and some kind of partial order to be comparable like this.

This betrays a lack of understanding of nominal versus real debt dynamics. (And also taxes, but I'll admit that my upbringing has probably given me a unique perspective on Caesar and what it means for us to be able to use his money.)

Of particular note, debt instruments are denominated in nominal dollars, and they're paid back in nominal dollars, but what concerns the creditor is the real value of those nominal payments. Economic growth has this pernicious habit of pushing nominal prices upward, and if the money supply and credit system don't grow commensurate with the resulting increased demand for liquidity, the real burden of existing nominal debts can rise sharply and unpredictably.

This means that borrowers can find themselves underwater on, e.g., mortgages while the nominal obligations remain fixed, and banks will swiftly foreclose on them and tighten credit when considering their balance sheets. Many of the panics of the 1800's included a lot of this very dynamic.

It's a very bad time.

And? As a risk management professional, I am nonetheless left scratching my head.

The illustration that Chrome uses Mach ports is a counterexample to an (implied) insinuation that nothing in running in macOS user space uses Mach ports. Outlier or not, the fact there is at least one (albeit also very useful and very popular) application that uses them ties back to the conversation about how they're a kernel feature worth having for Mac compatibility.

> If you ever owned a vacant home, not only are you still going to have to pay all the usual bills, you'll be dealing with vandalism, squatters, and that undetected roof leak that ruins the interior.

Maybe! That said, the improvements on the land (namely, the house itself) are only one piece of the puzzle.

Housing is usually modeled as an oligopoly, where there are relatively fewer sellers than there are buyers, and this model makes some interesting assumptions. The demand for housing is relatively inelastic, landlords have market power, dwellings are not fungible for seemingly myriad reasons, and the landlord is seeking to maximize profit over occupancy.

For smaller landlords, the opportunity cost inherent in leaving a property vacant is often pretty high for the reasons stated here. Paying out obligations, remediating illicit uses, remediating undetected faults in the structure, etc. For these folks, leaving the dwelling vacant is a losing proposition[0].

However, for the more institutional landlords[1] that have a lot of dwellings and, possibly, own most of the rental housing that exists in a particular market? There's an incentive to constrain the supply beyond what, e.g., exclusive-use zoning and arduous design review processes achieve. By the oligopoly model, the fewer of those dwellings they rent out, the more they can command in the price of rent.

Diamond cartels, for whatever it's worth, do the same thing. The same model applies. Welcome to market power.

Housing is a little more fucked in that there's also an appraisal angle, and for rental properties, the comparatives are based on the expected rent. If rents go up, the appraised value goes up. If rents go down, the appraised value goes down.

The appraised value has an outsized effect on the line of credit available to a landlord to continue their investments in real estate in that they have less to offer in terms of collateral on that leverage[2].

--

0: The only real exception that I can think of is when the housing supply is so constrained that the value of the land itself wildly outstrips the value of the improvements to that land. At that point, the land becomes a speculation vehicle, the house on it be damned. California's Proposition 13 exacerbates this by artificially lowering the opportunity cost of keeping the dwelling vacant or otherwise not redeveloping the land it sits on.

1: Real estate investment trusts count. Chances are good that anyone's retirement fund includes one of these.

2: The real estate developer Donald Trump is a notable exception to this rule. Banks, for reasons that baffle the shit out of me, continued to loan him money. There was no due diligence undertaken for most of the loans he received, and it seemed to be based entirely on vibes.

To this end, we can model the price level at some point in time in a way similar to how we model compounding interest: P(t) = P(0) * (1.0 + r)^t. Take P(0) = 1.0, r = 0.032, and t = 2026-1914 = 112. Then, P(112) = 1.0 * (1.032)^112 = 34.05. We've had an average 3.2% rate of inflation in the price level over the past 112 years, give or take.

We can make certain assumptions that the rate of inflation won't be far off from this when we evaluate certain financial risks.

Just as in modeling adjustable interest rates for compounding interest, we can make r depend on t, and at that point, it becomes an ODE problem: dP/dt = r(t) * P(t).

> and you kinda have to keep learning in software dev

It's less the constant need to keep learning that bothers me (my formal credentials involve a requirement to undertake a minimum of 30 hours of continuing education annually) than it is that the industry has focused on vendor-specific implementations of various concepts.

I've built a lot of control planes like Kubernetes and Cloud Foundry (and its accompanying BOSH) over my career as a system administrator, but hiring managers and staff engineers these days specifically want Kubernetes experience.

Likewise, I've built a lot of control planes like what backs AWS's EC2 (including multi-tenant accounting and quotas and MAC/RBAC!), but hiring managers and staff engineers specifically want production experience with EC2 (and, seemingly, the broader AWS ecosystem, which is nonetheless approachable for someone who has production distributed systems experience).

At my most tinfoil hat, I want to say that this is blatantly in the interest of setting up a consulting industry around these vendor solutions, and it creates something of a monopsony situation for someone with my skillset.

My formal credentials are as an actuary, so I can at least go back to pursuing claim severity modeling and pricing out when something has a hiccup as a primary responsibility, but this focus is almost certainly a deskilling move otherwise.

Oh, I mean, I'm fully aware of the emphasis on codebreaking and on areas like operations research that provide optimization tools for tacticians, not to mention the ITAR restrictions on crypto export.

(I was there when only 40-bit SSL was unrestricted! I very much remember those days! They sucked!)

I guess the hangup I have is that my forays into areas like differential geometry and algebraic topology are extremely abstract, and I cannot control anyone who isn't me who misuses those tools in decidedly not-sober ways. I can only make sure that I don't have a direct hand in that kind of misuse.

In the case of housing, agglomeration effects play the biggest role. A single family house in the middle of nowhere is very scarce, but it also has limited economic productivity[0]. Notably, that lone house isn't enough to support commerce.

It's also worth observing that land supply is a fixed constraint[1], and a platted parcel is what everyone should be looking at. The more inhabitants we can keep on a parcel, the more potential customers and employees we have for the businesses adjoining that parcel.

--

0: In some models, it's also said that this house has limited utility, but I'm reticent to put this front-and-center on HN. I've nonetheless left it here as a footnote to split the difference.

1: At least, it's fixed barring, like, land reclamation, which is extremely expensive.

> Are you sure your reasoning is flawless?

Oh, heavens no. I'm human, so I almost certainly make logic errors all the time, whether in terms of truth or validity.

However, one thing I am very careful about is that the definitions I use are falsifiable, which makes them minimally empirically testable. ("A hypothesis forbidding some observations means that hypothesis is testable.") In the best case, I can proceed a priori with my reasoning and entirely sidestep needing to approach a posteriori, but real life is very often not that nice.

I generally find capital-L Libertarian free markets less than useful for analysis because the syllogism tends to break down thus:

1. Free markets produce prosperity.

2. Prosperity indicates the presence of a free market.

3. Therefore, free markets are prosperous.

Like, that'd be a perfectly valid form of proving both directions of a biconditional if it were possible in this framework to show the presence of a free market that doesn't produce prosperity or that prosperity happens without a free market. Unfortunately, they're taken as axiomatic, so they don't admit the reasoning required for such a proof.

I want to be wrong though. Perhaps I'm missing something?

> Free markets correspond with prosperity, in case after case.

I'll caution that correspondence does not admit causal inferences.

Forgive me, but I don't exactly follow. This isn't obvious to me.

Thanks for sharing. I don't work on weapons as a matter of deep conviction, so I'll continue to be concerned about this.

The stuff in Airtable seems to be. The lion's share of what's there is research that has applications to, like, sensors and robotics and AI, even if the abstracts themselves don't state that, likely because the PIs behind those proposals aren't aware of those applications of their work.

(Just to illustrate how mathematicians can sometimes embody the archetype of Hans Guck-in-die-Luft: I had someone tell me once that there were no applications of ODEs or PDEs to the sciences or engineering.)

I might be biased, but everything under the Division of Mathematical Sciences in the MPS directorate is important. Math is one of those weird fields that tends to touch everything else even if not directly, and it's impossible to know prima facie which areas of research will be useful downstream to the sciences and engineering.

We didn't have this definition of correctness until 1990 (due to Steele and White), and it wasn't until the past decade or so that we got a published algorithm (most recently, I think, Dragonbox, due to Jeon when he was still working on his PhD at UCSD), with mathematical proof[0], that satisfied all of the requirements of being complete, being unique, and being deterministic.

(Though, Steele-White expresses these slightly differently, as preserving information, having the shortest possible output, and rounding correctly. Completeness, uniqueness, and determinism seem to be equivalent and are easier to prove.)

--

0: Available at <https://fmt.dev/papers/Dragonbox.pdf>. It's pretty easy to understand either with a numerical analysis background or after having worked through volume 2 of Knuth's TAoCP.

> I implemented float-to-string for my compiler's printf back sometime before 1985. It's a medium difficult problem.

It might have been medium difficult back in 1985, but the state of the art has progressed in the intervening 40 years. The constraint of correctness makes things trickier because it further imposes three requirements: completeness, uniqueness, and determinism.

Unless the implementer has the skill to prove these properties about the algorithm they've implemented, such an implementation is likely not to achieve those requirements.

It's something of a meme that's grounded in reality, but it's honestly less grounded in race and ethnicity than it is in socioeconomic status.

That said, there's a joke that the white people version of this is the little bug in the lower right corner of the Windows desktop background that tattles on you for not having activated.

Which, like, y'know, that's fair.

> The honestly very poor black community around the area? they fucking hate us white people. like actively, all the time, no what no matter what you have or haven’t done they don’t like your ass one bit.

Did you replace folks' smoke detector batteries for them?

The way I found to get around this prejudice was to go to church most Sundays and walk around with a sling bag full of 9V batteries and knock on doors to offer to maintain smoke detectors. (It probably didn't hurt that I was also good for bumming cigarettes off of and slowing down to have a chat.)

YMMV, but showing up and doing work was apparently what my neighbors in East St. Louis and Atlanta wanted.

Wait, we had socialism in the US?

> Thanks for the insights!

Most welcome!

> I generalized a bit too wide when I said "social stuff". I was thinking more along the social media angle (GitHub stars, follows, etc.)

I figured given the context, but I wanted to cover this base too just in the off chance that the intention was different. I overcorrected a little perhaps. (:

> I already foresee the inevitable "why weren't the changes propagated to the other repositories?" moment.

It's something of an ironic "What manages the configuration of the software configuration management system?" moment, really. My suggestion is to have a sensible chuckle at it.

> I think this will be my weekend project!

I hope to read more about it!

I'll add one more thing to this, given that my experience as a math major at SIU was that each hour of lecture was expected to carry four hours of studying: read and summarize the course's texts (additionally, for math courses, attempt the problems) before lecture.

I found that (1) I didn't need to take anywhere near as many notes during lecture and (2) I could ask way, way, way more relevant questions.

This also helped tremendously when it came to studying for the actuarial exams.

> I'm going to stop right there.

Um, OK.

> Perfect competition is unnecessary.

It is definitionally a precondition to applying the supply and demand model for analysis.

> Part of perfect competition is perfect information.

This is mostly correct. It's actually a biconditional: perfect competition iff perfect information.

> No transactions have perfect information.

This is correct.

> There is always "risk" involved. And risk contributes to determining the price of something.

Risk (due to Frank Knight, sometimes called Knightian risk) is a term of art that refers to known probability distributions of market outcomes. My professional training is literally to build pricing models against these.

> Ever wonder why used cars lose a third of their value when you drive a new one off the lot? It's because of the risk that a person might be selling that almost-new car with a hidden expensive fault in it.

No, I can't say I've ever wondered this; I've not been around a lot of new cars. (See also: the whole Mennonite thing.) I know that it's because of the term of art that we call uncertainty (again, due to Frank Knight, sometimes called Knightian uncertainty), which refers to those unknown probability distributions of outcomes that I listed as a precondition to applying this model for analysis. I can't price against those.

A used car market is what we call an Akerlof market, which is a market with a feedback system that has a whole lot of pathologies linked to information asymmetry. This also happens to be the textbook example of adverse selection, which is itself an example of how the supply and demand model fails to predict the market due to that information asymmetry.

> Risk is … absolutely a part …

This is correct with the words written here, but I'll caution that risk as a term of art has been misused.

I'll flip this around and instead elucidate when it does apply if that's alright.

The supply and demand model only applies to analysis of microeconomic systems with perfect competition, namely of systems that lack monopoly and monopsony; that have perfect information; that lack an ability of a single economic actor to affect the price in the market; that lack externalities[0]; that lack transaction costs; that lack unknowable probability distributions of outcomes; that lack economic actors that change their behaviors based on what they expect other actors to do; and that lack non-market rules or structures that restrict prices, quantities, or entry. These assumptions are all-or-nothing, and there are a lot of them. It's also an equilibrium model, and we aren't guaranteed to have equilibrium.

For macroeconomic analysis (i.e., analysis of multiple sectors of an economy or multiple economies), it's problematic because of a theorem from the 1970's due to Sonnenschein, Mantel, and Debreu. I've stated it elsewhere, but it also has the implication that the behaviors of the actors in a microeconomic system do not carry over to the macroeconomic context. This tends to violate many of the assumptions above. (Though, if it isn't obvious how or why, please ask.) Furthermore, in the macro context, I don't think we'll ever have equilibrium; at least, I'm having a hard time thinking of an example when it exists.

Let's bring in a physics analogy. While we can look at mechanics through the lens of elementary algebra, those tools really only give us the ability to look at static force-balance problems for springs and pulleys in equilibrium. To model such a system more closely to how it behaves in the universe, with damping, inertia, and feedback (and whatever other time-dependent behaviors I may be forgetting), we are best served moving to differential equations.

Economics is no different; the supply and demand model is a lot like Hooke's law in its simplicity and its applicability. (In fact, in my modeling, I very heavily used numerical solutions of PDEs.)

--

0: When the retail price of kerosene tanked by 70% between 1860 and 1880, this resulted in millions more kerosene lamps and lamps broadly being lit for longer periods of time. This had externalities in the form of more soot and particulate matter (and probably also respiratory illness, but I couldn't seem to find anything with observations from the time period).

Please forgive me. I happen to have a formal background in economics by way of training as an actuary, and I can't take seriously the implication that the standard model is a law. Supply and demand aren't invariant under all conditions.