I'm guessing: fewer people buying from the power companies/grid => the fixed costs of these companies are pushed onto the poorer customers, who already couldn't afford much.
But there is a bit more. Almost all power plants in Pakistan are built with state-backed dollar-denominated loans (reason govt incompetence+corruption). This means if grid demand goes down, power plants don't go out of business like they would in a market based system. Instead, they keep collecting dollar-denominated interest paid by the state, even if they produce zero power.
The state mitigates this by increasing electricity prices (in rupees). I have forgotten how this helps.
The reason power plants in Pakistan probably require this kind of financing is because Pakistan doesn't have the industrial capability to make the equipment that you need to build a power plant, so, dollars are a requirement.
Power companies in Pakistan also don't have easy access to international money markets, and thus, it makes sense for the government to back those strong currency loans as a subsidy on infrastructure.
This is not exclusive to Pakistan, this is the routine of infrastructure financing on developing countries. J.P. Morgan is not really eager to lend money for PakiPower Incorporated, but it is willing to lend to the government.
It is unfortunate that the government of Pakistan and their investors (China and the IMF) made poor investment decisions. They should feel free to go back to debt holders to renegotiate the debt, or default on it and hand the stranded assets back to creditors. The death spiral is of their own making, and will only accelerate as solar PV and battery cost declines continue. Electricity consumers will simply go off the grid. Such is the risk of unsophisticated investors not understanding the market in which they invest. Capital being at risk is an inherent component of investment.
My condolences and sympathy to the people of Pakistan caught in the mess. The global energy transition will be volatile.
Who owns the distressed fossil generation collateralized debt? China. Where is Pakistan importing cleantech from? China. There is some IMF debt in there as well, for accuracy.
> default on it and hand the stranded assets back to creditors
I doubt the debt is secured by the power assets. If anything, maybe China can assume ownership of the entire powerplant if Pakistan cannot pay. They have done that many times in Africa. See: "debt-trap diplomacy". Also, it is terrible advice to tell a country to default on external debt. See: Argentina!
Usually there are three parties in these agreements.
1. State of Pakistan
2. Someone with dollars (the investors)
3. Local businessman who are willing run the power plant.
The three parties come to an agreement on what the minimum returns should be on the investment. Say 10% annual. Then the investors give money to the businessman, who then import the power plant equipment and start operating it. The state-run electricity distribution companies buys from the power plant as needed and pays them the unit price set by the State of Pakistan. Part of this is converted into dollars at some pre-agreed rate and transferred to the investors.
In all this, if the total returns to the investor are above 10%, then all is good. However, if the grid demand has fallen, and the distribution company didn't buy a lot of units from the power plant, then the State of Pakistan has to step in and give the investors the difference to make up the 10% returns.
State capitalism like you described totally undermines the price system by replacing profit-and-loss–guided entrepreneurial calculation with political allocation of resources, thereby rendering economic calculation increasingly impossible and eroding the coordinating function of the market process.
Yes, but nobody has found a more effective way to build infrastructure in poor countries. State capitalism as described is how infrastructure development happened in Indonesia, Malaysia, Taiwan, Hong Kong, Korea, Japan, Vietnam, Thailand, etc.
The fact that infrastructure was built under state capitalism does not demonstrate the superiority of central planning, only that capital accumulation occurred despite intervention, often financed by prior scarcity, foreign savings, or coerced transfers; absent market prices and entrepreneurial profit-and-loss, the state cannot know whether the infrastructure created was the most value-productive use of scarce resources, only that concrete and steel were poured.
I think it demonstrates the increased variance of central planning. The Congo Free State was also centrally planned, and so was the Holocaust, the Holodomor, the Armenian Genocide, Suharto's mass murder of suspected PKI sympathizers, etc. But the expected outcome for poor countries is that they stay poor and don't develop into industrialized export giants the way my laundry list of countries did.
Higher variance isn’t a redeeming feature when the mechanism that generates it lacks rational calculation in the first place. Central direction can occasionally coincide with growth in poor countries because initial scarcity leaves many wasteful paths that still raise output, but that doesn’t establish a positive expected value
I’m not arguing that centralized or state-capitalist systems “never work” in the sense that nothing gets built, or that output can’t rise. Clearly roads, ports, power plants, and factories were constructed in many of the cases you listed.
The narrower point I’m making is about economic rationality. Without market prices for capital goods generated through profit-and-loss entrepreneurship, there is no way to know whether those projects were the best use of scarce resources, or merely a use that happened to raise output from a very low baseline.
In very poor countries, almost any large capital investment will increase measured output because there are so many unmet needs. That means growth can occur even under badly misallocated investment. The fact that development happened does not tell us whether it happened efficiently, or whether alternative decentralized uses of those same resources would have generated more value.
That’s also why I don’t find higher variance persuasive as a defense. Occasional success doesn’t validate a mechanism that lacks systematic feedback. Without prices and profits, planners can’t distinguish luck from competence, or learning from error. Things such as malinvestment and moral hazard result. You only know concrete and steel were poured, not whether society is richer than it otherwise would have been.
So my claim isn’t state capitalism always fails, nor is it a moral argument about atrocities. It’s that infrastructure success alone doesn’t answer the calculation problem. Growth from scarcity is compatible with irrational allocation, and therefore doesn’t establish a positive expected value for centralized direction as a general development strategy.
They surely were not the best use of scarce resources. However, I don't know if you've ever tried to run a business in a poor country. It turns out that the decentralized economic systems they have are also not economically rational, systematically failing to provide functioning market mechanisms that support long-term investments such as highway systems, reliable electric grids, municipal water treatment, etc., even when those things would be highly efficient uses of scarce resources. Rather, generally speaking, they systematically squander their resources in order to stabilize the existing socioeconomic power structures.
Generally speaking, if you invest your money in a historically kleptocratic country, you can expect your investment to get confiscated if it's profitable, and possibly even if it's not, which is what happened to my retirement savings. Even if you make your investment at a time when the country is governed by non-kleptocrats, you will probably lose it after the next coup or election in which new kleptocrats come to power.
In that environment, where private investment in long-term infrastructure projects is irrational and languishing in poverty for many generations is the normal state of affairs, state capitalism frequently works.
I don't think moral arguments about atrocities are somehow orthogonal here. Power plants and electrical grids are often worthwhile investments, not because building monuments to Westinghouse is a pious sacrifice that pleases the electrical gods, but because they promote human welfare by providing material abundance. That's how we measure whether society is richer, not, as you say, by the amount of concrete poured. If human welfare is your yardstick, the possibility of economic catastrophes like the Holodomor greatly diminishing human welfare must necessarily weigh on the negative side of the balance. The inhabitants of Auschwitz and the Congo Free State were not enjoying even the material abundance they had enjoyed previously.
So we know that central planning carries risks to human welfare that decentralized systems do not. However, it also has opportunities to promote human welfare that decentralized systems do not. The variance is larger. I don't think we know enough to measure the expectation.
Sure, running a business in kleptocratic hellholes is a nightmare. Confiscation risks kill private investment, especially for big infrastructure that needs stable property rights and enforceable contracts to make sense.
But that's not a bug of decentralized markets; it's the poison of political interference and weak institutions, which state capitalism only doubles down on by swapping entrepreneurial discovery for bureaucratic fiat, still without solving the calculation problem or providing systematic feedback beyond "stuff got built."
Look at Hong Kong in your list: it boomed precisely because of its hands-off, free-market approach with top-tier economic freedom rankings, unlike more interventionist tigers that rode credit-fueled waves but crashed in '97 or stagnated like Japan post-bubble.
Atrocities are the dark side of concentrated power touted as enabling opportunities, jacking up variance with no way to gauge if human welfare gains beat the unseen costs of foregone innovations.
Without prices and profits guiding resources, we're left guessing expectations, but history shows freer systems deliver sustainable abundance when institutions let them, not coerced escapes that often loop back to poverty or worse.
This is the right way to think about the problem if you're faced with the problem of what country to start an electric power company in. But it's not very helpful if you're faced with the problem of how to govern Pakistan, which is certainly not a hellhole but does suffer major corruption problems. The political interference, weak institutions, concentrated power, and official corruption are largely givens; you can work to change them incrementally, but you can't just import the government of Hong Kong. Even if you could, you couldn't convince investors that the reforms would stick this time, for real.
Perhaps, if you could obtain political power that you could retain stably for decades, you could make pretty big changes there, but only at the cost of further concentrating power, creating opportunities for even greater corruption profits for whoever can wrest power away from you. Nobody has ever held power for decades in Pakistan's history. Even Nawaz Sharif didn't make it to 10 years in power over his three (non-consecutive!) terms. If you simply liberalize economically without eliminating the confiscation risks that kill private investment, private investment will not magically materialize without the private investors, who are sensibly investing their savings in a 7-11 franchise in Cleveland, Ohio.
So, what can you, hypothetically governing the country, do under these constraints?
A proven strategy is public investment, like Airbus, like the military contracts that sustain Boeing, like the Hoover Dam, like the Rural Electrification Administration, like federally guaranteed student loans, like the interstate highway system, like Volkswagen, like Tupolev, like Rosatom, like Industrias Aeronáuticas y Mecánicas del Estado, like the Apollo Program, like Huawei, like Westinghouse's nuclear power division, like the ARPANet and NSFNet. Certainly the money won't be invested as wisely as if savvy entrepreneurs like Warren Buffett were directing it, and neither Warren Buffett nor Juan Perón is going to do a good job at investing in unforeseen innovations; but, even if most of that investment is wasted like Project Huemul, it can still dramatically augment the economic productivity of the country, under circumstances when private investment is unavailable. Often such productive capacity will eventually make the country more appealing to private investors, but that can take a long time.
In this situation, you aren't faced with the choice between state capitalism and regular capitalism. You're faced with the choice between state capitalism and no capitalism.
Thus, state-guaranteed loans to build power stations.
Catlover76 asks in a [dead]ed comment, "And China, right?" It's a reasonable question. It's debatable whether the infrastructure of the parts of China I didn't mention was built by state capitalism or by a straightforwardly Communist system of production, so I only mentioned the more clear-cut cases.
I heard that they are trying to restructure the billing in this way for next fiscal year (July 2026- ), but its really difficult to find a non-regressive scheme. Electricity per-unit prices in Pakistan are set by the government, they vary depending on how much you consume [1], and they play a pretty significant role in government popularity.
[1] There is a price for the first 50 units you consume, then a higher price for the next 150 units, etc. Similar system to income taxes.
Grids in Germany if you're not a "typical household/office" with therefore atypical grid usage bill for peak power and energy separately; the billing related peak power is measured by averaging power over 15 minute chunks, and taking the worst one of a year.
Alternatively it's also practical for such solar situations to bill for market rate price of the energy in each 15 minute chunk separately; this doesn't correctly attribute transformer and other transmission equipment expenses between solar houses and non-solar houses, but it's still handling the grid tie solar load on the grid's power plants during periods of very little sun.
> averaging power over 15 minute chunks, and taking the worst one of a year.
What an interesting metric. Wouldn't even a very cheap and small battery (definitely small enough to keep inside an appartment) provide enough smoothing to, like, halve this peak number? You could rig it to not even output energy until you are beyond the current year's peak usage... How much money would you save this way?
I just feel this number is so prone to small mistakes (grandma plugs in the wrong things at the wrong times) and hacks (like the above) that the relationship between users' reward/punishment and the grid's health seems wildly disproportionate.
> market rate price of the energy in each 15 minute chunk separately
I am currently on a plan with 5 minute market rates, can buy and sell in (sell prices can go negative - as can buy, actually), all automated. At least I feel we am working with the grid, not against it, and we make a small net profit (before depreciation).
> relationship between users' reward/punishment and the grid's health seems wildly disproportionate.
It's still much closer to the real costs for the grid operator than $/kWh. The fundamental problem that rooftop solar has revealed is that people think they are paying for the electricity, but they are not. Electricity is dirt cheap. Most of what they are paying for is the maintenance of the grid, and simple usage based billing crushes the system because of freeloader problem once rooftop solar is added.
Long term, the likely thing you pay for will be the size of the main fuse that connects you to the grid. Because that's the thing that scales with the costs you impose on the operator.
Actually the local cost is not the fuse size, but how much smaller the first transformer after you could be if you weren't there.
Though it's often more fair to determine such for each user; then take those as a relative scale, then split the transformer's actual TCO by the determined share sizes between the users. Because the first user needs the transformer to it's peak size; the second only by the instantaneous-added peak size, which is lower as they won't use it peak at the same time.
> the first user needs the transformer to it's peak size; the second only by the instantaneous-added peak size
Of course, how does the electricity company determine which user was first in this situation. A tariff that depends on the order of connection may not be practical for domestic situations, although it may be OK for very large users, e.g. factories, data-centres.
Using fuse size seems a more reasonable and fair proxy for cost, assuming the same load patterns as the rest of the users. Then again, consumers with EVs might argue that their load pattern is different to the average user (e.g. filling up with off-peak electricity). Also consumers with air conditioning might argue for special treatment given their usage correlates with solar output (except where it does not).
You don't; you let the second one pay more than what the marginal cost was, in order to make the first one not pay so disproportionally.
That only has to happen once there is a second one, though.
Punishing a mere "large enough to not worry about popping it" fuse by billing shared infrastructure based on it (not just billing the stub line from the main in the street to the fuse/meter box in one's home in relation to what wire gauge is needed based on the fuse choosen) is pretty stifling.
If e.g. your furnace fails in the middle of the winter and the repair guy tells you it needs replacing, you might want to get some space heaters and run them for a few days until your actually-wanted new furnace/heatpump/whatever can get delivered, instead of having to get installed whatever the HVAC guy has in the local storage, because if you wait more than on the order of 12 hours, you'll start to get frost damage from pipes and such.
Having to be beholden to an electricity company having time to upgrade your fuses on such short notice so that you can plug in the space heaters without blowing them might be a problem.
But paying say 300 bucks extra because you did that for like 3 days or so would easily be cheaper than the cost of temporarily installing an available loan furnace and then having to remove it again to make way for the actually-wanted one.
They do though bill you if you make them dig the street up to say pull a medium voltage line into your factory that previously just got low voltage from a shared street transformer, but now that you've plans to use a lot more, you'd need the higher feed. Then they bill you and if within like 10 years or so someone else orders service that can piggyback on what capex you paid for, then you'd get a proportional refund from them having to pay off part of your share.
But that's not for just getting normal basic electric service to a normal residential building in a city, that's for building a new farmhouse on the other end of some field where there never was electricity, or for getting unusual service that wouldn't be in the street if you didn't request it.
Merely sizing the transformers/substations to handle the aggregate current of the users attached is not typically handled by the above mechanism, especially because it only covers initial buildout.
This is how it works in Japan for the newer rate plans for consumers now (replacing the previous method of charging you based on the size of your main breaker), but checked in 30 minute increments rather than 15.
The steps are pretty coarse - on my rate plan there are just 3 steps: 0-10 kW, 11-15 kW, 15 kW+. You're not going to surpass peak 10 kW in an apartment anyway.
It's legacy tactics; against the hacking the comparable thing for internet connections has historically been iirc 5 minute chunks and then taking the 95th percentile (like, charging not the highest, but the one 5% away from the highest). Not sure about the 5 min aggregation tbh.
The 15 minute chunks are due to the German and much of the European grid market being in that chunk size.
I don't think it's necessarily impossible to get that billing model as a household; it's just not an interesting one to have as it's not competitive for the usage patterns of a household.
It's good until a Wednesday afternoon your home system dies and you have no electricity until Monday. I guess more people would prefer to pay a $10 or $20 monthly fee just in case.
> I guess more people would prefer to pay a $10 or $20 monthly fee just in case.
The grid becomes an insurance policy. In that case it is justified to ask for the insured party to pay their share of the system costs; both an energy fee and transmission/distribution/generation capacity fee.
I think most places the service is priced under the assumption that usage is enough to pay for the grid…
I’ve only ever rented though. Are connection fees something that homeworkers think about?
Possibly we will have to see changes to account for this sort of stuff at a more granular level, as the grid becomes more dynamic. But, that’s a future we should be actively looking to design for, as the energy supply mix is going to change whatever anybody thinks about that. Can’t beat energy falling from the sky, on price…
Is the €30 usage fee going directly to the producer of electricity, or is part of it a variable transmission fee that goes to the network operator?
My monthly electricity bill in Sweden, averaged over a year to 1600KWh/month, is approximately €90 production, €50 transmission fee, €25 fixed connection-size fee (25A, 400V), €70 national electricity tax and €50 VAT for a total of €285/month.
We'll be moved to yearly-peak-based transmission tariff in 2027 (European law), but for now I don't need to worry about plugging in the car to chargeon cold days or taking shower when someone is cooking.
Both, currently; notably it's mostly not what goes to your local grid, but rather mostly to the larger scale grid.
It's about a 60/40 to 70/30 split between production/"grid-usage-fee" ("Netznutzungsentgelt").
It basically pays off the grid stability provision bids for fast-response power, and the transmission itself.
It'd likely be helpful if the peak part could be regulated in a way that's more condusive to match the actual impact you create on transformer sizing, not the worst-case impact you might have.
Because there's a difference between a mostly-uncorrelated peak of shower+cooking vs. the car+cold day, because your neighbours don't shower the same time, but the several hours of charging do often overlap and the cold is the same across a neighbourhood that shares a local substation.
But yeah, for the most part, transformer size isn't that large of a contributor to overall electricity provision expenses, so I don't expect that to be a significant problem by that 2027 law.
its easily fixable, utility company can charge fee for fixed cost those who connected to the grid, and if all rich decided to disconnect, then they disconnect neighborhood eliminating fixed cost.
