At the end of each year, attention turns to the finance law, one of the main documents showing where the state is heading in the year ahead: expected revenue, planned spending and financing needs. In Tunisia the document matters even more because public finances have been in crisis for years, a crisis tied to structural imbalances that have built up over time.
In the current situation, though, it is hard to expect the state to take on the broad structural reforms that are now necessary, such as reviewing the subsidy system, opening the file of public enterprises, and reconsidering the size and cost of the civil service. Economic, social and political reasons make these reforms costly and complicated to carry out.
Earlier articles covered some of these problems: subsidies, public enterprises and public finance imbalances. This article takes a different angle and tries to assess how well the state can anticipate the path of its own public finances. We study the accuracy of Tunisia’s finance laws since 2014 by comparing what was planned at the start of each year with what was actually achieved, especially for revenue, spending and financing. We also try to identify the areas where forecasts were more accurate and the periods when the pattern of these errors changed.
Figure 1. Initial finance law and outturn, 2014-2025
Billions of dinars. Pick an aggregate, then hover over or tap a year. Green: revenue above the finance law, or spending, deficit, borrowing and debt below it. Red: the reverse.
Gap = (outturn − initial finance law) / initial finance law. The 2025 outturns may be provisional. Source: Ministry of Finance (finance laws and budget execution results); author's calculations.
Revenue
Taken together, taxes show a relatively small gap: its mean absolute value was 5.3% over 2014-2025. That does not mean revenue can always be estimated, or that the forecast held in the same way every year. Actual receipts were above estimates in 2014, 2018, 2019, 2021 and 2022 and below them in other years. In 2020 the Covid shock cut tax revenue by about 4.6 billion dinars relative to the forecast, which is understandable given how economic shocks change activity and the state’s resources.
The picture differs by tax. The mean absolute gap was about 6.5% for personal income tax, against 16.9% for corporate tax. Corporate tax follows company profits, which move with the business cycle, prices and the cost of financing. It also depends on the administration’s capacity to collect, and tax evasion adds to the inaccuracy of forecasts. VAT, a core revenue source, depends on the volume and composition of consumption, on prices and on tax compliance. The change in Tunisians’ consumption patterns in recent years may make it harder to estimate, and the reasons for that change deserve separate research. Supply disruptions and shortages of some goods may also change what is sold formally and what enters the tax base.
Figure 2. Tax revenue: gap with the initial finance law, by tax
Outturn − initial finance law, in billions of dinars. The number above each bar is the total gap. Hover over or tap a year for the breakdown.
The 2025 outturns may be provisional. Source: Ministry of Finance (finance laws and budget execution results); author's calculations.
A tax total close to the estimate can hide differences among its components: an increase in one source may offset a shortfall in another, and each affects households and firms differently. Non-tax revenue had a mean absolute gap of 16.7%. These receipts depend on varied factors, including public enterprise profits, transfers and exceptional revenues, so forecasting them better requires more transparency about their sources and the timing of their collection.
Expenditure
Spending excluding debt service had a mean absolute gap of 4.6%, but that average hides a path that says something about fiscal policy. Spending was below plan in 2014 and 2015, then exceeded it repeatedly between 2017 and 2022, by about 3.9 billion dinars in 2021 and 3.4 billion in 2022. Since 2023 spending has again come in under the estimates: by about 0.2 billion in 2023, 2 billion dinars in 2024 and 1 billion in 2025. This alone does not prove that the estimates became more conservative. It may also reflect pressure on execution, spending controls or postponed commitments. To judge, one has to look at the components of spending and at whether the cuts touched planned services or investments.
Wages had a mean gap of 2.5%, the lowest of any spending item. The state plans a large share of hiring, raises and promotions in advance, so the wage bill is clearer when the budget is prepared. Goods and services had a mean gap of 9.8%, and public interventions 18.5%. Interventions exceeded the planned amounts by 57% in 2021 and 41% in 2022. Their cost can change with prices and subsidy needs during the year, but the aggregate data do not show which programmes were behind the increase.
Figure 3. Spending excluding debt service: gap with the initial finance law, by category
Outturn − initial finance law, in billions of dinars. The number above each bar is the total gap. Hover over or tap a year for the breakdown.
The 2025 outturns may be provisional. Source: Ministry of Finance (finance laws and budget execution results); author's calculations.
Capital spending came in below its allocations in several years, including 2014 and 2015 and then the most recent years. Some of the projects written into the budget therefore did not turn into actual spending in the same year. The delay may be tied to project preparation, procedures or implementation capacity, and it may also reflect a change in priorities.
Financing
In financing, the mean absolute gap was 35.2% for the deficit, 34.2% for external borrowing and 62.9% for domestic borrowing. The deficit depends on the revenue collected, the spending the state carries out and the decisions it takes during the year. The actual deficit exceeded the estimate in seven years out of twelve, and the 2020 shock was among the main exceptional causes. The gap therefore shows how far the outcome moved from the plan, but it cannot by itself separate the contributions of forecasts, decisions and shocks.
Figure 4. Budget deficit: initial finance law forecast and outturn
% of GDP. The number above each year is the gap in points of GDP.
Average gap: +1.0 point of GDP (+0.5 excluding 2020); the deficit was above the forecast in 7 years out of 12. The 2025 outturns may be provisional. Source: Ministry of Finance (finance laws and budget execution results); author's calculations.
The change in borrowing sources is clearer. Since 2020 external borrowing has come in below plan, with gaps of −46% in 2020, −43% in 2021 and −38% in 2022, then −61% in 2023 and −79% in 2024. Domestic borrowing, by contrast, exceeded its estimates in most of these years: +364% in 2020, +43% in 2022 and +98% in 2024. The explanation is the state’s shift away from external borrowing toward growing reliance on domestic borrowing: it was forced to lean more on the local market, especially after negotiations with the IMF failed.
The public debt stock stayed relatively close to estimates, with a mean absolute gap of 4.2%, despite the change in borrowing sources. The year-end stock can land near the planned figure while the timing, cost and source of financing change. The composition of borrowing therefore matters when assessing the budget plan, alongside the size of the debt.
A change of pattern
Between 2014 and 2019 the gaps had no single clear direction: some revenues rose above the estimate one year and fell below it the next, and the same happened with the deficit and borrowing. Total spending was relatively close to plan, and below it in 2014 and 2015. The period looks more like year-to-year variation, except for the corporate tax deviations between 2015 and 2017, a reminder of how sensitive that revenue is to company profits and to the economic assumptions used in preparing the budget.
Figure 5. Gap between outturn and initial finance law, by item and year
% of the initial finance law. Green: revenue above the finance law, or spending, deficit, borrowing and debt below it. Red: the reverse.
Gap = (outturn − initial finance law) / initial finance law. The 2025 outturns may be provisional. Source: Ministry of Finance (finance laws and budget execution results); author's calculations.
The pattern changed from 2020. The pandemic was an exceptional shock, but some gaps persisted after it: between 2020 and 2023 the deficit exceeded its estimates repeatedly, as did public interventions, especially in 2021 and 2022. External borrowing stayed below plan from 2020 on, while reliance on domestic borrowing grew. Because the pattern has lasted, the question is wider than forecast accuracy. Did the finance laws reflect financing sources that could be counted on? Was spending reordered when those sources did not materialise? The data we have show the shift, but they are not enough on their own to weigh each cause.
Since 2023 spending has come in below plan, particularly wages and capital spending, and the actual deficit was lower than expected in 2024 and 2025. Debt stayed slightly below the estimate. Tax revenue, however, fell short of the forecast in 2023, 2024 and 2025, and external borrowing remained far below plan. These results point to lower estimates for some spending and for the deficit, while the same shift did not appear in revenue and the financing mix remained hard to predict. The 2025 figures may be provisional.
Conclusion
The comparison shows that tax totals, spending excluding debt service and the debt stock came relatively close to their estimates, while gaps widened in some taxes, public interventions, the deficit and borrowing sources. A total close to plan can also hide differences in the composition of revenue and spending. Assessing the budget therefore depends on how realistic the resources, allocations and financing plan are, and on how clearly differences are explained when they appear.
Figure 6. Yearly gaps between outturn and initial finance law, by item, 2014-2025
One dot per year, in % of the initial finance law. Items run from most to least accurate. Pick a year to highlight it on every row.
Gaps beyond ±60% are drawn at the edge of the chart; the tooltip gives the actual value. Gap = (outturn − initial finance law) / initial finance law. Source: Ministry of Finance (finance laws and budget execution results); author's calculations.
The practical question is whether expected revenue rests on assumptions that can be checked, whether the approved spending was carried out, and whether the change in financing is a passing circumstance or a recurring pattern. The total may come close to target while the mix of taxes and spending changes, projects are delayed, or domestic borrowing replaces external borrowing. These differences have different effects on households, services, investment and the cost of debt.
Narrowing the gap requires tying estimates to published assumptions about activity, consumption, profits and collection, providing an explanatory revision when conditions change, and tracking investment execution and financing sources during the year. On the policy side, subsidy reform, public enterprises, wage bill management and the fight against tax evasion all still affect how predictable and executable the budget is.
Read against recent IMF studies of forecast errors and budget credibility, the results for Tunisia look closer to a problem in executing some budget components and in the financing mix than to a general inaccuracy across all forecasts. This literature indicates that narrowing these gaps requires more realistic forecasts, stronger forecasting and monitoring capacity within the administration, better management of spending and liquidity, and a regularly published assessment of the difference between what was planned and what was achieved. Such an assessment would also help explain what is due to economic shocks and what is due to policy choices or implementation problems.