Impacts of crude oil market on global economy: Evidence from the Ukraine-Russia conflict via fuzzy models

The increasing Russia-Ukraine crisis is without a doubt Europe's most prominent conflict since World War II, changing the dynamics of the oil and other key markets. Because the oil market has traditionally interacted with other financial and commodity markets, it will be intriguing to examine how it interacts with substantial financial assets amid market volatility induced by a conflict. The goal of this study is to propose a fuzzy time series (FTS) model and to compare its competitiveness to existing fuzzy time series (FTS) models, Autoregressive Integrated Moving Average (ARIMA) model and some machine learning methods i.e. Artificial Neural Networks (ANN), Support Vector Machine (SVM) and XGBoost models. We considered changes in the partitioning universe of discourse, optimization of parameters method(s), and interval estimation to make the forecast accuracy more precise forecasting than traditional methods via MAPE. The event-based data results show the proposed fuzzy time series model is outperforming all the competitive methods in the study. Furthermore, the proposed model forecasting shows a future decline tendency in WTi market crude oil prices (US$/BBL) after being at the record highest level, which is good news for the worldwide economy.


Introduction
Crude oil is one of the most challenging markets to forecast because it is influenced by so many conflicting crosscurrents, such as supply and demand, global economic health, geopolitics, and the global monetary and regulatory environment.When there is a war in an oil-producing country, oil prices rise.It's one of the most foreseen market developments.In the short term, two major geopolitical developments could have an immediate and negative impact on the oil price: the first is an escalation of the Russia-Ukraine conflict, which could disrupt Russian oil and gas supplies to the European Union; and the second is a worsening of the situation in Iraq, which could affect its oil infrastructure and production (EU).In the long run, however, the major risk to global oil supply and the price of oil comes from the Arab Gulf nations' rapidly expanding domestic oil demand and lack of economic diversification.According to the research [1], an interruption in Iraq's oil output could boost oil prices to more than $140/barrel, while a disruption in Russian oil supply to the EU could easily add $20-$30 to the price of oil.It will also assert that unless the Arab Gulf countries drastically cut their domestic oil use or replace oil with nuclear and solar energy in power generation and water desalination, their oil exports would almost likely fail by 2032.The conflict has had a substantial influence on the global economy, causing growth to stagnate and expenses to rise.Aside from the pain and humanitarian issues brought about by Russia's incursion of Ukraine, the global economy will be affected from slower growth and higher pricing.The consequences will be felt through three major channels.One, higher commodity costs, such as food and energy, will push up inflation further, reducing buying power and influencing demand.Two economies, particularly those next to one other, will see disruptions in commerce, logistics processes, and remittances, as well as an unprecedented rise in refugee flow.Third, falling company confidence and rising investor concern would exert downward pressure on asset prices, tightening economic circumstances and perhaps promoting capital outflows from developing nations.Russia and Ukraine both have substantial resources.Russia and Ukraine are major resource producers, and supply disruptions have driven up worldwide prices, notably for oil and natural gas.Wheat has reached a record high, accounting for 30 % of Ukraine and Russia's global exports.Nations with direct trade, tourism, or financial exposure will suffer additional pressures in addition to global spillovers.Oil-importing nations may run greater budget and trade deficits, as well as face stronger inflationary pressures; nevertheless, certain exporters, particularly those in the Middle East and Africa, may benefit from higher prices.Higher food and fuel costs are expected to exacerbate the risk of instability in a number of nations, ranging from Sub-Saharan Africa and Latin America to the Caucasus and Central Asia, while food insecurity in Africa and the Middle East is expected to intensify.These repercussions are difficult to foresee, but we believe our growth forecasts will be revised downward next month when we provide a more comprehensive picture of our World Economic Outlook and regional reviews.Long-term, the conflict may fundamentally alter the global economic and geopolitical order as oil trade moves, supply lines reorganize, payment networks split, and governments review their reserve currency holdings.The danger of economic fragmentation, particularly in commerce and technology, grows as geopolitical conflict escalates.
In the intricate web of global economics, few events carry as profound an impact as geopolitical conflicts, and the ongoing Ukraine-Russia war stands as a testament to this reality.As the world watches with bated breath, the tendrils of this conflict extend far beyond national borders, intertwining with the complex tapestry of the WTI crude oil market and reverberating throughout the global economy.Crude oil, that lifeblood of modern industry, is uniquely sensitive to geopolitical tremors.Supply and demand dance a precarious duet, influenced not only by economic health but also by the unpredictable choreography of political forces.In this intricate ballet, a war in an oil-producing region serves as a crescendo, sending oil prices soaring in anticipation of disruption.Thus, it is no surprise that market experts are closely monitoring the Ukraine-Russia conflict, identifying key triggers that could resonate with global oil prices.In the immediate horizon, the conflict's escalation holds the potential for swift and unfavorable consequences.The specter of a Russia-Ukraine confrontation looms large, casting shadows over Russian oil and gas supplies to the European Union.Simultaneously, the fragility of Iraq's oil infrastructure underscores the delicate balance between geopolitical instability and oil production.The ripple effects of these potential disruptions are felt not merely as localized tremors, but as seismic shifts in the global oil market landscape.Peering into the distant horizon, the landscape becomes equally intriguing.Beyond the immediate price hikes lies a more formidable challengethe burgeoning oil demand of Arab Gulf nations.With domestic needs surging and economic diversification lagging, the global oil supply chain faces a reckoning.Research suggests that even an interruption in Iraq's oil output could jolt prices beyond $140 per barrel, while disruptions in Russian oil supplies could inject an additional $20-$30 premium.Meanwhile, the looming threat of dwindling Arab Gulf oil exports by 2032 underscores the urgency of energy diversification.But the impact of this conflict doesn't stay confined to the world of commodities and markets.Its tendrils wind through the corridors of economies and societies alike, triggering a symphony of repercussions.As growth stalls and costs rise, the global economy feels the tremors of uncertainty.A multi-faceted disruption unfolds: commodity costs surge, weighing down purchasing power and dampening demand.Trade routes falter, cascading into disruptions in logistics, commerce, and remittances, and swelling the ranks of displaced populations.
The psychological toll is equally undeniable.Plummeting business confidence and investor trepidation conspire to tighten the economic vise.Asset prices falter under the weight of uncertainty, potentially triggering capital outflows from emerging economies.The intertwined destinies of Russia and Ukraine, both resource-rich nations, come into stark focus.Disruptions in supply chains send shockwaves across global prices, leaving no corner of the world untouched.From the fertile plains of wheat to the intricate webs of trade, the impact is far-reaching, fostering a heightened sense of interdependence in a world already grappling with complexities.As economists scrutinize their forecasts and policymakers strategize their moves, a common thread emerges: uncertainty.Predicting the full extent of the impact remains an elusive endeavor, as geopolitical intricacies continue to intertwine with economic realities.Yet, one thing is certainthe echoes of the Ukraine-Russia conflict will continue to resonate through global oil markets and reverberate across economies, weaving an intricate narrative of interconnectivity in an ever-evolving world.The past studies relevant to the study is briefly discussed in the literature review section.

Literature review
The current and future research [2] explores the effects of economic shocks on areas such as Europe, the Caucasus, Central Asia, the Middle East, North Africa, Sub-Saharan Africa, the Western Hemisphere, Asia and the Pacific, and Global Shocks.Researchers [3,4] identified the ongoing Russia-Ukraine conflict as the most significant war in Europe since World War II, affecting the dynamics of oil and other important markets and negatively impacting the global economy as a whole.Their research was further enlightened by Ref. [5], who emphasized that the Eurozone manufacturing purchasing managers' index (PMI) fell in the month of the invasion and concluded as food and petroleum prices rose rapidly due to irreparable infrastructure damage.They express their sympathy to Ukraine, which suffered a more severe impact from the invasion than Russia and the entire Eurozone, and they extend it to the world economy.Researchers [6] examined the global GDP further and attempted to present 2023 GDP forecasts for various nations, which would have a detrimental impact on their economies.
The researchers [7], used an event research technique to assess the influence on stock markets across nations following the Russian invasion on February 24, 2022.They observed that EU nations have seen a large drop in cumulative anomalous returns, although enterprises in other countries outside of the war appear to be unaffected.They also believe that economic considerations will be the most compelling argument for Russian administrations to end their conflict in Ukraine.Their research was bolstered by Ref. [8], who urged international leaders to persuade Russia to leave Ukraine or face punitive sanctions.
The authors [9] investigated the transmission of returns and volatility in the commodities universe in the aftermath of the Ukraine war.The overall volatility spillover rises from 35 % to 85 %, surpassing the amount observed during the epidemic.Crude oil becomes a net transmitter of return spillovers, while wheat and soybeans become net receivers.Silver, gold, copper, platinum, aluminum, and sugar become net volatility transmitters.Geopolitical threat spillover indices are caused by Granger.High levels of return and volatility have been linked to high levels of geopolitical risk.Their research on geopolitical risk was expanded upon by Ref. [10], which investigated the impact of geopolitical risks caused by the Russian-Ukrainian conflict on Russia, European financial markets, and global commodity markets using time-and frequency-based time-varying parameter vector autoregression (TVP-VAR) approaches.They investigated the geopolitical aspects and repercussions of gas supply and demand in Europe during the ongoing war [11].
As an event case study [12], examined the impact of a Russian invasion of Ukraine in 2022, concluding that the invasion created negative cumulative abnormal returns for global stock market indices, albeit with diverse consequences.Economic globalization, as measured by GDP-scaled trade, is inversely related to event-day and post-event returns, according to a cross-sectional study.Markets in NATO nations showed better returns, consistent with the projected economic boost of military preparation.Their research is further mirrored in Ref. [13], where equities more vulnerable to the regulatory concerns of the transition to a low-carbon economy did better, implying that investors expect the change to decelerate.Analysts have also raised their profit expectations for these stocks.The stock price consequences for transition risk were especially substantial in the United States.The impacts on transition risk were less strong or perhaps inverse in Europe.Stocks with low-carbon transition potential gained, presumably because market players anticipate stronger governmental responses supporting renewable energy sources in the face of Europe's considerable reliance on Russian oil and gas.As a result, investors anticipate that the pace of transition to a low-carbon economy will differ between the United States and Europe.The analysis accounts for a variety of Environmental, Social, and Governance (ESG) indicators.Companies that mentioned inflation more frequently in analyst conference calls did badly.Internationally focused corporations performed poorly, and investors were particularly concerned about companies' exposure to China.
The researchers [13], investigated significant environmental contamination as well as several other livelihood concerns.They think that the Russian-Ukrainian conflict created a tsunami that had a significant influence on the global economy, geopolitics, and food security.Because of the grave humanitarian crisis, the environmental consequences have been disregarded.However, because of the fierce warfare, the consequences will be tremendous, resulting in an environmental disaster.The violence is already influencing places outside of Ukraine (explosions in Russia and Moldova territory).The prolonged battle causes significant air pollution and greenhouse gas emissions.In addition, combat actions were carried out near the Zaporizhzhia nuclear power station (the largest in Europe) and Chornobyl, heightening fears of radioactive leakage.Deforestation and habitat degradation are wreaking havoc on biodiversity, with serious consequences for species.Bombing, trenching, and tunnel excavations will almost certainly have a detrimental influence on soil deterioration and landscape morphology.This is especially important because Ukraine has some of the most fertile soils in the world (Chernozem), which affects food production.Water availability and quality are anticipated to suffer as a result of infrastructure destruction and pollution movement to water reservoirs.The ecosystem services provided will most likely be severely harmed because deforestation reduces ecosystems' ability to manage air pollution or temperature.Soil pollution will impede food production, and the erosion of landscape aesthetics, cultural assets, and social cohesiveness will have a significant impact on cultural services.Finally, the consequences for human health are already enormous.However, it can be significantly greater due to high levels of pollution and deterioration of sanitary conditions.The consequences of this conflict are quite unknown.
The key contributions of this study are as under.
• Proposed Fuzzy Time Series Model.

Research methodology
In this research study, we have proposed a new FTS model in comparison to FTS Stevenson and Porter Model referred to as the SP model [14] and Autoregressive Integrated Moving Average (ARIMA).By using the percentage change of the dataset for forecasting as UoD, on which fuzzy sets are given, the interval of variations of such indicators as the year-to-year percentage change rates, called "chain growth rates".
Consider a time series of an economic indicator, denoted by y i , i = 2, 3,…,n.For modeling this time series, the considered method uses the following indicator as chain growth rate in eq. ( 1): Hereafter, the model contains the following steps.
Step 1. Determine the UoD as the Set U, where U = [min T i ,max T i ], i = 1, 2, 3, …, n and divide into equal "m" intervals, obtained by eq. ( 2) as: Where "m" is the class interval and "N" is the total number of observations in "U".
Step 2. After frequency distribution, naturally, large classes are divided into smaller partitions for accuracy.In our dataset, we have divided the larger three classes into 5-3-11 sub-intervals respectively for improving accuracy.
Step 3. Fuzzy sets X k , k = 1, 2, 3, …, m on each partition interval as the triangular fuzzy number, which carriers are defined by three values: lower limit, midpoint, and upper limit.For the actual dataset, it is to determine which fuzzy set will describe each value.So that fuzzifies the dataset into initial series.
Step 4. Defuzzification of the fuzzy set to crisp values is done by eq. ( 3): Where a j− 1 , a j , and a j+1 are the midpoints of the interval of fuzzy sets X k− 1 , X k , and X k+1 carriers respectively.
Step 5. Determine the predicted values for each level of our proposed model as under in eq. ( 4): As in Ref. [14], the forecasting results by eq. ( 4) are supposed to have a better accuracy rate than other FTS models as well as conventional time series models.However, the accuracy may be further improved if the t j parameters rather than specified values as in eq.
(3) are estimated and the optimized values are used for forecasting the future value.In this study, we have implemented and will be discussed later.
Case Study: Crude Oil Prices ($/BBL) during Russia Ukraine War.In Fig. 1, international market WTi crude oil daily prices ($/BBL) 10 years data has been presented i.e. from Nov 26, 2012, to Nov 23, 2022.Before and after Ukraine-Russia War by putting a dissection line on February 17, 2022, to onward as war data which is our case study for this article.During COVID-19, prices recorded a historic low while consecutively high for crude oil international since the war break out, therefore its assessment and forecast are our interest area to investigate.
In Fig. 2, WTi market crude oil prices ($/BBL) during Ukraine-Russia war daily data has been presented for a clearer insight into the case data.During the war, we see that minimum crude oil prices ($/BBL) were recorded as 85.56 and maximum as 110.53.By eq. ( 1), we get the growth rate (T) as under.
In Table 1, Growth Rate is displayed against corresponding values, we observed that it is varying from − 0.12126 to 0.083544 as the   minimum and maximum values.We form a universe of discourse as U = [− 0.12126, 0.083544 ] and its frequency distribution using class boundaries limits system by using eq.( 2), huge contributing classes i.e 5th, 6th , and 7th having frequencies 44, 22, and 87 respectively were further divided into 5, 3, 11 sub-intervals, fuzzy sets X i , i = 1, 2, 3, …, 25 were applied in the fuzzification process.By using eq.( 3), we can fit our proposed model as in eq. ( 4) to defuzzified data as under: In Table 2, the proposed model fitted against observed values using defuzzified values and was compared graphically to Stevenson and Porter's (SP) model [14] in Fig. 3.
In Fig. 3, we see that the proposed model, the plot overlaps the SP model and as a result, shows a single curve, that fits the observed data well than the SP model.It is a clear indication of better forecast accuracy than the SP Model.However, the forecast accuracy may be increased if the parameter under (3) is estimated rather than using specified values as 0.5 each.The parameters γ 1 , γ 2 , δ 1 and δ 2 are to be optimized for said purpose.The general form of adjusted defuzzified t adj j is given as under in eq. ( 5): For optimization of these parameters, a simultaneous equation system through Maple 17 was used and got the optimized values as γ 1 = 0.5000000029, γ 2 = 0.499999996, δ 1 = 12.00840071 and δ 2 = 53.79319947.We observed a huge difference in the optimized values of the last two parameters as compared to specified values.Fig. 4 shows the comparison of all studied methods in this article.
As it is shown in Fig. 4, the plot of the proposed model with optimized parameters overlaps all other model plots and is more visible  than all methods in this scenario.So, it is evident that the proposed model improved the forecasting accuracy by having optimized parameters.So, models with specific parametric values have less accuracy than optimized ones.The model forecasting performance is given as a Fig. 5.In Fig. 5, we tried to check the forecasting performance of models i.e. from Nov 24, 2022, to Dec 31, 2022.As like Fig. 4, the plot of the proposed model with optimized parameters overlaps all other model plots and is more visible than all methods in this scenario.So, it is evident that the proposed model with optimized values is performing better model forecasting than all others.So we may conclude that in fuzzy time series, usually optimized parameters play important in improving the forecasting accuracy of the model.It shows that up to Dec 31, 2022, the price of crude oil ($/barrel) will keep lowering at a very slow rate during the Ukraine-Russia conflict and is good for the worldwide economy, if the tendency continues to fall for a long.
ARIMA (1,0,0) was utilized in the analysis as conventional time series through auto.arimafunction in Gretl, the significance of parameters was chosen by AIC and BIC criteria.The result of model forecasting with a complete summary is given in Tables 3 and 4.
From the above ARIMA (1,0,0) model, the autoregressive term is highly significant.Fig. 6 gives us a vibrant look into in-sample and out-of-sample forecasting with real-time forecasting with a 95 % confidence interval.

Results
The overall comparison summary of all models is given in Table 5, showing the MAPE of all models i.e.Proposed, SP, ARIMA, Artificial Neural Networks (ANN) [15], Support Vector Machine (SVM) [16] and XGBoost [17] models with respective relative efficiency, keeping the SP model as the base model with a relative efficiency of 100 %, it is clear from the table that the Proposed Model has higher relative efficiency (marked as bold) than all competitive methods in this study.Table 5 shows that the proposed model with adjusted parameters has greater forecasting accuracy than all competitive methods.

Discussion
In this article we tried to find out the best forecasting method for the crude oil price (USD/Barrel) WTi international market using the dataset from February 17, 2022, to Nov 23, 2022, in assessing the international market against Ukraine-Russia War which was started on February 24, 2022, and going on.We found that our proposed model is the best-performing forecasting model versus other fuzzy time series, ARIMA, ANN, SVM and XGBoost models.The in-sample and the out-sample forecast show Proposed Model has the least MAPE against the base model.The base SP model [14], which was outperformed by proposed model in both fitting and forecasting to real-time data in having specified parameters as well as with estimated optimized parameters.We further made its comparison to ARIMA(1,0,0), ANN, SVM and XGBoost models for better insight performance of the proposed model and found it better than it.Overall, we conclude that FTS models perform better than conventional time series models.The main issue in the FTS model in forming the universe of discourse and its partitioning.MAPE is highly dependent on it.MAPE will be least if the partitioning is least otherwise not.
Our finding, out of the sample forecast shows that the international market WTi crude oil prices (US $/BBL) will keep lowering with every passing day after hitting the record highest price in history.The lowering price effect may be seen in the world economy as initial war days were not so good for it is higher prices.Higher prices of crude oil have huge consequences on oil-importing countries and their GDP is directly linked to US Dollar.So, Ukraine-Russia war bears huge economic consequences for the whole world.

Conclusion
The key findings of the article are given below.
•The proposed method in fuzzy time series is performing better than its competitors for both specified and optimized parameters.
•There is no unique universe of discourse as in SP model claimed, i.e. in the previous literature.The choice of UoD is a subjective approach.
•Forecasting through interval estimation gives more information than point forecasting.
•Relative efficiency is a good addition in research for checking model accuracy.

Future perspective
The further research directions on same study are as under.
•Assessment of global economy in the presence of Ukraine-Russia War using Multivariate Fuzzy Time Series.
•Machine Learning Techniques may also be utilized for deep insight knowledge.•Comparative analysis of multivariate time series models versus fuzzy and machine learning methods can also be adopted.

Data availability statement
The data was taken from WTi website.i.e. https://www.investing.com/commodities/crude-oil.

M
. Bilal et al.

Fig. 3 .
Fig. 3. Comparison of proposed model to SP model against observed data.

Fig. 4 .
Fig. 4. Comparison of methods with specified and optimized parameters.

Table 1
Growth rate against price (USD/Barrel) in WTi market.

Table 2
Fitting proposed model (Y t ) against actual crude oil price (USD/BBL).

Table 5
Overall comparison of all models.
M. Bilal et al.