Harnessing Big Data Analytics to Enhance Cost Allocation and Budget Preparation: Evidence from Iraq

This work is licensed under a Creative Commons Attribution 4.0 International License. The license permits unrestricted use, distribution, and reproduction in any medium, on the condition that users give exact credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if they made any changes. Harnessing Big Data Analytics to Enhance Cost Allocation and Budget Preparation: Evidence from Iraq Noora Sabah Baha Aldin Alhalawi  Business Administration, College of Administration & Economics, Kirkuk University, Iraq Abstract This paper discusses how big data analytics can be used to enhance the cost allocation and budget preparation in the Iraqi system of public financial management. The descriptive statistics, regression modeling, and time-series approaches are used to analyze fiscal data in 2015-2024 to evaluate the performance of the budgeting process in Iraq and the usefulness of the advanced analytics in the specified setting. Findings indicate long-term problems: the average oil dependency at 88-93 and a budget variance of -14%. Accuracy of forecasting is also still behind international standards. Even though our regression models are limited with respect to numbers of observation, they still show that the variables of oil prices and political stability affect accuracy but the results are not statistically significant. The study was applied with an AR(1) specification, and the resulting MAPE of 13.3 was a more optimistic result than the actual results of Iraq however below the 812% range that represents OECD countries. Experiments conducted in the past indicate that big data analytics have a potential of enhancing accuracy in forecasts by 10-25, but such results vary depending on the situation and the readiness of the infrastructural systems. It is estimated that the implementation cost will be 415 million USD within 5 years, and its potential annual benefit of up to 5.3 billion USDa hypothetical ROI of more than 6000% will be achieved provided that key initiatives within the organization and technological barriers are overcome. The paper tests three hypotheses, namely (1) big data analytics has the potential to ameliorate the quality of forecasts where sufficient infrastructure has been established; (2) conventional line-item budgeting is not suitable to oil-dependent, volatile economies; and (3) institutional and technological impediments are a significant barrier to the implementation of analytics in Iraq. The findings suggest that the study recommends a progressive implementation process aimed at enhancing data quality, conducting analytics tools pilot projects within some of the ministries, developing technical capacity, and postponing massive investments until some firm baseline is achieved. Keywords: Big Data Analytics, Revenue forecasting, Budget variance, cost allocation, Iraq, Public financial management, oil dependent economies, ARIMA, Fiscal volatility. JEL Classification codes: H61, H68, O23, C53. Suggested citation: Alhalawi, N.S.B.A. (2025). Harnessing Big Data Analytics to Enhance Cost Allocation and Budget Preparation: Evidence from Iraq. European Journal of Management, Economics and Business, 2(6), 258-272. DOI: 10.59324/ejmeb.2025.2(6).19 EJMEB (ISSN 3041-2102) | VOLUME 2 | NUMBER 6 | 2025 259 Introduction The horizon of the economy management of the state is shifting quickly, and the governments of the world are experimenting with new sophisticated analytical tools to make the budgeting process more efficient and open. The capability of big data analytics, or, in other words, the ability to processing large and heterogeneous sets of data in real time, now allows more fiscal planning options than were imaginable only a generation ago. However, the application of such technologies in the resource-dependent developing nations with low institutions and unstable fiscal conditions remains an unexplored field. Iraq is a prime example. Iraq remains under a perpetual fiscal instability even though the country makes a profit of more than 100 billion USD in oil revenues on a yearly basis. With oil comprising 88-93% of state income, the financial system is characterized by a high level of reliance on hydrocarbons, and the country is very susceptible to the changes in the global prices (International Monetary Fund, 2024; Shafaq News, 2025). Meanwhile, Iraq continues to use the conventional line item budgeting which is ill adapted to the volatility the commodity exporters regularly encounter. All these difficulties have become apparent over the last few years. Toxic politics and frequent crises resulted in three years of non-approved budgets, including 2014, 2020 and 2022, which has interfered with the delivery of services to the public, as well as slowed development projects (Al Jazeera, 2023; Atlantic Council, 2024). Actual revenues continue to be lower than expected even with approved budgets and it is in the negative by an average of -14 in the last ten yearsnearly twice as much as seen in the Gulf Cooperation Council countries. The IMF (2024) has been clear about the poor fiscal data infrastructure of Iraq, which has loopholes in granularity, frequency, and timeliness that negate quality control and analysis. Low-quality data, low analytical capacity, and disintegrated information systems amongst ministries are significant barriers to using advanced analytics as a combination. But these very evils testify to the imperatio modernization. Although there is enough research on big data in developed sectors, little research has been conducted on the viability and utility of big data in developing economies, which are resource-constrained. This paper fills that gap by analyzing the empirical performance of the Iraqi fiscal, as well as estimating the way big data analytics would enhance budgeting and cost allocation in our case. Research Objectives This work, it aims at achieving three objectives that we believe are necessary to comprehend the situation in Iraq: Validated Records: We directly review the present budgeting culture within Iraq, and we quantify the exact inefficiency in the forecasting of revenues, also in the allocation of expenditure and the implementation of the budget. Analytical Evaluation: We estimate how the big data approaches may be practically and feasible in enhancing the accuracy of cost allocation and budget preparation in the context of Iraq. Implementation Analysis: We estimate institutional and technological obstacles to effective implementation and we generate evidence-based policy proposals based on them. Research Question and Hypotheses Research Questions: RQ1: To what extent can Iraq predict revenue based on the existing mechanisms that we have had and how can it be improved in the event of adoption of big data analytics? RQ2: How effective are these ancient line-item methods of budgetingdo they work well to address the fiscal instability in the oil-reliant economies like Iraq? RQ3: What are the institutional and technological obstacles that impact the adoption of big data analytics in the Iraq budgetary processes and how would we overcome them? Theoretical Framework and Development of Hypotheses. To construct our hypotheses concerning big data analytics in the Iraqi public financial management, we combine three theoretical frameworks which are complementary to one another. It suggests that to make technological adoption successful, you should have a match between technological capability, organizational readiness, and even environmental support in the EJMEB (ISSN 3041-2102) | VOLUME 2 | NUMBER 6 | 2025 260 Technology-Organization-Environment framework or TOE (Tornatzky and Fleischer, 1990). This implies the difficulties of Iraq, they cut across various levels, in terms of infrastructure, capacity, institutional insecurity, simultaneously. Next we there is the Resource-Based View (Barney, 1991). It is a framework that takes analytical abilities as strategic assets. The point here is that, however, they are only useful with the help of other assets that complement them. Quality data, qualified staff, favorable institutions. And as far as IMF estimates are concerned, Iraq is deprived of these resources at the moment. Another layer is the contingency theory (Lawrence and Lorsch, 1967). It puts forward that organizational effectiveness relies on the alignment between structures and demands of the environment. Therefore, the extreme fiscal volatility of Iraq theoretically demands complex analytics, but the little ability of our country has, perhaps simpler methods are more suitable, at least at the start. When we combine these frameworks, it implies that technology is useful, they do not apply everywhere. The technological solutions must be preceded by institutional requirements. H1: The quality and availability of data, the volatility of the source of revenues, and the analytical capacity of organizations are positively moderating the effectiveness of big data analytics in enhancing the quality of revenue forecasting. I would like to explain this hypothesis. It conflicts with this concept of technological determinism by suggesting conditional advantages as opposed to unequivocal ones of higher-order analytics. They discovered that machine learning only enhanced the prediction of volatile sources of revenue such as property tax (Chung, Williams, and Do 2022). In the local governments of the United States, the conventional approaches even worked better than ML to ensure the sustenance of revenues. Contingency theory, it also projects that the sophistication of techniques should be consistent with the characteristics of tasks. It requires minimal methods of simple, stable forecasting. Complicated, intricate work, they could use developed methods, though they must have the preconditions. In the case of Iraq in particular, these are the IMF (2024) records of these serious deficiencies in data provision of low granularity, frequency, and timeliness. This implies that in our case, data quality is a binding limits analytics effectiveness. Having improvements in forecasting, we believe, is likely to focus on oil revenue (which is quite volatile) as compared to non-oil revenue (which is quite stable). Data quality issues will be solved only when we consider the underlying data quality problems. This is a conditionally negative hypothesis, it openly denies the null hypothesis of unconditional benefits that we occasionally encounter in practitioner literature. We suggest instead that organizational and data infrastructure they should evolve with analytical sophistication. This is moderated by organizational capacity, which is in line with the complementary assets of RBV. It is not possible to create value solely using advanced algorithms without qualified analysts to use algorithms appropriately and quality data to train them. H2: The level of budget system flexibility is correlated with fiscal variance negatively in the oildependent economies, and that the correlation is stronger with the rise in the degree of oil dependency. This is based on the premise of contingency theory by suggesting that uncertainty in the environment requires organizational flexibility (Burns and Stalker, 1961). We use this on budget systems as a structure of allocation of resources by organizations. The traditional line-item budgeting can be described as such a structure which we refer to as mechanistic structure, which is appropriate in a stable environment. Instead, performance based or program budgeting is organic structure in turbulent environments. This generates a structural incongruity within the oil based economies that is both highly volatile and structured on fixed budget models. As per the IMF (2014) analysis, it is evident that the GCC countries underwent falling consumption by over 30% in the 1980s oil price crash. Recovery took over 25 years. Nations that had more liberal fiscal systems, they had a safer ride through these shocks. They established that GCC countries which had better fiscal regulations and medium-term expenditure models, they attained better fiscal sustainability results (Magazzino et al., 2022). This mismatch is very clear in Iraq. Three years EJMEB (ISSN 3041-2102) | VOLUME 2 | NUMBER 6 | 2025 261 without sanctioned budgetsthis indicates a high level of rigidity in our system. Mean budget variance of -14% as compared to -8.5% regional average, it is a sign of systematic forecasting failures. And 45.2% public sector salaries will make the structure inflexible. The system is not flexible to revenue losses. This persistence is explained by the resource curse literature (Auty, 1993; Ross, 2012) as a result of the mechanisms of political economy. They allow patronage expenditure through oil revenues, and this generates political opposition to flexibility reforms. Weak institutions, they decrease the motivation to develop adaptive capacity. The fiscal variance will be reduced only when Iraq enhances flexibility of the budget system which can be in terms of medium term frameworks, contingency reserves, scenario planning and increased power of mid year reallocations. H3: The effectiveness of the implementation of big data analytics in the management of public finances is an outcome of the interaction between technological preparedness, organizational capacity, and environmental support, where organizational capacity is the bottleneck in the development of the country environment. This supposition is a continuation of TOE. Though all three situational contexts have to coincide, we suggest that organizational contingencies should be, they are the tie that binds in settings of resource constraint. Technology is relatively easy to acquire or even to transfer. Nonetheless, organizational ability, it is built gradually via learning-by-doing. It describes how barriers are perpetuated, based on three pillars (North, 1990; Scott, 2014), the institutional theory. Regulative (weak laws of data governance), normative (patronage norms that are immune to transparency), and cognitive (mental models of control over efficiency). All these inhibit change. Information of Heeks (2003) demonstrates that in developing countries, there is 85 percent failure rate of eGovernment projects. It is because of the so-called design-reality gapsthe discrepancies between the assumptions of the technology design and the realities in the institutions on the ground. They report what Andrews et al. (2013) refer to as premature load bearing. This is where the developing countries embrace complex systems even before the basic capabilities are realized. What is obtained is isomorphic mimicry--the copying without the doing. In the case of Iraq, the IMF (2024) specifically determines the organizational constraints as binding more than the technological ones. They mention that they have limited technical capacity is part of larger problems-institutional fragmentation, brain drain, lack of proper coordination mechanism, political instability. Technological hurdles such as inefficient electricity and old systemsthese can be solved by investing in them within a period of 2-3 years. But organizational barriers? The development of analytical capacity, the creation of data governance, alignment of the work of ministries this takes 5-10 years of active work. Environmental barriers, they rely on the factors of the political economy that are outside the technological reach. The implementation of analytics will only be successful when organizational capacity building comes before or at least to go hand in hand with technology implementation. Training investments, institutional coordination, governance structures, these are more critical than the software purchase or a hardware upgrade is. Literature Review Big Data Analytics in Public Financial Management There was a need to explore Big Data Analytics in Public Financial Management. It is fairly an important step in the evolution of traditional methods of government financial management as it involves the use of the highly developed analytical tools. However, the data on effectivenessit is also mixed and highly contextual, not consistently positive as one might assume. They performed this rigorous analysis of machine learning applications in revenue forecasting at the local governments of the United States (Chung, Williams, and Do, 2022). What was discovered by them was against all expectations. They found that the traditional statistical forecasting approaches were superior to the ML algorithms in general (p. 1133). K-Nearest Neighbors was the only ML EJMEB (ISSN 3041-2102) | VOLUME 2 | NUMBER 6 | 2025 262 approach that performed better and this was only in the case of property tax revenue which is volatile and non-linear. This observation implies--and it is especially significant to our case--that the advantages of advanced analytics, they are highly dependent on the peculiarities of revenues and tendencies of the data. It is not an all-over betterment. This conclusion has been supported by later studies as a result of contextual differences. Similar analyses reproduced by Victoria Department of Treasury and Finance in Australia found that machine learning algorithms are not effective in predicting payroll taxes better than models based on simple autoregressive forecasting, but could be effective in predicting land transfer duty (Victoria Department of Treasury and Finance, 2023). This was attributed to the ability to reduce dimensions when performing a large number of correlated features. Makridakis, Spiliotis and Assimakopoulos ( 2018) gave even bigger context. They have shown that machine learning algorithms, they do not have high success rates of outperforming simple models such ARIMA in terms of time-series forecasting. They point in their research that complexity of the modelcomplexity does not imply a better accuracy. In many cases, less complicated solutions become stronger in case of data quality or quantity shortage. In the context of the situation in Iraq, these results indicate that the potential value of big data analytics, it may focus on the areas that have a high concentration, in particular, oil revenue forecasting. The conventional practices may be sufficient in the case of more stable sources of revenue. It seems that the key factor, and it should be data quality and infrastructure, rather than the sophistication of the algorithm. Budget Preparation in Economies that rely on oil. It is a system with resource-dependent economies that struggle with these unique fiscal managerial issues. Conventional budgeting models, to their credit, they have difficulty in dealing with them. It is the inherent incompatibility of fixed budget structures and extremely volatile turnover of revenues-this generates structural weaknesses which we time and time again are witness to. International Monetary Fund captures these issues in a detailed manner in the comprehensive study they have carried out regarding economic diversification in GCC countries (International Monetary Fund, 2014). GCC countriesDuring the 1980s oil price crash, the GCC countries registered consumption per capita losses that topped 30%. It would not recover until late in the 2000s. As this historical episode shows, it is evident that hydrocarbon prices are unstable and are one of the major causes of macroeconomic volatility (p. 8). These produce boom-bust cycles and they cannot be well managed by traditional budget frameworks. Recent studies by Magazzino et al. (2022) when they compared fiscal sustainability in all six GCC countries using panel data of 1990-2017 showed them mixed results on the long-run fiscal sustainability. It is especially challenging to Saudi Arabia, Bahrain, and Qatar. The study validates that increasing the levels of debt requires more fiscal effort (p. 385). However, they add, oil-reliant nations, they find it hard to make requisite fiscal changes. This is because of the limitation of political economy and lack of alternative source of revenue. These are the dynamics, which especially are sharp in the case of Iraq. As oil dependency has always been over 88% of government revenue, which is already higher than most of the GCC members, Iraq has severe susceptibility to price shocks (IMF, 2024). It has shot up and oil price breakeven in the country. In 2020, it was at 54 per barrel and in 2024, it was at 84 per barrel. This is an indicator of higher expenditure obligations as well as less fiscal flexibility (IMF, 2024). Implementation Problems in the Developing Economy Environment. Although the case of digital transformation in the field of public financial management seems to make a lot of sense, on the one hand, a theoretical level is rather convincing; on the other hand, in practice, the significance of the obstacles to its realization in developing states is significant. They get little attention in the academic literature, and we have encountered them through our experience. The Article IV Consultation of Iraq by IMF in 2024, is an unusually clear report on these problems that we encounter. These inadequate data provision shortcomings are observed in EJMEB (ISSN 3041-2102) | VOLUME 2 | NUMBER 6 | 2025 263 the report, leading to poor technical capacity and to the state of digitalization as well as low granularity, frequency and timeliness of fiscal statistics (IMF, 2024, p. 29). It is these inadequacies that in essence limit the performance of any sophisticated analytical setup. Correct, prompt, detailed information this is the key to analytics. This fact cannot be neglected in Iraq. In addition to the technical infrastructure problems, there are institutional aspectsthey pose no less challenge. The study of the GCC fiscal policy, it underlines the fact that the economies reliant on oil have specific difficulties in maintaining fiscal discipline. This is owed to the nature of political economy. It is resource wealth, it allows patronage expenditure which is politically hard to cut, despite fiscal prudence requiring it (Brookings Institution, 2020). In the case of Iraq in particular, the salaries of the people employees in the public sector are already spending 45.2 percent of the entire budget. This drives out productive investment and inflexibility that restricts fiscal policy (Rudaw Research Center, 2024). The Tornatzky and Fleischer (1990) developed Technology-OrganizationEnvironment framework is an effective theoretical framework to explain these multidimensional barriers. Effective technology implementation, it needs to be aligned in three areas. To begin with, technological preparedness (infrastructure, systems, tools). Second, the organizational capacity (skills, processes, culture). Third, environmental facilitation (political commitment, regulatory structure, stakeholder acceptance). The issues of Iraq are observed across all three dimensions. This requires holistic solutions, and not technical ones. Local Evidence from Iraq In the domestic context The small amount of empirical research evidence has been conducted on the degree to which the Iraqi organizations incorporate the analytical tools in their budgeting systems and financial management systems. Al-Dabagh and Al-Sayegh (2024) have shown that financial intelligence knowledge and behavioral factors have significant influence on the formation of the investment decisions and transparency in operation of the Iraqi Stock Exchange. Combined, these local-based studies have offered tangible evidence to support the main thesis of this study, which is that datadriven and technology-intensive practices are gradually transforming the management of public finances in Iraq and provide a good empirical foundation to the analysis of the presented case. Methodology The research design that we use in this study is the mixed-method. It consists of quantitative examination of the fiscal performance of Iraq and qualitative examination of institutional preparedness to technological change. The main sources of our primary data will be Iraqi Ministry of Finance Annual Budget Laws and Execution Reports, 2015-2024. In addition to that Central Bank of Iraq Monthly Statistical Bulletins, IMF Article IV Consultation Reports, World Bank MENA Economic Updates, and U.S. Energy Information Administration oil price information. These were complemented by regional reports of the Atlantic Council and other research centers. The data set has eight sets of annual observations2015-2024. We had to exclude 2014, 2020, and 2022 since no year had approved any budgets following the political unrest and crisis conditions. The information includes budgetary allocation, real revenues, oil prices, the World Bank Worldwide Governance Indicators on political stability and the GDP growth rates. Since the IMF (2024) explicitly reports serious failures in data provision such as low granularity, frequency and timeliness, we view all our results in the context of these limitations. We should be factual with the data constraints. It is an analytical process that is a combination of four approaches. First approach: descriptive statistic analysis. We compute the budget variance: (Actual Revenue Forecast Revenue)/Forecast Revenue x 100. Coefficient of variation and oil dependency also. Second: OLS regression equations. These are model revenue, which were financially equivalent to the oil price, political stability (measured by WGI index) and GDP growth. We incorporate extensive diagnostics Breusch-Pagan tests of heteroscedasticity, Durbin-Watson tests of autocorrelation, Variance EJMEB (ISSN 3041-2102) | VOLUME 2 | NUMBER 6 | 2025 264 Inflation Factors of multicollinearity and Jarque-Bra test of normality. Third approach is the time forecasting through AR(1) models. More complicated ARIMA or SARIMA specifications, they would need 36+ observations to have monthly seasonality. This is not possible using eight yearly data points. Fitting models whose parameters are more than a fifth of the existing observationsthis may overfit hence we do not. Fourth: comparison and benchmarking with the GCC peer countries based on IMF and World Bank data. This puts the performance of Iraq in perspective against other like economies, which rely on oil. In order to carry out the cost-benefit analysis we estimate the cost of implementation to be an amount of 415 million in 5 years. This is broken down as IT infrastructure $250M, software $75M, training $40M, consulting $50M. Our projected benefits due to annual increases in forecasting (530M), budget implementation (1162M), and corruption/leakage (3875M), are 5,300,000. The theoretical ROI of this is 6,331% though this is subject to an effective implementation and institutional change. Necessary methodological flaws-- we have to admit these. The highly restrictive sample (n=8) adversely affects statistical power. Setting up budgets that are missing years would result in structural gaps. The IMF reports data quality issues. Correlational analysis cannot be used to state causality. The external validity is minimized because of the extreme situation in the state of Iraq92% oil dependence, history of conflicts. And the counterfactual issue--we can not measure the results of other conditions except by experiment or quasi-experiment. All these imply that what we can do with findings is see them as suggestive patterns that should be supported, rather than concluding. Variable Operationalization The operationalization of the variable is as follows: Our major variables we operationalized in the following mannerwe were very careful, however, about consistency in measurement. Revenue Variance: this is a percentage difference between real and projected revenue. We compute it as (Actual - Forecast)/Forecast x 100 by using the reports of the Ministry of Finance. Budget Execution rate: this involves the calculation of the actual expenditure in comparison to the approved budget hence (Actual/Approved) x 100. Oil Price is the average annual Brent crude spot price in USD/barrel of U.S EIA. Our anticipation is that there will be a good relation to revenue performance here. Political Stability--we take the index of Political Stability and Absence of Violence index of the World Bank Worldwide Governance Indicators. It ranges from -2.5 to +2.5. To predict variance, we believe that a negative correlation is expected to predict errors since instability maximizes errors. GDP Growth is an indicator of the real GDP annual percentage change of World Bank World Development Indicators. We anticipate good association with financial results. Oil Dependency, this is the calculation of oil revenue as a percentage of total government revenue- (Oil Revenue/Total Revenue x 100) based on the ministry of finance and EIA statistics. We measured all the independent variables together with dependent variables at the same time. Diagnostic tests also confirm that there are no severe OLS assumption violations, despite the known limitations in the data that we have. Analytical Models In the model of determinants of the revenue variance, we will model the revenue variance as follows: RevenueVariancet = b0 + b1(OilPricet) + b2(PoliticalStabilityt) + b3(GDPGrowtht) + et. In this case, b1 should be negative, higher prices will decrease the errors in the forecasts, b2 should be negative, persistent undermines higher errors, b3 should be positive, growth enhances performance. Our standard OLS estimation has heteroscedasticity-robust standard errors. It is the time-series forecasting, which specifies AR(1): Revenuet = a + b Ravenget t -1 + et. The reason behind this decision is that more complicated ARIMA(p,d,q) or SARIMA models would demand a significantly larger amount of observations than we are able to provide, which are eight in our case. We do not take specifications in which parameters are more than about 20 percent of sample size, which is usual practice. This prevents overfitting. There are a number of tests in model EJMEB (ISSN 3041-2102) | VOLUME 2 | NUMBER 6 | 2025 265 diagnostics. Breusch-Pagan tests (H 0: homoscedastic errors). Durbin-watson statistics to test autocorrelation-2.0 portrays that there is no autocorrelation, less than 2 and more than 2 represent positive and negative respectively. Variance Inflation Factors to determine multicollinearityVIF above 10 is a matter of concern. Jarque-Braun statistics of residual normality. And analysis of residual plots of systematic patterns. We present all findings in a transparent manner and make corresponding corrections in case of violation of assumptions. Results Descriptive Analysis Combined, table 1 and table 2 sum up the fiscal performance of Iraq in 2015-2024. They demonstrate the trends and aggregate patterns of budget results, oil addiction, and such macroindicators. Table 1. Descriptive and Summary Statistics – Iraq's Fiscal Performance (2015–2024) Year Approved Budget (B USD) Actual Revenue (B USD) Budget Variance (%) Oil Dependency (%) 2015 87 67 -23.0 94 2016 88 71 -19.3 93 2017 95 78 -17.9 92 2018 106 97 -8.5 92 2019 112 105 -6.3 91 2021 90 92 +2.2 90 2023 153 138 -9.8 93 2024 163 115 -29.4 88 Note: Years 2014, 2020, and 2022 excluded due to political crises. Source: Iraqi Ministry of Finance; IMF Country Reports; Shafaq News. Table 2. Summary Statistics – Key Fiscal Indicators (Iraq, 2015–2024) Variable Obs Mean Std. Dev. Min Max Approved Budget (Billion USD) 8 111.75 31.21 87 163 Actual Revenue (Billion USD) 8 95.38 26.56 67 138 Budget Variance (%) 8 -14.00 10.24 -29.45 2.22 Oil Dependency (%) 8 91.63 1.85 88 94 Oil Price (USD per barrel) 8 65.26 14.98 43.7 83.5 Political Stability (Index) 8 -1.93 0.21 -2.2 -1.6 GDP Growth (%) 8 2.31 4.41 -2.2 11.0 The combination of the evidence provided by these two tables allows stating that the fiscal performance of the Republic of Iraq in the period between 2015 and 2024 was determined by the constant instability and this unrealistic reliance on oil revenues. The real revenues, they were lower than the projections in almost all years. This brought about an average budget variance of 14 percent. The only excess was in 2021 (+2.2 percent) as the oil rebounding took place postpandemic. It suffered the most acute deficit, and it was in 2024 (-29.4 percent). This was after OPEC+ production restricts and the decrease of oil prices worldwide. These fiscal imbalances, they also point out their severity through the summary statistics. Budget variance coefficient of variance (0.73) -this is much higher than that of OECD or GCC. It attests to extreme instability in our financial system. The average oil dependence stood at 91.6 with the variation being insignificant. This means that there is no physical diversification move, even after several reform EJMEB (ISSN 3041-2102) | VOLUME 2 | NUMBER 6 | 2025 266 programs that we have undertaken. Stability in politics was poor all the time (average = -1.93). It was volatility of GDP growth (-2.2 to 11 percent) which was indicative of repetitive shocks in the oil markets and security situation. Taken together, all these findings point to a fiscal system that is in some form of a high-volatility, low-diversification, equilibrium. This correlates with the previous literature (Agostino et al., 2021; Horngren et al., 2018). Their traditional budgeting processes lack resilience or predictability of future in our resource-driven environment. Correlation Analysis Correlation analysis was done before we ran the regression model. This was to determine the correlation between the important fiscal and macroeconomic variables in Iraq. Also to identify possible multicollinearity conditions. Table 3. Correlation Matrix – Iraq Fiscal Indicators (2015–2024) Variable Budget Variance Oil Price Political Stability GDP Growth Budget Variance 1.0000 Oil Price -0.3025 1.0000 Political Stability 0.3156 0.8924* 1.0000 GDP Growth -0.1947 0.6103 0.5249 1.0000 Note: *Correlation significant at the 0.05 level (two-tailed). Table 3 shows that there are some significant relationships between fiscal and economic variables. Yet it also underlines some econometric issues which we encounter. The expected negative relationship between the oil prices and the budget variance is evident (-0.30). This suggests that increased prices decrease errors in revenue forecasts but the connection, it is not very strong. It seems counterintuitive at the very beginning that there is a positive relationship between political stability and budget variance (0.32). However, do not forget that the political stability index is negative or the higher the negative, the less stable it is. So this positive relationship, it in fact validates the fact that instability exacerbates fiscal performance. The most problematic aspect of our regression modeling is that the correlation between oil prices and political stability is very high (0.89). This is an indication of high multicollinearity which will overinflate the standard errors and decrease the statistical power of any model having the two variables. It is not surprising intellectually that this multicollinearity should be. They increase the ability of the government to stay afloat with patronage expenditures, as oil revenues rise. But it makes it difficult to effectively extract their influences on fiscal performance. The medium degree of relationship between oil prices and the growth of GDP (0.61)- this again proves the point of Iraq being excessively dependent on the petroleum sector in its economy. Regression - Budget Variance Determinants It is Table 4 that provides the results of our OLS regression model. We hypothesized that oil price, political stability, and GDP growth are significant in affecting the budget variance in Iraq between the year 2015 and 2024. Strong standard errors were used to find possible heteroscedasticity within our small sample. Table 4. OLS Regression Results – Determinants of Budget Variance (Iraq, 2015–2024) Variable Coefficient Robust SE t p [95% Confidence Interval] Oil Price -0.148 1.024 -0.14 0.892 [-3.05, 2.76] Political Stability 8.782 58.218 0.15 0.887 [-154.8, 172.3] GDP Growth -1.090 2.271 -0.48 0.656 [-7.34, 5.16]