Power Reliability, Sleep, and the Rupiah: Electricity Outages as a Productivity and Household-Cost Channel

Rupiah Stability Watch · 2026-08-23

The premise

A household does not experience the rupiah only at the exchange counter. It experiences it through the price of cooling a room, the cost of keeping a small shop open, the diesel behind a generator, the reliability of refrigeration, and the attention available after a broken night of sleep.

That is the narrow reason to add power reliability to Rupiah Stability Watch’s operating-status ledger. In earlier work, Currency Stress and the Rupiah Wellbeing Channel treated depreciation as a household and workplace stress channel. Heat, Water, and Work Hours and Hourly Heat Load and the Rupiah treated climate and cooling demand as productivity channels. Vehicle-to-Grid and the Rupiah, Reuse Before Replacement, and 3T Energy Readiness and the Rupiah treated distributed energy as resilience only when it reduces imported-fuel dependence rather than importing a new fragility.

This piece connects those threads at the level of daily operating capacity. If power unreliability rises during heat, haze, fuel-price pressure, or local grid stress, the first signal may not be a currency move. It may be tired students, shorter shop hours, spoiled cold-chain goods, diesel purchases, and a wider gap between households that can buy backup power and those that cannot.

What the evidence supports

The strongest current prompt comes from South Africa, not Indonesia, and should be used carefully. A 2026 Sleep Health study on energy insecurity and load shedding in South Africa describes scheduled power outages as a sleep-health problem, acting through environmental disruption, psychological strain, and changed behavior. A separate Sleep Health item on university students, surfaced this week, frames load shedding as a problem for lighting, room temperature, generator noise, and student sleep. A 2024 South African university study in the Journal of Student Affairs in Africa similarly found missed lectures, limited resource access, assignment disruption, stress, anxiety, isolation, and disturbed sleep patterns among students facing load shedding.

None of this proves the same magnitudes in Indonesia. It does support a modest mechanism: electricity interruptions can reduce human operating capacity before they appear in national financial aggregates. The human channel is plausible where outages coincide with heat, online learning, evening study, refrigeration needs, or noise from backup generators.

Indonesia-specific firm evidence points in the same direction, though the freshest public data are not always detailed. World Bank Enterprise Survey data for Indonesia show the share of firms experiencing electrical outages fell materially over time: 45.1 percent in 2009, 22.5 percent in 2015, and 12.7 percent in 2023. That is a sign of improvement, not deterioration. Older Indonesia survey indicators still show why the residual problem matters: in 2015, affected firms reported losses equal to 1.9 percent of annual sales; 11.7 percent of firms owned or shared a generator; and among firms using generators, generator power supplied an average 16.5 percent of electricity.

That older generator evidence should not be over-read as the 2026 condition. It is useful because it names the channel: when grid reliability is imperfect, some firms replace public electricity with private fuel-backed electricity. That replacement can preserve operations, but it changes the cost base. If backup power relies on imported fuel or dollar-priced equipment, outage resilience itself can become a currency-exposure channel.

The fuel side matters because Indonesia remains exposed to oil and product imports. Public summaries of BPS data report 2025 crude-oil imports of 17.58 million tons worth US$9.31 billion, while BPS’s own petroleum import tables track crude and petroleum products by origin through 2024. Separate public reporting says Indonesia’s 2026 energy subsidy and compensation envelope remains large, with nearly Rp400 trillion initially allocated for energy subsidies and compensation. These figures do not say outages are moving the rupiah. They say fuel-linked backup, electricity compensation, and household energy bills sit in the same external-balance and fiscal neighborhood as the rupiah.

The electricity system is also structurally important because Indonesia’s power generation still rests heavily on coal. The IEA’s Indonesia electricity profile lists coal as the largest source of electricity generation in 2023, at 69 percent of total generation. Coal dominance does not mean the system is unreliable. It does mean that fuel logistics, plant availability, reserve margins, heat-driven demand, and policy-managed tariffs belong in the same operating picture.

The rupiah transmission channel

There are five channels worth watching, each with a different evidentiary threshold.

First is the imported-fuel backup channel. When households, shops, clinics, factories, or data rooms respond to outages with diesel generators, the private solution can preserve output while increasing fuel demand, maintenance costs, noise, and exposure to dollar-priced inputs. This channel is rupiah-relevant only if it scales, concentrates in strategic sectors, or overlaps with broader fuel-import stress.

Second is the heat and cooling channel. Outages during hot evenings have a different human cost than outages during mild hours. Cooling loss can reduce sleep quality, food safety, and workplace recovery. This is the point where the earlier heat-load ledger matters: hourly heat, not only daily average temperature, changes the severity of power interruptions.

Third is the small-firm continuity channel. A small business can lose sales without appearing in large industrial production data. A barber, food stall, cold drink seller, pharmacy, tailoring shop, or small processor may lose hours, stock quality, or customer confidence. The older World Bank indicators show that affected Indonesian firms once reported measurable sales losses from outages; the question for 2026 is where such losses still cluster.

Fourth is the learning and attention channel. The South African student evidence should not be imported mechanically, but it usefully widens the ledger. If evening outages disturb study routines, online access, room temperature, and sleep, the effect is not just educational inconvenience. It is next-day attention, exam readiness, workplace concentration, and family stress.

Fifth is the fiscal and tariff channel. Indonesia’s electricity and fuel prices are policy-managed in ways that can protect households. That protection can be socially stabilizing. It can also shift pressure onto the budget, PLN, Pertamina, or compensation timing when global fuel prices and the rupiah move together. The least-harm reading is not that subsidies are good or bad in the abstract. It is that the operating ledger should track where protection is paid for, when, and by whom.

What the evidence does not support

The evidence does not support saying that electricity outages are currently moving USD/IDR. The available public record supports a channel, not a market claim.

It also does not support treating South African load-shedding findings as an Indonesian estimate. South Africa’s outage regime, institutions, household backup patterns, student housing, and grid history differ from Indonesia’s. The useful transfer is conceptual: energy insecurity can affect sleep, learning, mental strain, and productivity.

The evidence does not support individual financial advice. A household or firm deciding whether to buy a generator, battery, rooftop solar, or efficiency upgrade faces location-specific reliability, cash-flow, safety, and maintenance constraints.

Finally, the evidence does not support a single national conclusion. Indonesia’s grid experience is geographically uneven. The rupiah-relevant question is not “does Indonesia have outages?” but “where are interruptions becoming costly enough to alter fuel demand, household stress, food safety, firm hours, or confidence?”

The least-harm reading

Power reliability should be monitored as a human operating-capacity variable. That is a narrower and more useful claim than calling it a currency shock.

A proportional reading would treat reliability data as early-warning evidence. If outage frequency and duration remain contained, the rupiah ledger records resilience. If interruptions cluster during heat, fuel-price pressure, logistics strain, or local disasters, the ledger should widen to include sleep, cold chains, small-firm hours, diesel backup, and public compensation.

This approach also avoids a false choice between grid investment and household adaptation. Some backup systems reduce suffering and keep essential services running. But backup that depends on imported diesel can reduce one fragility while adding another. Resilience is strongest when it protects the household, the clinic, the school, and the external balance at the same time.

Watchlist

For Rupiah Stability Watch, the observable watchlist is practical:

• PLN reliability metrics: SAIDI and SAIFI, by region where available, not only national averages.
• Outage timing: evening and hot-hour interruptions matter more for sleep, cooling, and study than equal-duration outages at lower-stress hours.
• Reserve margins and plant availability: especially where heat, coal logistics, maintenance, or local demand surges coincide.
• Diesel-generator use: fuel sales, generator imports, business backup-power spending, and noise complaints where data exist.
• Cooling load: air-conditioning demand, peak-load records, and household electricity bills during heat episodes.
• Cold-chain failures: food, vaccines, medicines, fisheries, and school-feeding logistics.
• School and work disruption: absenteeism, late assignments, reduced operating hours, and household survey signals.
• Fiscal pressure: fuel and electricity subsidy or compensation payments to Pertamina and PLN, especially when the rupiah weakens or global oil prices rise.

What I am uncertain about

The largest uncertainty is Indonesia’s current subnational outage pattern. National improvement can coexist with local fragility in islands, remote areas, industrial clusters, disaster zones, or fast-growing demand centers.

The second uncertainty is household adaptation. Some families and firms may already have batteries, generators, rooftop solar, or flexible routines. Others may have none. The welfare impact depends on that distribution.

The third uncertainty is the size of the sleep and attention channel in Indonesia. The South African evidence makes the channel visible, but Indonesia-specific sleep, schooling, and workplace data would be needed before assigning a number.

For now, the responsible conclusion is modest. Power reliability belongs on the rupiah operating ledger because it can convert exchange-rate, fuel, heat, and infrastructure stress into human fatigue and dollar-linked backup costs before a visible market break appears. It is an early-warning channel, not proof of a currency shock.