Policy brief

Sidi Bouzid's Solar Paradox: Diffusion Without a State

Seven points to understand how 3,201 solar pumps appeared in Sidi Bouzid without public subsidy, and what the 2025 Finance Law still needs to fix.

  • agriculture
  • water
  • governance
  • policymakers
  • researchers
  • tunisia
  • north-africa

Finding

In Sidi Bouzid, 95% of 3,201 solar pumps appeared in just five years (2021–2025) and left no trace in public subsidy statistics: the transition unfolded outside any institutional steering.

Recommendation

Article 81 of Tunisia's 2025 Finance Law must be paired with three simple requirements — meter, extraction quota tied to recharge, and a public solar-well cadastre — or regularization will legalize overexploitation instead of ending it.

Key findings

  • Between 2013 and 2025, satellite imagery detected 3,201 solar-equipped wells in Sidi Bouzid governorate alone.
  • 95% of them were installed between 2021 and 2025; the single year 2023 accounts for 1,704 installations — more than half the total.
  • 0% of farmers who installed a solar pump used bank credit (GIZ/ANME survey); 63% paid cash, 38% used informal supplier credit.
  • Statistical analysis attributes almost all the explanatory weight to a single factor — whether a nearby neighbor has already installed one — and finds the influence of institutional channels to be indistinguishable from zero.
  • In adoption clusters, deep-aquifer extraction reaches roughly 140% of the recharge rate; 8 farmers in 10 pump more after installing solar.
  • Article 81 of the 2025 Finance Law (December 2024) opens a fiscal regularization route (TND 2,000–2,500 per solar well) without imposing metering, quotas, or a cadastre.

Recommendations

For policymakers

Make eligibility for regularization conditional on the installation of a volumetric meter.

Tie each regularized well's extraction quota to the measured recharge rate of the relevant aquifer.

Build a public solar-well cadastre from regularization filings, cross-checked against the satellite imagery already available.

For researchers

Use the December 2024 turn as a natural experiment to measure whether the state becomes a meaningful actor in diffusion from 2025 onward.

For civil society

Document water budgets at the well-cluster level; effective regulation must operate below the governorate.

Methodology

Satellite census of solar pumps (Sentinel-2, 2013–2025) over 13,201 points (3,201 solar wells, 10,000 controls) in Sidi Bouzid governorate, combined with spatio-temporal survival analysis and a diffusion model. Neighborhood effects estimated within a 5 km radius. Results last verified 2026-01-24.

Questions this brief answers

  • How many solar pumps were installed in Sidi Bouzid, and when?
  • Who actually drives adoption — the state, the bank, or the neighbor?
  • What is the water cost of this transition?
  • What does the 2025 Finance Law still need to fix to turn regularization into governance?

Download PDF →

Executive summary

Between 2013 and 2025, satellite imagery counted 3,201 solar pumps in Sidi Bouzid governorate alone. Ninety-five percent of them appeared in five years, and more than half during 2023 alone. No public subsidy steered this movement: 0% of the farmers involved obtained bank credit, and statistical analysis attributes almost all the adoption decision to a single factor — has a nearby neighbor already installed one?. Where these pumps cluster, deep-aquifer extraction reaches roughly 140% of the recharge rate. Article 81 of the 2025 Finance Law opens, for the first time, a fiscal regularization route for solar wells. But without a meter, a quota, and a cadastre, that regularization will lock in overexploitation instead of ending it. The window to add those three requirements closes with the 2025–2026 implementation calendar.

1. The ground: an agriculture hanging from deep aquifers

Sidi Bouzid is one of central Tunisia’s agricultural centers. Private irrigated area there grew from 2,000 hectares in the 1960s to more than 47,000 hectares by 2010, almost entirely on deep boreholes. Officially, much of that infrastructure sits in safeguard zones where new permits have not been issued for years. In practice, the national inventory counts 32,575 unauthorized boreholes out of 48,762 deep boreholes (66.8%), withdrawing roughly 800.6 million cubic meters per year.

It is on this base of tolerated illegality that a solar transition has grafted itself. This brief does not invent Sidi Bouzid’s water crisis; it documents who is steering — or not steering — its new phase.

2. The surge: 3,201 pumps in twelve years, 1,704 of them in 2023

The yearly distribution of installations is highly skewed. Over the pioneer period 2013–2020, only 157 pumps were installed (4.9% of the total). The boom period 2021–2025 concentrates 3,044 pumps (95.1%), with a peak in 2023 (1,704 installations in a single year). This is not an ordinary diffusion trajectory; it is a transition that took place almost entirely in five years.

PeriodYearsSolar pumpsShare
Pioneer2013–20201574.9%
Boom2021–20253,04495.1%

No corresponding signal appears in ANME subsidy statistics or in agricultural credit portfolios. A transition of this scale, completed in five years, leaving no trace in public steering tools, is a political fact in itself.

3. Where they appear: not anywhere — in clusters

The pumps are not evenly distributed. They concentrate where land is flat, access is easy, and water is reachable — and above all where other pumps have already been installed. Three geographic variables show up in every model: slope, travel time to town, elevation. But their combined weight is smaller than that of the immediate neighborhood.

At the governorate scale, spatial clustering is statistically very strong. For regulation, this means no policy designed at the governorate level can be effective: the dynamics play out in clusters of a few kilometers.

4. The official channel: almost no one came through it

If the state had steered this transition, traces would show up in financial data. Instead:

  • 0% of farmers who installed a solar pump financed it through bank credit.
  • 63% paid cash.
  • 38% used informal supplier credit.
  • The vast majority of panel purchases moved through unofficial channels — including across borders.

The cause is known. The ANME and APIA subsidies are conditional on a Water Extraction Authorization issued by the regional agricultural development commissariat. In the regions where demand is highest, those authorizations have not been issued for years, because the aquifers are already classified as overexploited. The public subsidy is therefore tied to a door the administration has itself closed. The result is mechanical: farmers exit the formal channel and buy elsewhere.

5. What actually triggers adoption: the neighbor

Once the effect of terrain and accessibility is removed, the variable that best explains solar pump adoption is the presence of an already-equipped neighbor within five kilometers. The influence of institutional channels — campaigns, agencies, banks, extension services — is, conversely, statistically indistinguishable from zero.

The verdict, in the brief’s plain language:

The analysis adds a subtler point: past a local threshold, neighborhood density actually slows down further adoption. This is the saturation effect of the shared aquifer — there is no longer enough water for a new entrant, even if the technology is available. Social diffusion is thus self-limited by the resource itself.

6. The hidden cost: the water table falls faster than rainfall refills it

The technology works. That is precisely the problem. Once the solar pump is installed, energy is free at the margin. The GIZ/ANME survey confirms that 8 farmers in 10 increased their water use after installation, and that 88% report higher short-term yields.

In the most stressed aquifers, extraction reaches roughly 140% of the recharge rate. Put plainly: for every cubic meter that rainfall restores, about 1.4 cubic meters are pumped out. That gap is the booking of a water debt that will manifest, within five to ten years, as a sustained drop in the water table and the progressive failure of the shallowest wells — first those of the smallest farmers.

This is what one observes when an energy-transition instrument is deployed without a framework to govern the underlying resource.

7. The 2025 turn: recognition is not enough — a framework is needed

Article 81 of the 2025 Finance Law, adopted in December 2024, marks a doctrinal shift. For the first time, Tunisian law:

  • defines a distinct fiscal category for “solar-equipped wells”;
  • opens a retroactive regularization route against a fee of TND 2,000 to 2,500;
  • effectively recognizes the existence of infrastructure the administration had, until then, chosen not to see.

This turn is necessary. It is not sufficient. As written, Article 81 provides for no metering, no extraction quota, no cadastral integration. A purely fiscal regularization converts an illegal fait accompli into a legal fait accompli, without resolving anything about the water imbalance.

Three simple additions

For regularization to become a governance instrument rather than a one-off revenue measure, three requirements can be added in the implementing decree:

  1. Meter. Make eligibility conditional on the installation of a volumetric meter on each regularized well. The marginal cost is modest relative to the TND 2,000–2,500 fee already collected.
  2. Quota. Cap permitted extraction on the measured recharge rate of the relevant aquifer, with periodic revision. Without a quota, the meter is mere instrumentation.
  3. Cadastre. Build a public registry of solar wells out of regularization filings, cross-checked with the satellite imagery this study has produced. That cadastre is the basic infrastructure of a sub-governorate water policy.

None of these three additions requires a new law. They can be written into the implementing decree of Article 81. The 2025–2026 calendar is the window in which they can still be added without unwinding the political compromise that the article’s adoption represented.

Limitations

This analysis covers Sidi Bouzid governorate only. Generalization to Kairouan, Kasserine, or Cap Bon requires replication. Satellite detection of solar pumps, validated on a sub-sample, may under-detect installations under tree cover. The statistical models used establish the absence of the state in the observed diffusion; they do not, by themselves, establish the cause of that absence — the “authorization trap” diagnosis presented in section 4 rests on a complementary analysis of the administrative architecture.

Sources and data access

This brief is derived from the preprint:

Gasmi T., Guesmi R., Ben Abdelbari S., Hammami G. (2026). The Solar Paradox: Pure Social Diffusion and Competitive Resource Capture in Semi-Arid Irrigated Land Expansion. EarthArxiv. https://doi.org/10.31223/X5ST97

Data and code available from the authors. Additional sources:

  • Le Quotidien (2023), Prolifération des forages sauvages.
  • Cahiers Agricultures (2023), L’agriculture irriguée en Tunisie.
  • GIZ/ANME, SPIS survey, UNECE 2019 presentation.
  • ONAGRI (2020), EAU 2050 — Stage 3, Volume 2.
  • IWMI (2017), Groundwater Governance in the Arab World.
  • Republic of Tunisia, Article 81, 2025 Finance Law, adopted December 2024.
  • AI-native access to ONAGRI agricultural data: agridata MCP — the studio’s Model Context Protocol server that exposes the agridata.tn portal directly to LLM clients (source code on GitHub).