## 5. Tax Regimes and Probabilities of Exceeding Pre-Specified Fiscal Ratios

## Source details

**Canonical URL:** [5. Tax Regimes and Probabilities of Exceeding Pre-Specified Fiscal Ratios](https://www.imf.org/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2005/_wp05171.pdf)

## Other formats

- [Markdown version](/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2005/_wp05171.pdf.md)
- [Structured JSON version](/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2005/_wp05171.pdf.json)

---

### Introduction — motivation and central questions
- Paper motivated by two issues:
  - The extent to which high volatility of commodity prices affects fiscal performance by increasing the probability of under-performing on set targets.
  - The effects of the mode of sequencing of tax and expenditure policy decisions on fiscal performance in the context of transitory commodity price changes.
- Two broad tax regimes compared:
  - Pure tax regime: tax decisions are taken in advance and independent of expenditure decisions.
  - Expenditure-induced tax regime: expenditure decisions influence tax decisions.
- Sample: two oil-exporting countries (Kazakhstan and Russia) and two non-oil commodity exporters (Kyrgyz Republic and Tajikistan).
- Commodity price measures:
  - Oil price changes used for oil exporters.
  - Composite commodity price index (weighted average of oil and main export commodity prices using 2004 trade shares) for Kyrgyz Republic (Gold) and Tajikistan (Cotton and Aluminum).

### Key empirical findings and quantitative results
- General result: increases in commodity prices have significant effects on taxes, expenditures, and the fiscal balance for both country groups.
- Regime comparison:
  - Effects on taxes and expenditures are higher on average under the expenditure-induced tax regime than under the pure tax regime.
  - The expenditure-induced tax regime yields higher taxes than the pure tax regime and introduces more uncertainty into business decision-making.
  - Although the probability of exceeding tax floors is higher under the expenditure-induced tax regime, the likelihood of exceeding deficit ceilings tends to be higher under the expenditure-induced tax regime in the long run as positive commodity shocks dissipate—reflecting anticipated expenditure increases in expectation of transitory tax revenues.
- Impulse-response and probability analyses imply:
  - Expenditure-induced tax policy can yield higher deficits in the long run and increases output volatility.
  - Temporary booms in commodity prices can undermine longer-term fiscal performance if caution is not exercised in spending during boom periods.
- Country and fiscal statistics (preserved exactly as in source):
  - Russia: average fiscal surplus in 2001-04 of about 2.1 percent of GDP. Measured at a constant oil price of $20 per barrel, the general government fiscal surplus is projected to decline by 1 percent of GDP in 2005, to about 1.3 percent.
  - Kazakhstan: fiscal surplus of the general government averaged about 2 percent during 2001-04. Total revenues increased to 26 percent of GDP in 2004, from 22 percent in 2000, 7½ percentage points of which is derived from oil revenues.
  - Tajikistan: deficit including foreign financed PIP declined from 5.6 percent of GDP in 2000 to 2.7 percent in 2004. Public external debt reduced to 40 percent of GDP as of end-2004. On average, cotton and aluminum contributed 22 percent to growth in 1999-2004, and real GDP growth averaged 9 percent per year. The two commodities contributed an average 19 percent per year to tax revenues in the form of sales taxes.
  - Kyrgyz Republic: in 2003 the state-owned and a Canadian company (Centerra)–operated gold mine contributed around 10 percent of total taxes, and contributes some 40 percent to gross output. Real GDP growth was flat in 2002, yielding a moderate average rate of growth of 5.0 percent during 2000-04. Fiscal deficit of the general government declined by 6 percentage points of GDP to 4.2 percent of GDP during 2000-04. Non-gold exports increased by an average of 12 percent per year during 2000-04.

### Fiscal behavior and vulnerability to commodity price volatility
- Observed patterns:
  - International commodity prices (aluminum, cotton, and oil) exhibit strong volatility; gold prices relatively less volatile over the past two years but have seen large swings in the past.
  - Volatile commodity prices affect fiscal revenues and expenditure outlays in commodity-dependent countries; revenues and expenditures generally positively correlated with commodity prices, especially during price increases.
  - Pro-cyclical expenditures and entrenched spending programs during booms can lead to borrowing when revenues fall—raising the likelihood of costly later adjustments.
  - Some correlations between commodity prices and fiscal aggregates can become negative in downturns, indicating limited spending retrenchment.

### Data, model, and estimation details (SVAR approach)
- Variables used (logarithms of seasonally adjusted quarterly data):
  - ln(Pcom_t) — composite commodity price index (oil for Kazakhstan and Russia; weighted oil + main export commodity for Kyrgyz Republic and Tajikistan).
  - ln(Re_t / y_t) — revenue-to-GDP ratio.
  - ln(Exp_t / y_t) — expenditure-to-GDP ratio.
  - ln(y_t) — nominal GDP.
- Sample periods:
  - Kyrgyz Republic: 1995Q1–2004Q4.
  - Tajikistan: 1996Q1–2004Q4.
  - Kazakhstan: 1994Q1–2004Q2.
  - Russia: 1994Q2–2004Q2.
- Data sources: output, total revenues, and expenditures from IMF data sources and country authorities; oil, aluminum, gold, and cotton prices from Datastream.
- Time series properties:
  - ADF tests and ARIMA(p,1,0) specifications applied; logs of commodity price indices and output series generally follow unit root processes with possible drift; revenue/GDP and expenditure/GDP ratios are stationary in most cases (except Russia).
  - Johansen cointegration tests indicate 4 cointegrating vectors for each country dataset (null of no cointegration rejected at 5 percent level for both maximum eigen-value and trace criteria).
- Structural identification:
  - Structural VAR (SVAR) framework following Blanchard and Perotti-style identification to separate automatic stabilizers from discretionary fiscal shocks.
  - Assumptions include: commodity price shocks are strictly exogenous; government tax and spending decisions do not react contemporaneously to unexpected movements in real activity within the same quarter (cyclical components filtered out); parameters α, β, γ, χ, θ capture responses of revenues, expenditures, and output to commodity price shocks and discretionary fiscal shocks.
  - Initial reduced-form VAR lag length set to 6 quarters and reduced based on Schwartz and Akaike criteria; final structured system has 11 unknowns (7 parameters and 4 variances) estimated from reduced-form variance-covariance information.

### Fiscal regimes and identification — Pure Tax and Expenditure-Induced Tax
- Pure Tax Regime:
  - Characterized by weak exogeneity of government spending decisions on tax revenues (0
β=).
  - Two estimated versions:
    - Model I A (exactly identified): both tax revenues and government spending decisions have immediate effects on output (Model I A: 0
β=).
    - Model I B (over-identified): both decisions do not have immediate effects on output (Model I B: 0β= and 23
0θθ==).
  - Political economy interpretation: tax decisions are made prior to and independent of expenditure decisions.
- Expenditure-Induced Tax Regime:
  - Expenditure decisions are taken without contemporaneous feedback for tax performance (0
χ=), while tax decisions take into account expenditure shocks.
  - Two estimated versions:
    - Model III (exactly-identified): immediate effect of tax and expenditure decisions on output (Model III: 0
χ=).
    - Model IV (over-identified): absence of contemporaneous relationships (Model IV: 0χ= and 23
0θθ==).
  - Political economy interpretation: tax decisions are subordinated to spending programs.
- Model comparison: except for Tajikistan, likelihood ratio tests could not reject equality between two models within each fiscal regime at the 5 percent significance level; subsequent discussion focuses on Model I (pure tax) and Model III (expenditure-induced tax).

### Contemporaneous effects of commodity price shocks (key empirical findings)
- Revenues:
  - Commodity price shocks have a generally significant positive impact on revenue/GDP innovations (α).
  - Russia: a one percent increase in commodity prices results in an estimated 0.9 percent increase in revenue/GDP innovations.
  - Kyrgyz Republic: estimated responses around 0.5 under both tax regimes.
  - Tajikistan: revenue/GDP innovations jump from around 0.7 under the pure tax regime to 3.0 under the expenditure-induced tax regime.
  - Kazakhstan: positive but statistically insignificant responses of revenue/GDP innovations.
- Expenditures:
  - γ estimates indicate significant and positive contemporaneous effects of commodity price shocks on expenditures across tax regimes for all sample countries except under the pure tax regime in Kazakhstan.
  - Increase in expenditure/GDP innovations to a 1 percent commodity price shock is around 0.5 percent in Russia and the Kyrgyz Republic.
  - Tajikistan: response averaging 0.5 percent under the expenditure-induced tax regime and averaging 1.3 percent under the pure tax regime.
  - Kazakhstan: response insignificant and lowest.
- Tax–expenditure interactions:
  - χ indicates a positive response of expenditures to discretionary tax increases for all countries.
- Output responses (within-quarter):
  - Oil-exporting countries and the Kyrgyz Republic: positive commodity price shocks tend to increase output.
  - Tajikistan: positive commodity price shocks reduce output significantly (explained by inclusion of oil in composite commodity price index).
  - Discretionary revenue shocks show Keynesian-type effects: tax increases reduce output; the tax multiplier averages –0.3 in Russia.
  - Output response to expenditure shocks is mixed.

### Dynamic effects (impulse responses and group averages)
- Simulation scenario: 1 percent transitory shock to the commodity price index traced over 12 quarters.
- Oil-exporting countries (Kazakhstan and Russia):
  - Following a 1 percent shock to the oil price, revenues increase leading to budget surpluses.
  - Average response: revenues at the 5-percent significance level higher than expenditures over forecast horizon, producing higher fiscal surpluses particularly short-term.
- Non-oil-exporting commodity countries (Kyrgyz Republic and Tajikistan):
  - Composite commodity price increases lead to increases in revenues and expenditures but also increase the budget deficit (reflects incremental cost of oil imports).
  - Average response: expenditures higher than revenues, implying worsening fiscal performance after the shock.
- Regime comparison:
  - Overall change in fiscal balance is more pronounced under the expenditure-induced tax regime than under the pure tax regime.
  - Output volatility: lower when tax policy leads and is independent of expenditures (pure tax regime) than when expenditure decisions are taken in anticipation of tax receipts.
- Timing and persistence:
  - Output rises during the first year after a positive commodity price shock, reaches peak around four quarters, then reverts; long-run effects on output disappear.

### Stochastic simulations and probabilities (method and illustrative application)
- Simulation approach:
  - Structural VAR is simulated using a bootstrapping approach with 2,000 replications of the estimated SVAR.
  - From replications, median impulse-responses are derived and cumulative distributions constructed for each tax regime over forecast horizon k.
  - Horizons illustrated: k = 0, k = 4, k = 8, k = 12 (quarters after the shock).
  - Illustrative baseline/target revenue and expenditure ratios set at 5 percent.
- Interpretation example:
  - Revenue panel: probability of exceeding the 5 percent revenue baseline at k = 0 is about 80 percent.
  - Expenditure panel: probability of exceeding the 5 percent expenditure baseline at k = 0 is about 90 percent.
  - The 5-percent significance level intersections indicate the likely levels of fiscal aggregates at that significance level for each horizon.
- Uses of simulations:
  - Derive 1
α−significance levels of fiscal aggregates and probabilities of exceeding specified revenue, expenditure, and deficit/surplus baselines following a 1 percent commodity price shock.
  - Inform the size of fiscal buffers required to protect against unexpected shocks driving fiscal aggregates outside specified bands.

### Odds on fiscal performance and illustrative quantitative results
- General cross-country pattern:
  - Oil-exporting countries: simulations point to increases in fiscal surpluses (improved fiscal performance) following a positive commodity price shock.
  - Non-oil-exporting commodity-intensive countries: simulations point to higher expenditures than revenues, increasing deficits (worsening fiscal performance).
- Tajikistan (5-percent significance level, following a positive unanticipated commodity price shock):
  - Average annual fiscal deficit one year after the price shock: about 4.3 percent of GDP under the pure-tax regime.
  - Two years after the shock: approximately 4.7 percent of GDP.
  - Three years after the shock: approximately 4.9 percent of GDP.
  - Interpretation: gradual decline in fiscal performance reflects temporary nature of the positive commodity price shock and absence of fiscal tightening in anticipation of reversion.
- Kyrgyz Republic:
  - Simulations indicate a likely increase in deficit level (specific numeric progression for Kyrgyz Republic not fully reported in the supplied excerpt).

### Simulated fiscal aggregate levels (5 percent significance level)
- For oil-exporting countries, the oil price shock yields revenue gains that could strengthen fiscal surpluses:
  - Kazakhstan: likely fiscal surplus level of around 1 percent of GDP over the forecast horizon.
  - Russia: likely fiscal surplus level of around 4.5 percent of GDP over the forecast horizon.
- Simulations indicate a short-term increase in deficits that gradually rises as one-off effects dissipate; specific cited path: 4.7 percent of GDP in the first year to around 5.5 percent at the end of the third year.
- The likely levels of fiscal balances at the 5 percent significance level are generally higher under the expenditure-induced tax regime than under the pure tax regime.

### Probabilities of exceeding fiscal floors and ceilings (conditional on a 1 percent positive shock to the commodity price index)
- Simulations estimate probabilities of:
  - Exceeding fiscal floors and ceilings for Kyrgyz Republic and Tajikistan.
  - Performing better than recent trend (exceeding surplus) or worse than recent trend for Kazakhstan and Russia.
- Key comparative findings between regimes:
  - A priori expectation: expenditure-induced tax regime increases probability of over-performing revenue floors relative to the pure tax regime because of additional revenue from autonomous expenditure increases. Simulations confirm this for revenues but not uniformly for surpluses/deficits.
  - Expenditure-induced tax regime is generally associated with higher expenditures than the pure tax regime, affecting net probabilities of exceeding surplus/floor/ceiling targets.
- Temporal pattern:
  - Fiscal floors and ceilings are easily exceeded during the first year of the positive commodity price shock; performance deteriorates shortly after the transitory shock.
  - After the transitory shock dissipates, probabilities of exceeding targets decline significantly.
- Country-specific probability results:
  - Tajikistan:
    - Immediately after the price shock, probability of over-performing on revenues is high for a short period, leading to corresponding increases in expenditures.
    - Because spending is not immediately adjusted after the temporary price increase, likelihood of meeting the fiscal deficit target deteriorates initially; after growth effects and eventual adjustment of spending it becomes more likely to meet fiscal targets.
    - The expenditure-induced tax regime yields higher probabilities in the second and third years following the commodity price shock.
  - Kyrgyz Republic:
    - Odds against exceeding the deficit ceiling rise sharply to between 0.9 and 1 in the first quarter after the price shock.
    - These odds decline to between zero and 0.1 in subsequent quarters.
    - The expenditure-induced tax regime yields higher probabilities of exceeding the ceiling on the fiscal deficit, and also yields higher probabilities of exceeding the tax floors.
  - Kazakhstan:
    - Probabilities of exceeding the revenue target and the expenditure target decline, in particular in the second to fourth quarters after the transitory commodity price shock.
    - The probability of exceeding the floor on the fiscal surplus remains flat under both tax regimes in the short run.
    - In the long run, probability of exceeding the pre-specified expenditure ceiling rises faster than that of exceeding the revenue floor, leading to a decline in the probability of exceeding the fiscal surplus—particularly under the expenditure-induced tax regime.
  - Russia:
    - The expenditure-induced tax regime yields higher probabilities of exceeding the surplus.
    - Simulated odds on fiscal performance do not vary much across tax regimes; this invariance could be due to lack of statistical evidence to reject equality between the two model variants.

### Methodological notes and caveats
- Bootstrapping details:
  - Procedure draws error terms from estimated residuals, generates VAR variables using estimated coefficients, and re-runs VAR 2,000 times.
  - Estimated impulse-responses reported are median responses from 2,000 replications.
  - Confidence bands from this approach indicate uncertainty around parameter estimates but should not be interpreted as strict confidence intervals due to time-dependence assumptions.
- Limitations and further research:
  - Seeming invariance of revenue/GDP responses to commodity price shocks across tax regimes could reflect improper specification of data generation process or specific tax treatment of oil extraction and petroleum trade.
  - Inclusion of oil in composite commodity price indices may drive negative output responses for non-oil-exporting countries; excluding oil from the index yields output increases for Tajikistan and the Kyrgyz Republic.
  - Authors underscore need for further research on channels through which symmetric and non-symmetric commodity price shocks affect output.

### Policy implications and recommendations (concluding remarks)
- Sequencing of expenditure and tax policies is crucial for fiscal performance in an uncertain commodity price environment.
- For PRGF sample countries (non-oil-exporting, commodity-dependent):
  - Higher probabilities of meeting fiscal targets when following a conservative tax regime (a pure tax regime) rather than allowing revenue targets to be determined by expenditure commitments.
- For oil-producing non-PRGF countries:
  - Expenditure restraint is beneficial when rising oil prices are transitory because the probability of over-performing recent trends in fiscal surpluses declines once revenue-enhancing effects dissipate.
- Additional policy recommendations:
  - Increase diversification of the economy (particularly for non-oil exporting commodity-intensive small economies) to enhance resilience to commodity price shocks.
  - Advance the private sector development agenda to widen the tax base and enhance tax administration to reduce vulnerability of public finances to commodity price volatility.
  - Build flexibility into medium-term budget frameworks for commodity-dependent economies, because projected revenues may not be realized due to exogenous shocks and strict commitment could delay fiscal adjustment.
  - Recognize that increases in spending are difficult to reverse once entrenched; significant reductions in tax revenues could result in higher fiscal deficits (or payment arrears) that may compromise macroeconomic stability.

*Source: IMF Working Paper — Chapter/Section “5. Tax Regimes and Probabilities of Exceeding Pre-Specified Fiscal Ratios” (content as provided).*

### 1.     Contemporaneous Effects of Shocks .......................................................15

### 1.     Contemporaneous Effects of Shocks .......................................................15

### Major sections and structure
- Section titles and page references:
  - "1.     Contemporaneous Effects of Shocks .......................................................15"
  - "2.     Dynamic Effects of Commodity Price Shocks ........................................17"
  - "B. The Odds on Fiscal Performance ............................................................................20"
  - "1. 
1
α−significance Levels of Fiscal Aggregates.........................................22"
  - "2. Probability of Exceeding Fiscal Floors and Ceilings...............................25"
  - "V. Concluding Remarks...........................................................................................................28"
  - "References............................................................................................................................... 29"
  - "Appendix"
  - "Simulated 
1
α−significance Levels of Fiscal Aggregates and Probabilities  
   of Exceeding Pre-Specified Fiscal Ratios.............................................................................31"

### Empirical components (tables and figures listed)
- Tables:
  - "1. Dickey-Fuller Tests for Stationarity ....................................................................................10"
  - "2. Structural Parameter Estimates ............................................................................................17"
- Figures:
  - "1. Correlations between Commodity Prices and Fiscal Aggregates ..........................................6"
  - "2. Tax Regimes and Responses to Commodity Price Shocks..................................................19"
  - "3. Illustrative Probability Distribution of Fiscal Aggregates ...................................................22"
  - "4. Simulated 5-percent Significant Levels of Fiscal Aggregates .............................................24"

### Themes indicated by section headings
- Contemporaneous effects of shocks:
  - Focus on immediate (contemporaneous) relationships between shocks and fiscal variables (title indicates coverage on page 15).
- Dynamic effects of commodity price shocks:
  - Coverage of how commodity price shocks evolve and affect variables over time (title indicates coverage on page 17).
- Odds on fiscal performance:
  - Analysis framed around probabilities and odds (section "B. The Odds on Fiscal Performance" on page 20).
- Significance levels and probability thresholds:
  - Discussion of "1
α−significance Levels of Fiscal Aggregates" (page 22) and "Probability of Exceeding Fiscal Floors and Ceilings" (page 25).
- Appendix simulations:
  - Appendix contains simulated "1
α−significance Levels of Fiscal Aggregates and Probabilities of Exceeding Pre-Specified Fiscal Ratios" (page 31).

### Key numeric items preserved from source listing
- Page numbers: 6, 10, 15, 17, 19, 20, 22, 24, 25, 28, 29, 31.
- Table numbers: 1, 2.
- Figure numbers: 1, 2, 3, 4.
- Numeric expression: "5-percent" (in "Simulated 5-percent Significant Levels of Fiscal Aggregates").
- Symbolic expression preserved exactly: "1
α−significance Levels of Fiscal Aggregates".

*Source: _wp05171 - 1.     Contemporaneous Effects of Shocks .......................................................15*

### 5. Tax Regimes and Probabilities of Exceeding Pre-Specified Fiscal Ratios..........................26

### 5. Tax Regimes and Probabilities of Exceeding Pre-Specified Fiscal Ratios

### Introduction — motivation and central questions
- Paper motivated by two issues:
  - The extent to which high volatility of commodity prices affects fiscal performance by increasing the probability of under-performing on set targets.
  - The effects of the mode of sequencing of tax and expenditure policy decisions on fiscal performance in the context of transitory commodity price changes.
- Two broad tax regimes compared:
  - Pure tax regime: tax decisions are taken in advance and independent of expenditure decisions.
  - Expenditure-induced tax regime: expenditure decisions influence tax decisions.
- Sample: two oil-exporting countries (Kazakhstan and Russia) and two non-oil commodity exporters (Kyrgyz Republic and Tajikistan).
- Commodity price measures:
  - Oil price changes used for oil exporters.
  - Composite commodity price index (weighted average of oil and main export commodity prices using 2004 trade shares) for Kyrgyz Republic (Gold) and Tajikistan (Cotton and Aluminum).

### Key empirical findings and quantitative results
- General result: increases in commodity prices have significant effects on taxes, expenditures, and the fiscal balance for both country groups.
- Regime comparison:
  - Effects on taxes and expenditures are higher on average under the expenditure-induced tax regime than under the pure tax regime.
  - The expenditure-induced tax regime yields higher taxes than the pure tax regime and introduces more uncertainty into business decision-making.
  - Although the probability of exceeding tax floors is higher under the expenditure-induced tax regime, the likelihood of exceeding deficit ceilings tends to be higher under the expenditure-induced tax regime in the long run as positive commodity shocks dissipate—reflecting anticipated expenditure increases in expectation of transitory tax revenues.
- Implications from impulse-response and probability analyses:
  - Expenditure-induced tax policy can yield higher deficits in the long run and increases output volatility.
  - Temporary booms in commodity prices can undermine longer-term fiscal performance if caution is not exercised in spending during boom periods.

- Country and fiscal statistics (preserved exactly as in source):
  - Russia: average fiscal surplus in 2001-04 of about 2.1 percent of GDP. Measured at a constant oil price of $20 per barrel, the general government fiscal surplus is projected to decline by 1 percent of GDP in 2005, to about 1.3 percent.
  - Policy changes under consideration in Russia: cutting the unified social tax (UST); replacing many existing in-kind benefits with monetary benefits estimated to cost about 2 percent of GDP; debate on reducing the VAT tax rate from 18 to 13 percent in 2005.
  - Kazakhstan: fiscal surplus of the general government averaged about 2 percent during 2001-04. Total revenues increased to 26 percent of GDP in 2004, from 22 percent in 2000, 7½ percentage points of which is derived from oil revenues.
  - Tajikistan: deficit including foreign financed PIP declined from 5.6 percent of GDP in 2000 to 2.7 percent in 2004. Public external debt reduced to 40 percent of GDP as of end-2004. On average, cotton and aluminum contributed 22 percent to growth in 1999-2004, and real GDP growth averaged 9 percent per year. The two commodities contributed an average 19 percent per year to tax revenues in the form of sales taxes.
  - Kyrgyz Republic: in 2003 the state-owned and a Canadian company (Centerra)–operated gold mine contributed around 10 percent of total taxes, and contributes some 40 percent to gross output. Real GDP growth was flat in 2002, yielding a moderate average rate of growth of 5.0 percent during 2000-04. Fiscal deficit of the general government declined by 6 percentage points of GDP to 4.2 percent of GDP during 2000-04. Non-gold exports increased by an average of 12 percent per year during 2000-04.

### Fiscal behavior and vulnerability to commodity price volatility
- Observed patterns:
  - International commodity prices (aluminum, cotton, and oil) exhibit strong volatility; gold prices relatively less volatile over the past two years but have seen large swings in the past.
  - Volatile commodity prices affect fiscal revenues and expenditure outlays in commodity-dependent countries; revenues and expenditures generally positively correlated with commodity prices, especially during price increases.
  - Pro-cyclical expenditures and entrenched spending programs during booms can lead to borrowing when revenues fall—raising the likelihood of costly later adjustments.
  - Some correlations between commodity prices and fiscal aggregates can become negative in downturns, indicating limited spending retrenchment.

### Data, model, and estimation details (SVAR approach)
- Variables used (logarithms of seasonally adjusted quarterly data):
  - ln(Pcom_t) — composite commodity price index (oil for Kazakhstan and Russia; weighted oil + main export commodity for Kyrgyz Republic and Tajikistan).
  - ln(Re_t / y_t) — revenue-to-GDP ratio.
  - ln(Exp_t / y_t) — expenditure-to-GDP ratio.
  - ln(y_t) — nominal GDP.
- Sample periods:
  - Kyrgyz Republic: 1995Q1–2004Q4.
  - Tajikistan: 1996Q1–2004Q4.
  - Kazakhstan: 1994Q1–2004Q2.
  - Russia: 1994Q2–2004Q2.
- Data sources: output, total revenues, and expenditures from IMF data sources and country authorities; oil, aluminum, gold, and cotton prices from Datastream.
- Time series properties:
  - ADF tests and ARIMA(p,1,0) specifications applied; logs of commodity price indices and output series generally follow unit root processes with possible drift; revenue/GDP and expenditure/GDP ratios are stationary in most cases (except Russia).
  - Johansen cointegration tests indicate 4 cointegrating vectors for each country dataset (null of no cointegration rejected at 5 percent level for both maximum eigen-value and trace criteria).
- Structural identification:
  - Structural VAR (SVAR) framework following Blanchard and Perotti-style identification to separate automatic stabilizers from discretionary fiscal shocks.
  - Assumptions include: commodity price shocks are strictly exogenous; government tax and spending decisions do not react contemporaneously to unexpected movements in real activity within the same quarter (cyclical components filtered out); parameters α, β, γ, χ, θ capture responses of revenues, expenditures, and output to commodity price shocks and discretionary fiscal shocks.
  - Initial reduced-form VAR lag length set to 6 quarters and reduced based on Schwartz and Akaike criteria; final structured system has 11 unknowns (7 parameters and 4 variances) estimated from reduced-form variance-covariance information.

### Policy implications and recommendations (from empirical results)
- Main messages:
  - Uncertainty and volatility of commodity prices complicate fiscal policy management; the challenge is to absorb commodity price volatility without adopting measures that could undermine long-term growth.
  - Temporary commodity price booms can undermine longer-term fiscal performance if spending is not cautious during booms.
  - Expenditure-induced tax policy may require additional financing, which might adversely affect macroeconomic stability, since it yields higher long-run deficits and increases output volatility.
  - Indicates the need for caution in projecting expenditure paths—especially when tax revenues are volatile—and in expanding social sector outlays based on transitory revenue increases.

*Source: IMF Working Paper — Chapter/Section “5. Tax Regimes and Probabilities of Exceeding Pre-Specified Fiscal Ratios” (content as provided).*

### 1. Pure Tax Regime. This regime is characterized by weak exogeneity of government

### _wp05171 - 1. Pure Tax Regime. This regime is characterized by weak exogeneity of government

### Fiscal regimes and identification
- Pure Tax Regime: characterized by weak exogeneity of government spending decisions on tax revenues (0
β=). Two estimated versions:
  - Model I A (exactly identified): both tax revenues and government spending decisions have immediate effects on output (Model I A: 0
β=).
  - Model I B (over-identified): both decisions do not have immediate effects on output (Model I B: 0β= and 23
0θθ==).
- Expenditure-Induced Tax Regime: expenditure decisions are taken without contemporaneous feedback for tax performance (0
χ=), while tax decisions take into account expenditure shocks. Two estimated versions:
  - Model III (exactly-identified): immediate effect of tax and expenditure decisions on output (Model III: 0
χ=).
  - Model IV (over-identified): absence of contemporaneous relationships (Model IV: 0χ= and 23
0θθ==).
- Political economy interpretation:
  - Pure tax regime: tax decisions are made prior to and independent of expenditure decisions.
  - Expenditure-induced tax regime: tax decisions are subordinated to spending programs.

### Contemporaneous effects of commodity price shocks (key empirical findings)
- General result: commodity price shocks have a generally significant positive impact on revenue/GDP innovations (α).
- Country-specific highlights:
  - Russia: a one percent increase in commodity prices results in an estimated 0.9 percent increase in revenue/GDP innovations.
  - Kyrgyz Republic: estimated responses around 0.5 under both tax regimes.
  - Tajikistan: revenue/GDP innovations jump from around 0.7 under the pure tax regime to 3.0 under the expenditure-induced tax regime.
  - Kazakhstan: positive but statistically insignificant responses of revenue/GDP innovations; estimates have correct sign but not statistically significant.
- Expenditure/GDP innovations:
  - γ estimates indicate significant and positive contemporaneous effects of commodity price shocks on expenditures across tax regimes for all sample countries except under the pure tax regime in Kazakhstan.
  - Increase in expenditure/GDP innovations to a 1 percent commodity price shock is around 0.5 percent in Russia and the Kyrgyz Republic.
  - Tajikistan: response averaging 0.5 percent under the expenditure-induced tax regime and averaging 1.3 percent under the pure tax regime.
  - Kazakhstan: response insignificant and lowest; explanations include recent start of oil exploitation and prudent expenditure policy through end 2003.
- Tax–expenditure interactions:
  - For all countries, the estimate of χ indicates a positive response of expenditures to discretionary tax increases.
- Output responses (within-quarter):
  - Oil-exporting countries and the Kyrgyz Republic: positive commodity price shocks tend to increase output.
  - Tajikistan: positive commodity price shocks reduce output significantly (explained by inclusion of oil in composite commodity price index).
  - Discretionary revenue shocks show Keynesian-type effects: tax increases reduce output; the tax multiplier averages –0.3 in Russia.
  - Output response to expenditure shocks is mixed, possibly reflecting inadequate variation in discretionary expenditures after excluding the autonomous expenditure component.
- Model comparison:
  - Except for Tajikistan, the likelihood ratio test for equality between two models within each fiscal regime could not reject the null at the 5 percent significance level, suggesting a single model can represent each fiscal regime. Consequently, subsequent discussion focuses on Model I (pure tax) and Model III (expenditure-induced tax).

### Dynamic effects (impulse responses and group averages)
- Method: simulations consider a 1 percent transitory shock to the commodity price index and trace impacts on revenue/GDP, expenditure/GDP, and fiscal balance/GDP over 12 quarters.
- Oil-exporting countries (Kazakhstan and Russia):
  - Following a 1 percent shock to the oil price, revenues increase leading to budget surpluses.
  - Average response: revenues at the 5-percent significance level higher than expenditures over forecast horizon, producing higher fiscal surpluses particularly short-term.
- Non-oil-exporting commodity countries (Kyrgyz Republic and Tajikistan):
  - Composite commodity price increases lead to increases in revenues and expenditures but also increase the budget deficit (reflects incremental cost of oil imports).
  - Average response: expenditures higher than revenues, implying worsening fiscal performance after the shock.
- Regime comparison:
  - Overall change in fiscal balance is more pronounced under the expenditure-induced tax regime than under the pure tax regime.
  - Output volatility: lower when tax policy leads and is independent of expenditures (pure tax regime) than when expenditure decisions are taken in anticipation of tax receipts (expenditure-induced tax regime).
- Timing and persistence:
  - Country-specific impulse responses show output rises during the first year after a positive commodity price shock, reaches peak around four quarters, then reverts; long-run effects on output disappear.

### Stochastic simulations and probabilities (method and illustrative application)
- Simulation approach:
  - Structural VAR is simulated using a bootstrapping approach with 2,000 replications of the estimated SVAR.
  - From replications, median impulse-responses are derived and cumulative distributions constructed for each tax regime over forecast horizon k.
  - Horizons illustrated: k = 0, k = 4, k = 8, k = 12 (quarters after the shock).
  - Illustrative baseline/target revenue and expenditure ratios set at 5 percent.
- Interpretation example from Figure 3:
  - Revenue panel: probability of exceeding the 5 percent revenue baseline at k = 0 is about 80 percent.
  - Expenditure panel: probability of exceeding the 5 percent expenditure baseline at k = 0 is about 90 percent.
  - The 5-percent significance level intersections indicate the likely levels of fiscal aggregates at that significance level for each horizon.
- Uses of simulations:
  - Derive 1
α−significance levels of fiscal aggregates and probabilities of exceeding specified revenue, expenditure, and deficit/surplus baselines following a 1 percent commodity price shock.
  - Inform the size of fiscal buffers required to protect against unexpected shocks driving fiscal aggregates outside specified bands.

### Odds on fiscal performance and illustrative quantitative results
- General cross-country pattern:
  - Oil-exporting countries: simulations point to increases in fiscal surpluses (improved fiscal performance) following a positive commodity price shock.
  - Non-oil-exporting commodity-intensive countries: simulations point to higher expenditures than revenues, increasing deficits (worsening fiscal performance).
- Tajikistan (5-percent significance level, following a positive unanticipated commodity price shock):
  - Average annual fiscal deficit one year after the price shock: about 4.3 percent of GDP under the pure-tax regime.
  - Two years after the shock: approximately 4.7 percent of GDP.
  - Three years after the shock: approximately 4.9 percent of GDP.
  - Interpretation: gradual decline in fiscal performance reflects temporary nature of the positive commodity price shock and absence of fiscal tightening in anticipation of reversion.
- Kyrgyz Republic: similar simulations indicate a likely increase in deficit level (specific numeric progression for Kyrgyz Republic not fully reported in the supplied excerpt).

### Methodological notes and caveats
- Bootstrapping details:
  - Procedure draws error terms from estimated residuals, generates VAR variables using estimated coefficients, and re-runs VAR 2,000 times.
  - Estimated impulse-responses reported are median responses from 2,000 replications.
  - Confidence bands from this approach indicate uncertainty around parameter estimates but should not be interpreted as strict confidence intervals due to time-dependence assumptions.
- Limitations and further research:
  - Seeming invariance of revenue/GDP responses to commodity price shocks across tax regimes could reflect improper specification of data generation process or specific tax treatment of oil extraction and petroleum trade.
  - Inclusion of oil in composite commodity price indices may drive negative output responses for non-oil-exporting countries; excluding oil from the index yields output increases for Tajikistan and the Kyrgyz Republic.
  - Authors underscore need for further research on channels through which symmetric and non-symmetric commodity price shocks affect output.

*Source: Authors' simulations and empirical results reported in the supplied content of _wp05171 - 1. Pure Tax Regime. This regime is characterized by weak exogeneity of government spending decisions on tax revenues (0β=).*

### 4.7 percent of GDP in the first year to around 5.5 percent at the end of the third year.

### _wp05171 - 4.7 percent of GDP in the first year to around 5.5 percent at the end of the third year.

### Simulated fiscal aggregate levels (5 percent significance level)
- Simulations use estimated impulse-response functions and report likely fiscal aggregate levels at the 5 percent significance level.
- For oil-exporting countries, the oil price shock yields revenue gains that could strengthen fiscal surpluses:
  - Kazakhstan: likely fiscal surplus level of around 1 percent of GDP over the forecast horizon.
  - Russia: likely fiscal surplus level of around 4.5 percent of GDP over the forecast horizon.
- Simulations indicate a short-term increase in deficits that gradually rises as one-off effects dissipate; specific cited path: 4.7 percent of GDP in the first year to around 5.5 percent at the end of the third year.
- The likely levels of fiscal balances at the 5 percent significance level are generally higher under the expenditure-induced tax regime than under the pure tax regime.

### Probabilities of exceeding fiscal floors and ceilings (conditional on a 1 percent positive shock to the commodity price index)
- Simulations estimate probabilities of:
  - Exceeding fiscal floors and ceilings for Kyrgyz Republic and Tajikistan.
  - Performing better than recent trend (exceeding surplus) or worse than recent trend for Kazakhstan and Russia.
- Key comparative findings between regimes:
  - A priori expectation: expenditure-induced tax regime increases probability of over-performing revenue floors relative to the pure tax regime because of additional revenue from autonomous expenditure increases. Simulations confirm this for revenues but not uniformly for surpluses/deficits.
  - Expenditure-induced tax regime is generally associated with higher expenditures than the pure tax regime, affecting net probabilities of exceeding surplus/floor/ceiling targets.
- Temporal pattern:
  - Fiscal floors and ceilings are easily exceeded during the first year of the positive commodity price shock; performance deteriorates shortly after the transitory shock.
  - After the transitory shock dissipates, probabilities of exceeding targets decline significantly.
- Country-specific probability results (preserve original phrasing and numeric ranges):
  - Tajikistan:
    - Immediately after the price shock, probability of over-performing on revenues is high for a short period, leading to corresponding increases in expenditures.
    - Because spending is not immediately adjusted after the temporary price increase, likelihood of meeting the fiscal deficit target deteriorates initially; after growth effects and eventual adjustment of spending it becomes more likely to meet fiscal targets.
    - The expenditure-induced tax regime yields higher probabilities in the second and third years following the commodity price shock.
  - Kyrgyz Republic:
    - Odds against exceeding the deficit ceiling rise sharply to between 0.9 and 1 in the first quarter after the price shock.
    - These odds decline to between zero and 0.1 in subsequent quarters.
    - The expenditure-induced tax regime yields higher probabilities of exceeding the ceiling on the fiscal deficit, and also yields higher probabilities of exceeding the tax floors.
  - Kazakhstan:
    - Probabilities of exceeding the revenue target and the expenditure target decline, in particular in the second to fourth quarters after the transitory commodity price shock.
    - The probability of exceeding the floor on the fiscal surplus remains flat under both tax regimes in the short run.
    - In the long run, probability of exceeding the pre-specified expenditure ceiling rises faster than that of exceeding the revenue floor, leading to a decline in the probability of exceeding the fiscal surplus—particularly under the expenditure-induced tax regime.
  - Russia:
    - The expenditure-induced tax regime yields higher probabilities of exceeding the surplus.
    - Simulated odds on fiscal performance do not vary much across tax regimes; this invariance could be due to lack of statistical evidence to reject equality between the two model variants.

### Policy analysis and recommendations (concluding remarks)
- Sequencing of expenditure and tax policies is crucial for fiscal performance in an uncertain commodity price environment.
- For PRGF sample countries (non-oil-exporting, commodity-dependent):
  - Higher probabilities of meeting fiscal targets when following a conservative tax regime (a pure tax regime) rather than allowing revenue targets to be determined by expenditure commitments.
- For oil-producing non-PRGF countries:
  - Expenditure restraint is beneficial when rising oil prices are transitory because the probability of over-performing recent trends in fiscal surpluses declines once revenue-enhancing effects dissipate.
- Additional policy recommendations:
  - Increase diversification of the economy (particularly for non-oil exporting commodity-intensive small economies) to enhance resilience to commodity price shocks.
  - Advance the private sector development agenda to widen the tax base and enhance tax administration to reduce vulnerability of public finances to commodity price volatility.
  - Build flexibility into medium-term budget frameworks for commodity-dependent economies, because projected revenues may not be realized due to exogenous shocks and strict commitment could delay fiscal adjustment.
  - Recognize that increases in spending are difficult to reverse once entrenched; significant reductions in tax revenues could result in higher fiscal deficits (or payment arrears) that may compromise macroeconomic stability.

*Source: Authors' simulations and bootstrapping simulations based on estimated Structural VAR and estimated impulse-response functions (content as provided).*

---


_Source: https://www.imf.org/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2005/_wp05171.pdf_