The Global Precariat: The View from Da Nang

Article 3 of 3. Series on financial inequality.

On any given morning at a café on An Thuong street in Da Nang, three economies sit within ten meters of each other, none of them quite aware they are looking at the same thing: the global precariat, up close.

A Grab driver is parked at the curb, helmet still on, refreshing the app and eating nothing because the first ride of the day has not come. Inside, a Russian couple in their early thirties work off two laptops, paying for the coffee in a currency none of the three of us can hold in a normal bank account without a story attached. At the next table a German with a Notion board open is on a video call to a client he will never meet, billing in euros, spending in dong, on a tourist visa that technically does not permit any of it. I am the fourth, also on a laptop, also Russian-born, also a decade outside the country I was born in, running a business that lives mostly on servers and in wallets rather than in any single jurisdiction.

The four of us are in radically different financial situations. By the end of this article I want to convince you we are also four settings of the same machine. We all earn outside a stable employer-and-state relationship. We all carry risk a previous generation would have pooled. Almost none of Article 2's catalogue reaches us cleanly, because every tool in it assumes a worker the state can find, tax, and protect, and we have, by accident or by design, become hard to find. The courier is the partial exception, and watching his state try to reach him is one of the more revealing threads in this story.

That is the subject. Article 1 described the machinery that concentrates wealth: five transmission channels (monetary policy, housing finance, the borrow-against-assets strategy at the top, the shift from collective to individual retirement risk, and stock buybacks) compounding the gap between the return on capital and the growth of wages over 45 years. Article 2 mapped the tools that exist to slow that concentration, and ended on a finding I want to start from: the redistribution fight has moved from the tool to the container. Every instrument, from the minimum wage to a global billionaire tax, depends on a stable relationship between a worker, a state, and a residence. Capital is trying to leave that container from above, with advisers and second passports. The precariat is already outside it from below, with none of the advantages.

The word is not mine. The economist Guy Standing named the precariat in 2011, defining it less by income than by the loss of the securities a stable job used to bundle: no employment security, no income predictability, no occupational identity that lasts a career.[1] He added a second distinction that turns out to be the hinge of this article. Standing separates citizens, who hold the full set of rights where they live, from denizens, who hold only a narrower set. A migrant is a denizen. So, it turns out, is a relocant (the post-2022 Russian term for an exile-by-relocation) on a residence permit, a nomad on a tourist visa, and increasingly a gig worker the law has not decided how to classify.

One number reframes the whole picture before we start. The worker-state-residence container that the policy toolkit assumes was never the global norm. The ILO counts roughly 2 billion people in informal employment, about 58% of the world's workers, the highest absolute number on record even as the rate slowly falls; in Asia and the Pacific the informal share is around 68%.[2] For most of the planet, the secure formal job tied to a protecting state was always the exception, a mid-century rich-world arrangement that a minority of humanity ever lived inside. So this is not a Western institution breaking down in isolation. It is a rich-world minority drifting toward the condition most workers were always in, while the poorest economies try, unevenly, to build the protections it never had. This is the view from the overlap, from one vantage point: ten years of self-employed income in Southeast Asia, watching three versions of life outside the container happen on the same street.

What the five channels feel like from outside

Article 1's five channels operate everywhere. They are not Western. A Vietnamese family priced out of a Da Nang apartment by yield-chasing capital and a San Francisco renter priced out by the same flows are on opposite ends of the same housing channel. The mechanics are global. What is not global is how the channels are experienced, and that is what the policy literature, written mostly from inside rich welfare states, keeps missing.

The same structural exposure produces different lives depending on five conditions. These are not new channels. They are the dials that decide how hard Article 1's channels land on any given person:

Family network. Is there a household, a village, a parental apartment that absorbs a bad month? This is the oldest welfare system on earth and the only one most of the planet has ever had. It is also the most gendered of the five: for the male courier it is a cushion, for the woman in domestic or care work it is often an obligation, the unpaid second shift, as much as a safety net. Cost-of-living arbitrage. Does income arrive in a strong currency and leave in a weak one, or the reverse? The same skill gives a Western nomad a comfortable life in Da Nang and traps a local courier below a living wage in the same city. Currency stability. Can savings sit in the local money without melting? For a Vietnamese saver, mostly yes. For an Argentine, a Nigerian, or a post-2022 Russian, no, and that one fact reorganizes everything downstream. Jurisdictional rights (citizen or denizen). Full citizen, taxed resident, gray-zone tourist, or stateless-by-circumstance? Standing's distinction decides which of Article 2's tools can even reach you. Healthcare access. National system, employer insurance, private cover, or out of pocket? Article 2 called health-coupled-to-employment a second catastrophic risk stacked on income volatility. Outside a national system, a serious diagnosis becomes a question of which country will treat you and whether you can get there.

Over all five sits one accelerant that is not a sixth dial, because it moves all of them at once: AI and the remote-work economy built on it. It can lift a person up every dial (cheaper capability, global clients, location freedom) or strip the floor out, and which one it does is decided largely by where the other dials already sit. It gets its own section.

The five channels are the weather. These dials are whether you have a roof. AI is the thing currently rebuilding some roofs and removing others, fast.

The developing-economy precariat is not one thing

Start here, because it is the majority case globally and the one Western inequality writing reaches least. It is also not a single figure. It has at least three layers, and they score differently, sometimes oppositely.

The urban gig worker. Vietnam's gig economy is young and already structural. Grab launched motorbike ride-hailing and delivery here in 2014; the rideshare market was worth around 727 million dollars by 2023 and is projected to grow about two and a half times by 2029, with Be, Gojek, ShopeeFood, and AhaMove filling out the rest.[3] The drivers are not employees. Grab calls them "partners," the word the platforms use everywhere to set the route and the price while owing none of an employer's obligations and keeping more than an employer's power: an opaque rating can deactivate a driver with no appeal.[4] A survey by the Vietnam Labour Confederation with Oxfam found the average Grab motorbike driver earning around 7 million dong a month (roughly 285 dollars), with 95% working six to twelve hours a day, often no days off, two-thirds married and supporting at least two dependents.[5] Through the strongest typhoon in decades in 2024 and Ho Chi Minh City's longest heat wave in thirty years, the drivers kept working, because the app does not pause for weather and there is no paid day to take.[6] One driver's line to a reporter is the whole condition: "I cannot do this forever."[6]

The rural dispossessed. Behind the courier is an older form the cities absorbed. Article 1 noted Thailand's top 1% controlling roughly 35% of titled land by value: the farmer-tenant squeezed off the land by concentration is the precariat that existed before the apps, and the urban gig economy is partly where that displaced rural labor went.[7] The motorbike courier is often one generation removed from a smallholding the family no longer controls.

The woman in informal and care work. The precariat in most accounts, including the first drafts of this one, is implicitly male and mobile. It is not. Of the roughly 2 billion informal workers worldwide, just over 740 million are women, concentrated in domestic work, market trading, and care.[2] She scores the dials differently from the courier beside her: the family-network dial that cushions him is, for her, frequently an unpaid-care obligation that limits the hours she can sell; her currency and jurisdiction may be stable while her time is not her own. The five-dial frame is calibrated, by default, to a mobile worker with no care load, and a more complete reading has to flag that the same dial reads as protection for one worker and constraint for another.

The formal worker who is precarious anyway. This is the case that breaks the assumption that a real contract means safety. The Philippines is the world's call-center capital, its business-process-outsourcing sector employing close to two million people, generating around 35 to 40 billion dollars a year, and contributing roughly 7% of GDP, on par with remittances.[8] Justin Cunanan, 33, works a nine-hour night shift for a US company with a five-day week, premium pay for night hours, and health insurance for his family, on about 45,000 pesos (roughly 770 dollars) a month.[9] By every dial above he is doing well: family present, currency stable, full citizenship, employer healthcare. His exposure comes from something none of the five dials captures: his whole job exists on a single bet, that it is cheaper to route the work through Manila than to do it in the US, and that bet is under attack from two directions. An IMF working paper finds about one-third of Philippine jobs highly exposed to AI, and identifies BPO as the single sector with the highest share of jobs at risk of displacement.[10] At the same time, US bills (the proposed HIRE Act, which would tax American firms' offshoring payments, and the Keep Call Centers in America Act) threaten the arbitrage from the demand side.[11] Automation arbitrage is overtaking labor arbitrage.

The honest reading of that IMF paper is more interesting than doom, and I want it in here because the precision cuts against the easy story. Nationally, about 61% of the highly-exposed Philippine jobs are also highly complementary to AI, meaning the technology is more likely to raise the worker's productivity than replace them; only around 14% of the workforce sits in the high-exposure, low-complementarity zone where displacement is the real risk.[10] BPO is in that zone, but it is only about 3% of the workforce, and the sector is still adding jobs and trying to climb out of it, moving up into analytics, fraud detection, and AI-supervision work.[10] So the BPO case does not prove that AI is flattening Filipino labor. It proves something narrower and, for this series, sharper: formality is not the dial that protects you. Cunanan has the contract, the benefits, the citizenship, and sits in the one zone where the contract cannot help, while the courier with none of those things may be more adaptable precisely because he was never betting on a single arbitrage holding.

And here is the part the cruder version of this argument (including my own first draft) gets wrong by implying the toolkit reaches no one in this world. The state here is building the container in real time, not dissolving it. Vietnam's amended Social Insurance Law (Law 41/2024/QH15) took effect on 1 July 2025, extending mandatory coverage to part-time workers, household-business owners, and people on non-labour "service contracts" (the exact device platforms use to dodge obligations), and cutting the pension-vesting period from 20 years to 15.[12] The revised Health Insurance Law took effect the same day.[12] This is the gig-reclassification fight from Article 2 arriving, on schedule, contested. It comes with the honest backfire the series keeps finding: the new restriction on lump-sum withdrawals triggered a pre-deadline rush to cash out, and low-income workers are resisting enrollment, asking to stay classified as contractors to keep their liquidity, which can push them further into informality rather than out of it.[12] The container is being welded together here, leaks and all, while it comes apart elsewhere. That contrast runs through the rest of the article.

The post-2022 diaspora: denizens overnight

The Russian couple at the next table belong to the most sudden denizen-making event of the decade, and it is the one I am closest to, so I will separate what I have lived from what I have watched.

The wave is large and it has stayed gone. The Bell counts at least 650,000 people who left Russia after February 2022 and were still abroad more than two years later, the largest exodus from the country in three decades.[13] The OutRush project, surveying more than 8,500 emigrants, found fewer than 10% had returned, against early relocation-industry claims of 45%, and only 5% planning to return within a year.[14] The first month alone took 50,000 to 70,000 IT workers, the exact cognitive-labor segment that can work from anywhere.[15]

The defining feature of this diaspora is not where it went. It is what it lost on the way out, and the loss is specifically financial-infrastructural. Sanctions severed the ordinary plumbing: a Russian card stopped working at a Tbilisi ATM, a Wise account got frozen on a compliance flag, a transfer that used to clear in a day stopped arriving. OutRush found 63% reporting declining trust in host institutions, much of it banking friction, and 45% reporting discrimination, well above the migrant baseline.[14][16] This is the cleanest live example of Standing's denizen: a class with currency exposure, no reliable banking, and no fallback country. A Ukrainian refugee has, at terrible cost, a clear legal status and broad sympathy. A post-2022 Russian who opposes the war often has neither, too compromised by passport to be welcomed, too opposed to the regime to go back. (One caution on the data: surveys of emigrants who left and stayed gone cannot see the ones who ran out of money and returned quietly, so even the diaspora's own self-report skews toward those for whom it worked.)

Score the dials and you get nearly the inverse of the courier. Family: severed, the household in a country you cannot easily return to or move money out of. Arbitrage: often positive, for the IT worker billing a foreign client from Da Nang or Tbilisi, the one thing working in their favor. Currency: catastrophic, the ruble useless as a savings vehicle abroad and even holding hard currency an operational problem. Jurisdiction: degraded, the textbook denizen, visa runs and multi-year residence permits and the second-passport question hanging over everything. Healthcare: private or absent.

This is the world my Russian-language channel and my own history sit beside rather than fully inside. I left Russia roughly a decade ago, not in 2022, so I had already built a life on rails outside the Russian banking system before those rails were cut. I watched the 2022 wave arrive into the Asia I already lived in, in Phangan and then Da Nang, carrying the banking panic and the currency anxiety with it.

The Western precariat, and the nomads who arbitrage it

The German at the third table is two things at once, and the confusion between them is the point.

There is a Western precariat that never leaves the West. In the EU, 8.2% of working adults were at risk of poverty in 2024, a figure that has held around 8 to 9% for years, running as high as 13.4% in Luxembourg.[17] In the US, the adjunct teaching four courses across three campuses with no health insurance, or the worker holding more than one job (a group steady at roughly 5% of the employed), is the white-collar face of the same condition.[18] Their channels from Article 1 are identical to everyone else's, but their dials are a specific bad combination: family: weak, because the atomized Western household does not pool three generations under one roof; arbitrage: negative, earning and spending in the same expensive place; currency: good; jurisdiction: full, which is the one thing in their favor, because it means Article 2's tools (the EITC, the minimum wage, gig reclassification) are at least aimed at them. They are the population the toolkit was built for, and they are still struggling, which is Article 2's sobering point about the limits of raising the floor.

Then there is the part of that same population that did the arithmetic and left. The digital nomad is, structurally, a Western precariat worker who solved the cost dial by relocating into someone else's. This is the move I made and the move the German made. You take a Western income, often unstable and benefit-poor, and you spend it in Da Nang, where the same money buys a life the same income could not buy in Berlin. The arbitrage is real and I will not pretend otherwise: it is the most effective unilateral move available to a precarious knowledge worker, and it requires no policy, no permission, no employer.

It also requires becoming a denizen. The nomad infrastructure has formalized fast (more than 50 countries now run some version of a digital nomad visa, and Spain alone issued around 32,000 with applications up 40% year on year), but it sits on a deliberate legal fiction.[19] Thailand's DTV, launched in 2024, runs five years on 180-day entries and is explicitly a long-stay tourist visa, structured so the holder does not become a tax resident and cannot work for local clients.[20] Vietnam, where I actually live, has no dedicated nomad visa at all: the e-visa caps at 90 days with limited extensions, so the entire Da Nang remote-work economy runs on tourist workarounds.[21] Cross the 183-day line almost anywhere and you become a tax resident of a country whose system was never designed for your situation.

And the arbitrage is not free for the place that hosts it, which is the part nomad writing, mine included, has been slow to say. The same An Thuong grid where the four of us are sitting is now the most expensive and fastest-rising rental market in Da Nang, the beachside blocks where international tenants concentrate.[22] Da Nang has been named Southeast Asia's fastest-growing nomad hub, and the gap it produces is visible in individual lives: a Vietnamese resident interviewed by a reporter had her rent raised after a few months in the center and moved twenty miles south to Hoi An, where rents were 30% cheaper, while for the foreigner two tables over the same apartment at 455 dollars a month reads as a steal.[23] From her side, the nomad's "strongly positive" arbitrage dial is a transfer: a strong currency bidding up the rent on a weak one. The private exit that solves my cost problem helps create a new local precariat layer the five dials do not score, made up of the people who were here first.

So the nomad's dials: family: chosen and thin (a Slack channel and a coworking space are not a village). Arbitrage: strongly positive, the reason the model works, and not costless to the host. Currency: good, if the income currency is hard. Jurisdiction: deliberately ambiguous, present enough to stay, absent enough to owe and be owed little, a denizen by choice rather than by exile. Healthcare: private cover, or a flight home. The nomad looks like the winner of this comparison, and in absolute terms often is. But "solved the cost dial by moving into a country on a visa designed to give you no claim on it" is a precarious solution dressed as a free one. It works until a visa rule changes, a tax authority gets curious, or the income currency wobbles. It is precarity with better weather.

The global precariat scorecard

Dial	VN gig courier	PH BPO worker	Russian relocant	Western nomad
Family network	Strong (cushion)	Strong	Severed	Thin, chosen
Cost-of-living arbitrage	Negative	Neutral	Often positive	Strongly positive
Currency stability	Good	Good	Catastrophic	Good (hard income)
Jurisdictional rights	Citizen, thin state	Citizen	Degraded denizen	Denizen by choice
Healthcare	Partial, expanding	Employer	Private or absent	Private, or flight home
AI exposure (accelerant)	Rising, double-edged	High displacement risk	High, but climbing	Maximal, both ways

Laid side by side, the same five dials produce four different lives from one structural exposure. The accelerant row is AI, acting on all of them.

Read down the columns and "precariat" stops being a slogan. The courier and the nomad share almost no dial settings yet sit on the same structural exposure. The BPO worker scores best on the stable dials and worst on the one that now decides his future. The relocant is the courier's mirror image. None of the four is reachable by the same policy, which is the objection the whole frame invites: a category whose members share no settings is pitched at a high level of abstraction. That is deliberate. The shared thing is structural, the split of citizen from worker from resident, and structural is exactly the level at which the policy toolkit operates and fails.

The accelerant: AI moving every dial at once

Here the three worlds stop being parallel and start colliding, because the same technology is an on-ramp for one person and a demolition crew for the next, often inside the same month.

I build with this technology for a living. I run AI agents on my own VPS, I build lead bots and automation for clients, my whole practice exists because a non-developer can now ship software by writing prompts. So when I say AI lowered the barrier to entry, that is not theory: it is the thing that lets a freelancer in Tbilisi or Da Nang produce work that used to need a team. The clearest winners are not even in my three worlds. They are the local Global-South professionals whose dials are all set well, the Vietnamese or Filipino developer with family, citizenship, and a stable currency who now picks up the same AI on-ramp as the Western nomad without giving anything up for it. For that person the technology is pure lift.

The demolition is just as real, and the 2025 to 2026 data makes it specific. A study by Demirci, Hannane, and Zhu (Imperial College Business School, Harvard Business School, and DIW Berlin) in Management Science found that within eight months of ChatGPT's launch, postings for automation-prone freelance work in writing and coding fell about 21% relative to manual work, with writing the hardest hit, and image-creation postings fell about 17% after image generators arrived; competition rose as more freelancers chased fewer of the remaining (and more complex) jobs.[24] A Brookings analysis (Hui and colleagues) found AI-exposed freelancers losing roughly 2% of contracts and 5% of monthly earnings, with experienced, higher-priced workers hit too, not just beginners.[25] More recent platform data shows writing projects on Upwork down 32% year over year in 2025 and entry-level availability collapsing from 15% to below 9%.[26] The bluntest single exhibit is the Ramp "Payrolls to Prompts" study (January 2026), which tracked real company spending: the share going to freelance marketplaces fell from 0.66% in late 2021 to 0.14% by late 2025, while AI-model spending rose from zero to 2.85%, with more than half the firms that used freelance platforms stopping entirely.[27] The authors are careful to say the evidence is associational rather than strictly causal, but the direction is hard to miss.

The cruelty is in the distribution. The work that vanished first (basic content, simple translation, data entry, boilerplate, customer-support scripting) is exactly the entry-level and mid-tier work the Global South precariat depends on most, because that work was the on-ramp through wage arbitrage. The World Bank counts between 154 and 435 million online gig workers globally, the slice of the precariat most exposed to task automation, growing fastest in developing countries, and concentrated in the Philippines, India, Nigeria, and Bangladesh, the places with the fewest alternative jobs.[28] There is a darker symmetry underneath: the writing and translation that trained these models was scraped from the same kind of freelancers the models now underprice. The labor helped build the tool that replaced it. Meanwhile AI-related freelance work crossed 300 million dollars in annualized value on Upwork by late 2025, and freelancers who specialized in AI earned roughly 40% to 60% more per hour.[26][24] The ceiling rose and the floor fell, in the same labor market, at the same time.

This connects to my earlier AI essays: AI is at once the cheapest capability-leveling tool the precariat has ever had and the fastest wage-compression engine aimed at exactly the cognitive work the precariat sells. Which side you land on depends on where your dials already sit. The relocant with a strong technical base climbs. The courier whose plan B was online piecework finds plan B already automated. Same tool, opposite effect, sorted by who you already were.

The strongest case that this is good news

The objection the whole framing invites, stated at full strength, because the capital-flight version of it is the weak one and I owe you the strong one.

The optimist case rests on revealed preference, and it is not stupid. Nobody conscripted these people. The Vietnamese courier left a rice paddy or a factory line for the app because the app paid better and gave him control over his hours. The relocant chose exile over a regime that would jail or draft him, and most report improved wellbeing abroad. The nomad chose Da Nang over Berlin and lives materially better for it. The freelancer in Tbilisi bills clients on five continents that his predecessors could never have reached. On the absolute-level metric Article 1 took seriously, every one of them is ahead of where their parents stood, and AI, stablecoins, and remote work are the technologies that made the leap possible. From this angle, "precariat" is a pessimist's relabeling of what the people living it experience as freedom.

I take that seriously, and it is partly right, so the rebuttal has to be precise rather than dismissive. Three things, and they do not all apply to all four people, which is itself the point.

First, for the bottom of the structure, revealed preference confuses "chose the best available option" with "the option is good." The courier did not choose gig work over a secure salaried job with a pension; he chose it over subsistence farming or a worse factory. Choosing the best of bad options is not evidence the options are good.

Second, for the top of the structure (the nomad, and me), the choice is good, and that is exactly where the trap hides. The nomad's option is not the best of a bad set; it is a real improvement. But it is an improvement bought by becoming a denizen, holding the upside of mobility and arbitrage while quietly shedding every claim a citizen has on a state. That is not precarity disguised as freedom. It is freedom that is also precarity, the same arrangement seen from two sides, and it works beautifully right up until the single point of failure (a visa, a tax ruling, a currency) moves.

Third, and underneath both, the optimist case and the precarity case are describing the same facts, not competing ones. The technologies that democratize are the technologies that compress and dollarize. The AI that gives the freelancer global reach is the AI removing the rung he climbed in on. The stablecoin that saves the relocant's money is the instrument draining his host country's monetary control. Standing's whole argument was that this class trades security for a particular kind of autonomy; the honest synthesis is not that the optimists are wrong but that they are pricing the upside of that trade while its downside is still unbilled.

What fills the gap, and why the toolkit cannot reach it

Article 2 handed this article a question: why can a Western toolkit not simply be transplanted into a largely informal economy, and what fills the gap when it cannot? Watching it happen, the answer is that little public fills the gap. It gets filled privately, improvised and unequally, by four things.

Family, first and largest. For the developing-economy precariat the household does the job the welfare state does in Oslo. It is the most effective anti-poverty mechanism on the street outside my window, it appears in no policy catalogue because it cannot be legislated, and it is precisely what the diaspora and the nomad gave up to get the other dials. Its institutionalized, cross-border form is the remittance, and that flow is enormous: around 685 billion dollars in 2024 to low- and middle-income countries, larger than foreign direct investment and official aid to those countries combined, private family welfare wired home one transfer at a time.[29]

Cost-of-living arbitrage, second. Moving your costs into a cheaper jurisdiction is a private substitute for a raise, and the entire nomad economy runs on it. It is unavailable to anyone tied to a place by family, language, or a job that requires presence, which is most of the planet, and as the Da Nang rents show, it exports the problem to the host.

Stablecoins, third, and this is the one from an area I work in, so I have to be most careful to separate the real from the hype. For people the formal banking system has locked out, or whose currency is melting, dollar stablecoins have become a functioning financial rail rather than a speculation. The honest figures matter here. A16z's State of Crypto 2025 reports stablecoins moving 46 trillion dollars in gross transaction volume over the year, but most of that is high-frequency financial flow; on an adjusted basis that strips out bots and circular activity, the figure is about 9 trillion dollars (up 87%), and it is that adjusted number, running uncorrelated with crypto trading, that signals real use.[30] Nigeria runs tens of billions in stablecoin volume against currency devaluation; Argentina and Venezuela use dollar tokens as de facto savings accounts.[31] For the post-2022 Russian who cannot hold a working card, a USDT wallet is sometimes the account that does not get frozen.

I want to be precise rather than promotional, the same way I was about quadratic funding (the matching-fund mechanism behind public-goods grants) in Article 2, and the precision cuts against my own interest. This is a survival instrument for people the formal system shut out, not a redistribution tool, and it carries costs the boosters skip. Brett Scott's Cloudmoney is the corrective: digital money is also surveillable, freezable, intermediated money, and routing your life through dollar tokens trades one exclusion for a dependence on private issuers and the dollar.[32] The European Central Bank notes the evidence for systematic remittance use is thinner than the adoption headlines imply, and retail use is a small share of total volume.[33] And dollar stablecoins quietly dollarize the economies that adopt them, draining local monetary control and, in a crisis, accelerating the capital flight that breaks weak currencies further.[31] The system is no longer merely tolerating this exit; it is institutionalizing it. The US GENIUS Act, signed in July 2025, set the first federal rules for dollar stablecoins and was sold explicitly as a way to entrench the dollar's global role, which turns the precariat's improvised dollar-exit into an arm of US monetary strategy.[36] Weak-currency states can see the threat: the EU's MiCA rules cap dollar-stablecoin circulation, and the Philippine central bank is piloting a peso-pegged coin, defensive moves against a dollarization they cannot otherwise stop.[36] Both things are true at once: stablecoins route around a broken system for the individual and can deepen the dollarization that weakens it for everyone who stayed in the local currency. A private exit has public costs.

Jurisdiction shopping, fourth, and here the inequality shows up as a price, though not the way I first read it. The precariat's version of jurisdiction shopping is a SIM card, a tourist visa, and a wallet that does not freeze. Capital's version has a price list, and the list is rising, but it is worth being exact about why, because the easy reading ("prices rose, so demand rose") is wrong. The Caribbean citizenship-by-investment floor doubled to 200,000 dollars in 2024 not because buyers bid it up. Four of the five island states signed a coordinated price-floor agreement in March 2024 (the OECS Memorandum, with the fifth joining later) to end a race to the bottom, under heavy due-diligence pressure from the US and EU.[34] The EU closures are politics, not a demand curve: Spain ended its golden visa over housing affordability, Portugal removed the real-estate route, Greece raised thresholds.[35] The demand signal sits elsewhere and is clearer: digital-nomad-visa applications are surging (Spain's up 40% year on year), and more than 50 countries now compete for remote income.[19] Put the two together and the inequality sharpens. Demand for exit is rising across the board, while the cheap routes are being deliberately closed and the expensive ones professionalized. Escape is becoming a thing that real capital does through a rising, regulated price list, and that everyone else improvises with a 90-day stamp.

The structural answer to Article 2 follows from this. Every tool in the catalogue needs a state that can find you, a residence that ties you to it, and a formal income record it can act on, and the three worlds fail that test in three directions. The developing-economy precariat has the state and the residence but a thinner formal record, which is why the rich-world tools arrive late and the tool that actually fits informal economies (the direct cash transfer) goes ignored in Western debate, and it is the one world where the container is being built rather than dissolved. The diaspora has the formal skills but no state that claims them. The nomad has arranged, deliberately, to be claimable by no one. The toolkit is not wrong; it is drawn for a citizen-worker-residence relationship that, as the ILO numbers show, only ever covered a minority of the world, that two of these worlds have stepped outside of, and that the third is now trying to extend to people who never had it. To say little public fills the gap is not to say nothing could: migrant-portability agreements, social-security treaties, and the cash-transfer infrastructure that already works in informal economies are real options, just thin and rarely aimed here. What actually reaches all three worlds today is infrastructure, not policy: a phone, a wallet, a transfer that clears in seconds across a border a bank wire would take a week to cross. It is the plumbing the formal system stopped providing, rebuilt privately by the people it stopped serving, a thinner thing than a welfare state and the only thing currently on offer.

(This is the view from three worlds I can see from one street, not the only map. The largest denizen systems on earth, China's internal migrants under the hukou household-registration system and the Gulf states' kafala labor regimes, are bigger and harsher than anything here, and a complete account of the global precariat would have to include them. This article does not claim to be that account. It claims the structure is visible even from a café in Da Nang.)

My own position, stated plainly

This is the article where my vantage is the instrument, so I owe you an accounting of where I sit, because it is flattering and I will not hide that.

I am partial precariat with capital's option set. No pension, no national safety net I can reach, no employer past the current contract, which puts me on the same side of Article 1's asymmetry as the courier and the relocant. But the dials that matter are set in my favor: income in hard currency and crypto, costs in dong, a decade of arbitrage banked, the technical base to ride the AI wave up rather than be compressed by it, and enough mobility that no single jurisdiction's rule change ends me. I am the version of this story that worked, which is exactly the kind of example I have flagged before as needing a survivorship warning. So here it is: for everyone in my position who built something portable, there are many who tried the same move and ran out of runway in a country that owed them nothing. Reading only the nomads who made it flatters the model the way reading only the post-war welfare states flattered the toolkit in Article 2.

What the vantage buys is the thing no single-country observer has. The American adjunct, the Tbilisi relocant, the Manila call-center worker, and the Da Nang courier are not four separate problems. They are one structure, the splitting of citizen from worker from resident, experienced through different dial settings. From the café on An Thuong street, between all of them, you can see it is one picture. That is the reason this article had to come from here rather than from a university, and also the reason to distrust it a little: the man describing the structure is sitting in its most comfortable seat.

What this analysis produced

Most of this series mapped other people's mechanics and other people's tools. Drawing the view from the bottom produced three findings, and these are the ones I will carry into shorter pieces.

One: the same structural exposure produces opposite lives, and five dials decide which. Family network, cost arbitrage, currency stability, jurisdictional rights, and healthcare access are the settings that determine whether Article 1's channels land as catastrophe or as inconvenience, and formality is not among them, which is why a Manila call-center worker with a contract and family health insurance can be more exposed than a courier with neither. The precariat is a class defined not by income but by which dials you do not control.

Two: AI sorts the precariat by who it already was. The same tool that lets a non-developer ship software is the tool compressing the wage of the cognitive work the precariat sells. It is an on-ramp and a demolition crew at once, and which one you get is decided before you touch it, by your existing technical base, capital, and dial settings. AI does not lift or sink the precariat as a class. It widens the gap inside it, and the entry-level rung it removed was the one the Global South climbed on.

Three: when the public container is missing, exit privatizes, and the price of exit is the clearest inequality signal we have. Little public fills the gap; family, arbitrage, stablecoins, and jurisdiction shopping fill it instead, all private and all unequal. Capital exits through a rising, regulated price list (a Caribbean floor doubled to 200,000 dollars, EU golden visas closing). The precariat exits through a 90-day stamp and a wallet that does not freeze. The instinct to escape a jurisdiction is the same up and down the structure; the means are four orders of magnitude apart, and the cheap routes are the ones being closed.

The series ends on the same street it opened on. The four of us at that café are not a metaphor; we are four settings of one machine, and the thing none of us has, that our parents' generation mostly took for granted, is a place obligated to take us back. The courier's state is building him that floor now, late and leaky. The relocant's took it away. The nomad signed it away for the arbitrage. I am betting the rails I built privately outlast the public ones I never had.

Article 1 said the machinery is not a law of nature. Article 2 said the tools to slow it exist, several work, and most cannot reach the people who increasingly need them. This is what that looks like from inside the gap: not a class waiting for a policy, but a growing population wiring its own plumbing out of phones and wallets and cheaper countries. The map the rest of the debate is arguing over describes a container most of the world never lived inside to begin with.