Modelling asymmetric and nonlinear features in the natural resource wealth-economic complexity nexus: empirical insights from Nigeria

Abstract

Contrary to previous research, this study makes a unique contribution to the global discussion by incorporating asymmetric structure and nonlinearity into the analysis of how changes in natural resource wealth affect economic complexity. To achieve this objective, the study uses a nonlinear autoregressive distributed lag (ARDL) and a fully modified ordinary least squares estimator, utilizing data from Nigeria spanning the period 1984–2021. Unlike earlier studies, this study establishes robust evidence of nonlinearity and asymmetry in the sensitivity of economic complexity to changes in natural resource rents in the short and long run. The cumulative increases (positive shock components) in natural resource wealth provide strong stimuli and incentives that promote economic complexity in the short run, while the cumulative decreases (negative shock components) deteriorate economic complexity upgrades. Meanwhile, long-run estimates indicate that both positive and negative shock components are catalysts that impede Nigeria’s manufacturing structures’ ability to improve technological innovation and knowledge-based productive capacity for producing sophisticated and globally competitive exports. These findings imply that the resource curse phenomenon holds true regarding economic complexity in Nigeria in the long run. In conclusion, this study finds that Nigeria’s natural resource endowments breed complacency, racketeering, shrewdness, corrupt practices, and opportunistic behaviour, which impair innovative initiatives that spur economic complexity. This study outlines the policy implications and insights from the findings.

Introduction

Academics, policymakers, and development experts regard the economic complexity index as a comprehensive measure of economic development (Ajide 2022; Aslam et al. 2022; Ketu and Ningaye 2024; Lapatinas 2019; Nguyen and Su 2021a, b; Olaniyi and Odhiambo 2023a). This index assesses the level of technological sophistication and knowledge-based productive capacities embedded in the productive structure (Ashraf et al. 2023; Avom et al. 2022; Balland et al. 2022; Gnangnon 2022; Oumbé et al. 2023). Economic complexity provides more objective explanations for why some countries are more developed than others (Antonietti and Franco 2021; Shahmoradi et al. 2023; Hartmann et al. 2017; Hidalgo and Hausmann 2009; Njangang and Nvuh-Njoya 2023; Olaniyi and Odhiambo 2023a). While certain advanced countries manufacture a variety of high-quality and globally competitive exports, many others produce less sophisticated goods with a low chance of penetrating the international market. This suggests that different levels of economic complexity in the productive structure explain the varying levels of economic development across countries worldwide. In addition to being broader in scope than economic growth, economic complexity is a superior indicator of economic development. It is a metric that encompasses an economy’s technological and knowledge-based productive capacity to produce sophisticated and competitive exports, along with its explanation of economic components (Ajide, Osinubi Ojeyinka, 2023). Given its relevance and significance to development priorities, this index continues to be subject to empirical scrutiny. Numerous studies have emerged to explore the various factors that contribute to economic complexity. These factors include natural resource endowment, financial development, institutional quality, foreign direct investment, foreign aid, trade openness, remittances, and more. This study specifically focuses on one such factor of economic complexity: natural resource wealth. The intricacies often arise in the natural resource wealth-economic complexity nexus due to the resource curse phenomenon. When a country heavily relies on its natural resources endowment, it often neglects to use resource income to engage in diversification into productive and innovative sectors. This situation becomes aggravated as shrewd practices, such as racketeering, corruption, and opportunistic inclinations, thrive in the management of resource income. This problem often stifles the efficient allocation of resource wealth to finance investments in initiatives and components that could enhance knowledge intensity and technological advancement to produce sophisticated and competitive exports.

Natural resource endowments should offer resource-rich countries the opportunity to use their resource wealth as capital to diversify into innovative and productive industries. This natural resource income could also serve as capital to invest in strategic initiatives such as the development of human capital, innovation and technology, research and development (R&D), education, high-quality infrastructure, entrepreneurial prospects, and other components that can enhance technological innovation and knowledge-based productive capabilities, thereby facilitating the production of a wide range of sophisticated exports. This suggests that efficient management and allocation of natural resource wealth can foster diversification and generate capital to drive economic complexity upgrades in countries abundant in resources. Unfortunately, many resource-rich countries are struggling with unethical practices caused by weak institutions, resulting in racketeering, opportunism, rent-seeking, corruption, and political interference in the management of natural resource wealth. These practices discourage the utilization of resource income to promote improvements in economic complexity. As a result, many resource-rich countries find themselves trapped in low-innovative industries with weak technology, limiting the level of knowledge intensity and technological sophistication in their economies. This generalization may not apply to every resource-rich country. Countries like Canada, Norway, Russia, the United Kingdom, and the United States, which possess abundant natural resources, tend to have high scores in the global ranking of economic complexity. However, there are also countries such as Angola, Libya, Nigeria, and Saudi Arabia that have abundant resources but perform poorly in the global ranking of economic complexity. Studies have started investigating the impact of natural resource wealth on economic complexity. These research efforts help validate or refute the existence of the resource curse hypothesis in resource-rich countries.

Existing studies’ outcomes predominantly show that natural resource wealth impedes economic complexity upgrades (Valentine et al. 2024; Zhang et al. 2023; Yalta and Yalta 2021; Njangang and Nvuh-Njoya 2023; Nguea 2024; Ndoya and Bakouan 2023; Oumbé et al. 2023; Mayer et al. 2023; Ketu and Ningaye 2024; Kamguia et al. 2023; Inoua 2023; Ketu et al. 2022; Hoang et al. 2023; Chu 2023; Camargo and Gala 2017; Avom et al. 2022; Avom and Ndoya 2024; Arpaci-Ayhan 2023; Ajide 2022). This signals the existence of the resource curse syndrome. It implies that overreliance on resource wealth kills innovative initiatives and the drive to build technological innovation and knowledge-based productive capacities to manufacture a wide range of sophisticated exports. It equally highlights that resource richness breeds complacency and deep corruption. These phenomena cause the neglect of initiatives that could enhance knowledge intensity and technological innovation to produce sophisticated, high-quality and internationally competitive exports. Meanwhile, one major drawback of the existing research is that it overlooks asymmetric and nonlinear features in the sensitivities of economic complexity to changes in natural resource wealth. The data distribution of natural resource rents often exhibits nonlinear and asymmetric features that align more with socioeconomic realities and real-world fundamentals. This study adds novelty in this regard to the global discussion, enhancing the understanding of the nexus between natural resource wealth and economic complexity.

The study differs from previous studies by examining nonlinear and asymmetric effects of natural resource rents on economic complexity within a nonlinear autoregressive distributed lag estimator proposed by Shin et al. (2014). All existing studies assume symmetric and linear impacts of natural resource rents on economic complexity. Recent advances in empirical and econometric analysis show the impracticality of linearity’s and symmetry’s oversimplified assumptions (Olaniyi et al. 2023a; Olaniyi and Odhiambo 2023b, 2024b; Olayeni et al. 2021; Hatemi-J and El-Khatib 2020; Olaniyi and Ologundudu. 2022; Olaniyi and Olayeni 2020; Olaniyi 2020, 2019; Xu 2018a, b). The symmetric and linear assumptions underlying the natural resource rent-economic complexity nexus in existing research do not explain the potential hidden information and issues in natural resource rent dynamics and management. Thus, symmetric approaches provide restrictive information and limited or rigid policy options. Using linear methods may not clearly explain certain real-world socioeconomic realities, such as unethical manipulations, racketeering, and complexities that often ensue in the management of natural resource incomes in resource-rich countries. Natural resource wealth management in most resource-rich countries involves shrewdness in many large-scale corruptions and political inferences that might be difficult to explain on the surface. As a result, a more sophisticated approach to unraveling hidden features and fundamentals might be appropriate for better practical policy options. Linear and symmetric approaches in the extant literature may be inadequate at capturing the resource curse phenomenon. Unlike the linear approach, the nonlinear method separates cumulative increases in natural resource rents from decreases. This process divulges hidden information and provides more flexible policy options to reveal resource-curse or blessing in the nexus. It, therefore, allows the differential sensitivities of economic complexity to positive and negative shock components in natural resource rents for better and richer research outcomes and policy perspectives. Thus, a nonlinear method explains the resource curse hypothesis much better by revealing the differential impacts of positive and negative components of natural resource rents on economic complexity. The asymmetric approach provides more explicit explanations and policy options on how economic complexity responds to the differential impacts of separate cumulative increases and decreases in natural resource rents. This study heads up and expands the knowledge space by examining the nonlinear and asymmetric effects of natural resource rents on economic complexity.

This study’s method of analysis has an edge over existing ones in the following ways: One, it relaxes the assumption of the monotony of economic complexity’s responses to natural resource rent changes prevailing in existing studies. Using a nonlinear approach, we capture economic complexity’s potential nonmonotonic response to natural resource rent changes. Two, it captures the nonlinear and asymmetric characteristics of the data (Jin and Xu 2024a, b; Olaniyi et al. 2023a; He 2020; Ikizlerli et al. 2019; Olaniyi and Ologundudu. 2022; Olaniyi and Odhiambo 2023b, 2024a) as well as the volatile nature of natural resource prices (Olayungbo 2019). Three, it gives the opportunity to examine the differential impacts of negative and positive shock components of natural resource rents on economic complexity. Unlike previous studies, which assumed limited and symmetric responses, this study allows economic complexity’s responses and sensitivities to positive and negative changes in natural resource rents to vary. Unlike previous studies, this helps to determine whether cumulative increases or decreases in natural resource rents are a blessing or a curse to economic complexity. Four, it unravels hidden information and allows more flexibility in policy dimensions (Olaniyi and Ologundudu 2022; Olaniyi 2020; Olaniyi and Olayeni 2020; Olaniyi and Odhiambo 2023b, 2024a, b). The asymmetric and nonlinear approach is more practical and insightful in explaining the resource curse phenomenon in the relationship between natural resource wealth and economic complexity. It helps to understand how economic complexity responds differently to cumulative increases in resource wealth compared to cumulative decreases in resource income. As a result, this approach provides a more insightful perspective for addressing the probable resource curse or blessing proposition in the natural resource rent-economic complexity nexus. Five, unlike symmetric methods in existing studies, it gives more insightful and socioeconomically driven information to guide policy directions. The separation of the impacts of increases in natural resource rents from declines in natural resource rents on economic complexity gives better policy perspectives on how to use natural resource wealth to spur economic complexity improvements. Nigeria’s data on economic complexity and natural resource wealth provide some fascinating features, facts, and intricacies reflect the main arguments of this study. The critical issues that challenge the efficient management of natural resource wealth and their probable contributions to Nigeria’s abysmal performance in the global ranking of economic complexity are significant. These issues highlight the need for Nigeria’s specific study to uncover hidden information and propose practical solutions.

There are robust structural deficiencies inherent in the technological capabilities and productive knowledge of Nigeria’s production system to produce globally competitive products for export. Nigeria consistently ranks at the bottom of the global ranking of economic complexity. Despite the rich endowment of natural resources and huge earnings from resource proceeds, Nigeria remains on the negative side of economic complexity rankings worldwide (Tabash et al. 2022; Olaniyi and Odhiambo 2023a). This situation reveals that Nigeria’s economy lacks productive knowledge and a sophisticated technology-based production system to produce quality exports. The average performances of Nigeria’s economic complexity over the years consistently indicate that the country has the most negative values among the African countries (Olaniyi and Odhiambo 2023a: -1.89 for the period 1995–2020; Tabash et al. 2022: -2.03 for the period 1995–2017; Ajide 2022, 2024: -1.91 for the period 1995–2018). These statistics indicate that Nigeria is the least sophisticated economy in Africa in terms of productivity knowledge and technical capacities embedded in the production system. In a more explicit explanation to portray Nigeria’s weak economic complexity, Olaniyi and Odhiambo (2023a) highlight that Nigeria has the most negative value in the economic complexity ranking in Africa (see Fig. 1). This statistic exhibits and reaffirms Nigeria’s position as the least sophisticated economy in Africa. Few studies have also shown that Nigeria ranks lowest in African and global rankings of economic complexity, despite the country’s richness in natural resources (Ajide 2022; Olaniyi and Odhiambo 2023a; Ajide and Osinubi 2024). It highlights Nigeria’s slow diffusion of advanced technology and productive knowledge in the production system to produce globally competitive products. This circumstance raises a question as to whether resource incomes are utilized to diversify the country’s productive base by investing in innovation, research and development, human capital development, and entrepreneurial inclinations to upgrade economic complexity. These phenomena suggest that Nigeria’s economic complexity is resource-cursed, and the weak institutional framework might have deepened and exacerbated the resource curse syndrome in the country.

Fig. 1

Average Performance of African Countries in Economic Complexity

Baseline and recent studies have consistently described Nigeria as a typical example of a resource-cursed country (Sala-i-Martin and Subramanian 2013; Tabash et al. 2022; Ajide 2022; Olayungbo 2019; Abdulahi et al. 2019; Amundsen 2017). Nigeria’s abundant natural resources have not translated into better performance in economic fundamentals and outcomes (Ondoa and Andela 2023), including economic complexity. The country is notorious for exhibiting large-scale corruption and mismanagement of natural resource proceeds (Abdulahi et al. 2019). The income from natural resources in Nigeria has engendered opportunistic behaviour and rent-seeking activities, which are instrumental in causing economic distortion (Transparency International 2020; Fagbemi and Kotey 2022). Research by Sala-i-Martin and Subramanian (2013) demonstrates the prevalence of excessive corruption, rent-seeking, and the resource curse syndrome in the management of natural resource rents in Nigeria. These phenomena make Nigeria a resource-cursed country. Most Nigerian technocrats, stakeholders, and politicians see natural resource rents as free money and manna from heaven. Thus, there is massive maladministration, widespread sharp practices, and corruption in the management of natural resources in Nigeria. The natural resource endowments breed endemic corruption and complacency. They impede innovative power, diversification of the economic base, technological abilities, and productivity knowledge needed to drive a sophisticated manufacturing process. Thus, converting natural resource blessings into finance and building up the technical capabilities and productive knowledge in the production system (economic complexity) to produce globally competitive products for exports is highly difficult in Nigeria’s case. Despite the puzzles and complexities surrounding the relationship between natural resource wealth and economic complexity, there have been no specific empirical studies on Nigeria to uncover the critical role of natural resource wealth in economic complexity and the relevant issues for comprehensive economic policies.

Nigerian data on natural resource wealth and economic complexity reveal robust nonlinear features and asymmetric structures in data spread and distribution (see Fig. 2). Nigeria’s data decompositions clearly show positive and negative components. Thus, assuming symmetry and nonlinearity is impractical. Unlike existing research that assumes an impractical, cosmetic, and oversimplified supposition of linearity and symmetry in the natural resource wealth-economic complexity nexus, this study captures the asymmetric and nonlinear characteristics of the data and delivers more valid and robust empirical estimates and enhanced policy options. To advance the current state of knowledge and global discussion, this study adds novelty by using Nigerian data to examine the asymmetric and nonlinear effects of natural resource wealth on economic complexity.

Compared to earlier studies, this study introduces novelties to enrich existing research by examining the asymmetric and nonlinear features in the nexus between natural resource wealth and economic complexity. Asymmetric and nonlinear characteristics reveal hidden information, provide more flexible policy options, and highlight more robust and pragmatic ways to unravel the resource curse syndrome regarding economic complexity. Also, this study is the first to examine the role of natural resource rents in economic complexity in Nigeria. This study becomes more important as the country suffers from the resource curse phenomenon and has the least sophisticated economy in Africa.

The remaining aspects of the paper are arranged as follows: Section 2 discusses theoretical and empirical literature. Section 3 focuses on the data description, sources, and empirical strategies adopted in the study. Section 4 presents preliminary analyses, empirical findings, and discussions, while Section. 5 highlights the study’s practical novelties and policy implications. Section 6 provides the study’s summary and conclusion. The study concludes with the study’s limitations and suggestions for future research.

Fig. 2

Actual data of natural resource rent and economic complexity index and their respective positive ($\:{NRR}^{+}\&\:{ECI}^{+}$) and negative ($\:{NRR}^{-}\&\:{ECI}^{-}$) components

Theoretical and empirical literature

Theoretical insights

The idea of economic complexity (ECI), first introduced by Hidalgo and Hausmann (2009), has become well known in the economic literature as a comprehensive indicator of economic development. ECI measures a nation’s technological innovation and knowledge-based productive capacities to produce a range of goods that are competitive on a worldwide scale. This measure takes into account components such as productivity knowledge, technological innovation, and knowledge intensity in manufacturing. As an essential component of an economy, some scholars have studied many factors that contribute to advances in ECI, with varying findings. A strand of literature highlights natural resource wealth as a crucial factor that explains the extent of a country’s technological innovation and knowledge-based productive capabilities to produce a range of globally competitive exports. The stream of income accumulated from natural resource endowments could provide leverage to engender diversification into diverse innovative sectors. It also helps secure funding to invest in programmes and initiatives that enhance knowledge intensity and technological innovation in the production system. Natural resource income could be useful in crystallizing investment in economic complexity-enhancing components, like the development of human capital, research and development, high-powered infrastructure, innovation and technology, entrepreneurial inclinations, and other initiatives that are instrumental in the diffusion of sophisticated production techniques. Through these processes, resource income promotes advancement in technological innovation and knowledge-based productive capacities embedded in an economy’s productive base. This facilitates the production of a wide range of sophisticated and globally competitive exports.

Meanwhile, most resource-rich countries get trapped in low-technology syndrome and resource curse phenomenon. These countries become complacent and lack the drive to diversify into innovative sectors and invest in initiatives that could stimulate economic complexity due to continuous flows of resource income. The abundance of natural resources pushes the stakeholders and politicians to become complacent and neglect serious pursuits in investing in components and initiatives that enhance technological innovation, knowledge-based productive capacities, and entrepreneurial prospects (Papyrakis and Gerlagh 2004; Ajide 2022). Because of these obstacles, it is difficult to channel resource wealth to fund technological advancements, knowledge-based productive capacity development, innovation, research and development support, entrepreneurship encouragement, human capital enhancement, and other necessary initiatives to help advance manufacturing sophistication for producing a range of exports that are capable of competing on a global scale.

The natural resource curse theory serves as the theoretical foundation for this study. According to this theory, an economy’s overreliance on resource wealth could breed complacency and cause stakeholders and policymakers to neglect investment in diversification, forward-looking innovation, technological sophistication, and knowledge-based productive capabilities to produce ubiquitous and high-quality exports (Mayer et al. 2023). Hence, a country with an abundance of natural resources might have a lower ranking in economic complexity compared to a resource-scarce country (Li et al. 2024; Sachs and Warner 2001). The idea that resource income is free money encourages politicians and stakeholders to engage in rent-seeking activities and spend natural resource revenues on unproductive projects. This stifles the development of sophisticated technologies, forward-looking innovation, and productive knowledge, which in turn mitigates the complexity of the economy’s manufacturing base and makes it more difficult to produce quality and competitive exports. This phenomenon implies that the abundance of natural resources prevents economic complexity upgrades. This signals the presence of resource curse syndrome in the natural resource wealth-economic complexity nexus. The present study introduces asymmetry and nonlinearity into the nexus by enriching the theoretical standpoint. The resource curse theory shows a blind spot to the likelihood of asymmetry and nonlinearity. The socioeconomic realities of asymmetries and nonlinearities are too obvious to deny in the dynamics and administration of natural resource wealth. Natural resource wealth management in most resource-rich countries involves intricacies, hidden information, racketeering, rent-seeking, opportunism, political interference, and sharp practices. Thus, a more advanced estimator is needed to unravel these intricate fundamentals that might be difficult for linear and symmetric approaches in existing research to address. Also, natural resource wealth data often manifest deep features of asymmetry and nonlinearity that could distort empirical analysis and policy implications if left unaccounted for.

Empirical perspectives

The body of research on how the abundance of natural resources either fosters or inhibits economic complexity is still in its nascent stages. In most countries with an abundance of natural resources, the majority of research confirm the existence of the resource curse phenomena in the relationship between natural resource wealth and economic complexity. In a sample of 32 African countries, Ajide (2022) uses a generalized method of moments (GMM) estimator and finds that total natural resource rents impede economic complexity. These research outputs suggest that African countries use their income from natural resources to finance unproductive initiatives and crude means of production. This hinders their ability to promote technological innovation, knowledge-based productive capacities, and sophisticated methods of production. These elements are necessary for creating complex manufacturing systems that can support the production of goods and exports that are competitive on a global scale. Therefore, Africa’s natural resource wealth acts as a drag that weakens upgrades to their Economic Complexity Index (ECI). These findings align with the research outcomes of Avom et al. (2022), which also used the system GMM to analyze 108 countries, covering the period 1995–2017.

These conclusions align with the study by Mayer et al. (2023) in Iraq, which uses estimators such as Canonical cointegration regression, dynamic ordinary least squares, and fully modified least squares. These estimators consistently affirm that natural resource wealth reduces the extent of technological innovation and knowledge intensity needed to produce sophisticated and internationally competitive exports. Njangang and Nvuh-Njoya (2023) also established that natural resource rents had an adverse effect on economic complexity in a sample of 106 countries. These findings imply that rather than offering stimulus that could effectively support ECI, countries endowed with natural resources but lacking robust institutions, low R&D investment, and low-quality human capital are more likely to divert their resource wealth to finance less productive activities. Yalta and Yalta (2021) carried out a panel analysis of Middle East and North Africa (MENA) countries from 1970 to 2015 using the generalized method of moments (GMM). The findings from their study demonstrate that natural resource rents have a diminishing effect on ECI. The findings indicate that when a country has an abundance of natural resources, it tends to get complacent and give rent-seeking priority over diversification into productive sectors that support technical prowess, innovation, creative thinking, and entrepreneurial endeavours—all of which are ways of enhancing ECI. These findings follow past research, which attest to the existence of “resource curse” phenomenon. It suggests that the endowment of natural resources with ineffective institutions may encourage corruption and ineffective management of resource income. These human-induced misfits lead stakeholders and policymakers placing high premium on unproductive activities over initiatives and innovative ventures that may increase ECI. Meanwhile, a study by Zhang et al. (2023) reported mixed findings in a sample of 116 countries using a spatial regression approach. Their findings for total natural resource endowment are in line with the resource curse theory, demonstrating that the total endowment of natural resources has a detrimental effect on ECI. It is likewise proven for rents from oil and gas, but not for rents from coal and minerals, which increase ECI. These findings provide a novel perspective on the diverse ways that different types of natural resource rents can affect ECI.

Another pertinent study by Hoang et al. (2023) indicates that total natural resource endowment constitutes a drag that threatens upgrades in ECI in the empirical analysis of 19 emerging and developed countries. The study adopted estimators like feasible generalized least squares (FGLS), panel-corrected standard error (PCSE), and a generalized quantile regression approach. Similar outputs by Oumbé et al. (2023) suggest that natural resource rents are impediments to an improvement in ECI. The study adopted estimators like FMOLS and DOLS to analyze a panel dataset of 112 countries. Furthermore, a study on 45 African countries between 2003 and 2016 by Kamguia et al. (2023) uses the system GMM estimator to account for endogeneity problems. The study also supports the idea that natural resource rents have a detrimental impact on economic sophistication. These results are in line with a study by Owjimehr and Jamshidi (2024), which used a pooled mean group (PMG) in a sample of 20 resource-rich nations covering the years 2000–2020 and established that natural resource income had a negative impact on ECI. Arpaci-Ayhan (2023) conducts another empirical effort that confirms a decreasing effect of natural resource wealth on economic complexity. The study uses two-stage least squares to analyze a panel analysis of 86 countries.

Another notable contribution by Ketu and Ningaye (2024) uses system GMM to scrutinize the panel dataset of 27 African countries from 1996 to 2017. The study establishes that natural resource endowments constitute an impediment that hinders ECI upgrades in African countries. Avom and Ndoya (2024) adopts the same estimator to empirically analyse dataset 118 countries and also finds evidence in support of resource curse theory that natural resource wealth weakens the drive to improve economic complexity. Utilizing similar estimator, Ndoya and Bakouan (2023) to examine the nexus between natural resource rents and ECI using the sample of 27 African countries for the period 1995–2018. The study affirms that natural resource wealth reduces economic complexity, signalling the existence of resource curse syndrome regarding economic complexity in Africa’s continent. An increase in natural resource rent is also established as a factor that inhibits economic complexity upgrades by Chu (2023). Nguea (2024) used Driscoll-Kraay regression and the GMM estimator to analyze 27 African countries from 1996 to 2017 and discovered a similar outcome. The study discovered that natural resource rents had a diminishing impact on economic complexity. This research outcome affirms the existence of resource curse phenomena among African countries. Rather than diversifying into innovative sectors of their economies and making investments in initiatives that could enhance technological innovation and knowledge-based productive capacities, these countries depend only on their natural resource endowments to survive. The results of a more recent study by Valentine et al. (2024), which used a system GMM estimator to analyse data for a sample of 100 countries over the years 1995–2020, also support the idea that natural resources slow down the pace of economic complexity upgrade. It implies that natural resource endowment reduces ECI, translating the curse of natural resources into economic complexity.

The highlights from the literature reviews predominantly confirm that natural resource wealth reduces economic complexity in most resource-rich countries. This signals the curse of natural resource endowment, which translates to backward economic complexity. It suggests that resource-poor countries have more robust technological innovation and knowledge-based productive capacities than resource-rich countries. Some countries with abundant resources are exceptions because they have sophisticated manufacturing structures to produce high-quality and globally competitive exports. This set of circumstances raises questions about why certain countries with abundant resources have more sophisticated knowledge-based production and technological capacities in their productive bases than do others. Factors like strong or weak institutions and prudent or poor management of resource wealth could incentivize or disincentivize channelling resource income to initiatives that enhance economic complexity. The synopses of theoretical and empirical literature on the effect of natural resource wealth on economic complexity reveal some fascinating facts and intriguing highlights. All existing research assume symmetry and linearity in the sensitivity of economic complexity to changes in natural resource rents. Meanwhile, recent advances in econometrics and empirical research have raised issues against the supposition of linearity and symmetric structure in financial and economic variables. The assumptions do not align with real-world fundamentals and socioeconomic realities, which emphasizes pragmatism of asymmetric structures and nonlinearities. We introduce asymmetric and nonlinear features into the natural resource wealth-economic complexity nexus to give a better explanation of the resource curse phenomenon. It gives the opportunity to examine the differential impacts of cumulative increases in natural resource wealth on economic complexity compared to cumulative decreases in resource incomes. Interestingly, no research on Nigeria has investigated how rich natural resources contribute to economic complexity. Despite the nation’s reliance on natural resources and weak economic complexity, the only study that examined factors influencing economic complexity (Osinubi et al. 2024) excluded natural resource endowments.

Methodological strategies and data description

Data sources and descriptions

This study uses Nigerian annual dataset for period 1984–2021. Data availability determines the study’s scope. This study gleans data on variables such as GDP per capita (constant 2015 US$), $\:gdpp$, total natural resources rents (% of GDP), $\:nrr$, from the World Bank database called World Development Indicator (WDI). We obtain data on a broad measure of financial development from the International Monetary Fund’s (IMF) financial statistics database. This metric offers comprehensive information and statistics that capture the accessibility, depth, and efficiency of financial institutions and markets.

This study gleans the data on institutional quality from ICRG (International Country Risk Guide). We use control of corruption as a measure of institutional quality because it is more strategic to the efficient management of natural resource wealth and its channelization to components and initiatives that enhance the amount of technological innovation and knowledge-based productive capacities to manufacture sophisticated, high-quality and globally competitive exports. Besides, Nigeria’s performance in controlling corruption during the study period was abysmally weak. The average performance of corruption control as an institutional quality measure attests to Nigeria’s woeful institutional development, as the country scores 1.59 on an ordinal scale of 0–6. This represents a 26.5% average performance. This score could give more information on the extent of Nigeria’s corruption control to curtail manipulative tendencies and corrupt practices in natural resource rent management. Hence, this study places a high premium on corruption control as a major measure of institutional quality in Nigeria. Furthermore, this study uses data on the economic complexity index from the MIT Media Lab’s Observatory of Economic Complexity (https://oec.world/en/rankings/eci/hs6/hs96). The data rely on international trade statistics, which link nations to the quantity and quality of their exports. Following the explanation of Hausmann and Hidalgo (2011), ECI gives information beyond the technological capabilities and knowledge-based productivity in the production base; it also provides growth-related data on the economy’s education, institutions, competitiveness, and other development-related factors.

Modeling strategies and estimators’ procedures

This study follows the modeling procedures in studies that examine the impact of natural resource wealth on economic complexity (Avom and Ndoya 2024; Nguea 2024; Valentine et al. 2024; Ketu and Ningaye 2024; Owjimehr and Jamshidi 2024; Arpaci-Ayhan 2023; Chu 2023; Mayer et al. 2023; Inoua 2023; Hoang et al. 2023; Zhang et al. 2023; Oumbé et al. 2023; Kamguia et al. 2023; Avom et al. 2022; Ketu et al. 2022; Ndoya and Bakouan 2023; Yalta and Yalta 2021; Ajide 2022; Camargo and Gala 2017; Njangang and Nvuh-Njoya 2023). We augment the model with all relevant variables to align with existing research and incorporate the study’s peculiarities and novelty.

$$\:{eci}_{t}=\:{\varnothing\:}_{0}+{\varnothing\:}_{1}{nrr}_{t}+{\varnothing\:}_{2}{fd}_{t}{\gamma\:}_{2}+{\varnothing\:}_{3}{gdpp}_{t}+{\varnothing\:}_{4}{inst}_{t}+{\epsilon\:}_{it}$$

(1)

where $\:nrr,\:fd,\:gdpp,\:and\:inst$ are total natural resource rents, financial development, real GDP per capita, and institutional quality. The parameters $\:{\varnothing\:}_{1},\dots\:,\:{\varnothing\:}_{4}$ are long-run estimates of the respective variables while $\:{\varnothing\:}_{0}$ is the shift parameter. If the long-run coefficients are jointly significant, it implies the existence of long-run relationship. Following the influential Pesaran et al. (2001), unrestricted variant error-version of autoregressive distributed lag (ARDL) is presented in Eq. (2). This study explains the long-run coefficients in an error-correction model (ECM) version consistent with the work of Bahmani-Oskooee et al. (2020). This explains how shifts in economic complexity relate to previous periods of divergence and disequilibrium and any shifts brought about by adjustments to other regressors. Equation (2) explains these processes as follows:

$$\:{\varDelta\:eci}_{t}={\partial\:}_{t}+\sum\limits_{i=1}^{{\gamma\:}_{1}}{\omega\:}_{i}\varDelta\:{eci}_{t-i}+\sum\limits_{i=0}^{{\gamma\:}_{2\:}}{\theta\:}_{i}\varDelta\:{nrr}_{t-i}+\sum\limits_{i=0}^{{\gamma\:}_{3}}{\mu\:}_{i}\varDelta\:{fd}_{t-i}+\sum\limits_{i=0}^{{\gamma\:}_{4}}{\pi\:}_{i}\varDelta\:{gdpp}_{t-i}+\sum\limits_{i=0}^{{\gamma\:}_{5}}{\mu\:}_{i}\varDelta\:{inst}_{t-i}+{\forall\:}_{0}{eci}_{t-1}+{\forall\:}_{1}{nrr}_{t-1}+{\forall\:}_{2}{fd}_{t-1}+{\forall\:}_{3}{gdpp}_{t-1}+{\forall\:}_{4}{inst}_{t-1}+{\epsilon}_{t}$$

(2)

A bounds’ testing approach to cointegration was proposed by Pesaran et al. (2001) incorporates both long- and short-run effects simultaneously. The coefficients of first-differenced variables are short-run effects, while the parameters $\:{\forall\:}_{2},\dots\:,\:{\forall\:}_{4}$ normalized by $\:{\forall\:}_{1}$ explain the long-run components of Eq. (2) (Bahmani-Oskooee et al. 2020; Ullah et al. 2020; Olaniyi et al. 2023a; Olaniyi and Oladeji 2022). We present the normalization process as follows:

$$\:{\varnothing\:}_{1}=\frac{{\forall\:}_{1}}{{-\forall\:}_{0}},\:\:\:{\varnothing\:}_{2}=\frac{{\forall\:}_{2}}{{-\forall\:}_{0}},\:\:{\varnothing\:}_{3}=\frac{{\forall\:}_{3}}{{-\forall\:}_{0}},\:\:{\varnothing\:}_{4}=\frac{{\forall\:}_{4}}{{-\forall\:}_{0}}$$

(3)

The existence of long-run relationship among the variables depends on the joint significance of the long-run coefficients in Eq. (2). This process generates $\:F-statistic$. Using Pesaran et al.‘s (2001) method, we compare the calculated value of the F-statistic to the two asymptotic critical value bounds (upper and lower bounds). Lower bounds rely on the assumption that all the variables are stationary at level, $\:\left[I\left(0\right)\right]$, while upper bounds require that all the variables are stationary at the first difference, $\:\left[I\left(1\right)\right]$ (Olaniyi et al. 2023a). Cointegration among the variables exists only if the calculated F-statistic is above the upper critical values, regardless of the orders of integration, either $\:\left[I\left(0\right)\right]$ or $\:\left[I\left(1\right)\right]$. The approach does not require a pre-test of the unit root. Meanwhile, the approach breaks down in the presence of variables that are stationary at the second difference, $\:\left[I\left(2\right)\right]$.

Following the novelty discussed earlier in the study, all existing research that examines the effects of natural resource rents on economic complexity assumes linearity and symmetry. Prevalently symmetric and linear approaches are unrealistic for explaining certain fundamentals and revealing real-world phenomena and socioeconomic realities. The common dynamism of economic and financial time series, like natural resource rents, often exhibits asymmetric and nonlinear characteristics. As a result, this study extends the knowledge space by incorporating asymmetric and nonlinear features. We adopt a nonlinear ARDL recently developed by Shin et al. (2014). This method’s distinctiveness makes it easier to investigate how diverse natural resource rent shocks—positive and negative—affect economic complexity. It provides a novel perspective to explain the resource curse phenomenon with regard to economic complexity. In alignment with existing studies such as Hatemi-J (2012), Olaniyi (2019), (2020); Olaniyi and Olayeni (2020); Olaniyi and Odhiambo 2023b, 2024a), this study adopts cumulative partial sum approach to separate changes in natural resource rents, $\:nrr,$ into positive changes, $\:{nrr}^{+}$ (cumulative increases in natural resource rents) and negative changes, $\:{nrr}^{-}$ (cumulative decreases in natural resource rents). We define the two components as follows:

$$\:nrr=\:{nrr}_{0}+{nrr}_{t}^{+}+{nrr}_{t}^{-}$$

(4)

where $\:{nrr}_{t}^{-}$ and $\:{nrr}_{t}^{+}$ stand for partial cumulative sums of negative and positive changes in $\:nrr$, respectively. This study constructs these two components as follows:

$$\:{nrr}_{t}^{+}=\sum\limits_{j=1}^{t}\varDelta\:{nrr}_{j}^{+}=\:\sum\limits_{j=1}^{t}\text{max}\left(\varDelta\:{nrr}_{j},\:0\right)$$

(5)

$$\:{nrr}_{t}^{-}=\sum\limits_{j=1}^{t}\varDelta\:{nrr}_{j}^{-}=\:\sum\limits_{j=1}^{t}\text{min}\left(\varDelta\:{nrr}_{j},\:0\right)$$

(6)

To explore the nonlinear and asymmetric effects of natural resource rents on economic complexity, the constructed components, $\:{nrr}^{+}and\:{nrr}^{-},\:$of $\:nrr$ replace natural resource rents in Eq. (2). In line with these adjustments, we respecify Eq. (2) as follows

$$\:{\varDelta\:eci}_{t}={a}_{t}+\sum\limits_{i=1}^{{\gamma\:}_{1}}{b}_{i}\varDelta\:{eci}_{t-i}+\sum\limits_{i=0}^{{\gamma\:}_{2\:}}{c}_{i}\varDelta\:{nrr}_{t-i}^{+}+\sum\limits_{i=0}^{{\gamma\:}_{3\:}}{d}_{i}\varDelta\:{nrr}_{t-i}^{-}+\sum\limits_{i=0}^{{\gamma\:}_{4}}{e}_{i}\varDelta\:{fd}_{t-i}+\sum\limits_{i=0}^{{\gamma\:}_{5}}{f}_{i}\varDelta\:{gdpp}_{t-i}+\sum\limits_{i=0}^{{\gamma\:}_{6}}{g}_{i}\varDelta\:{inst}_{t-i}+{\forall\:}_{0}{eci}_{t-1}+{\forall\:}_{1}{nrr}_{t-1}^{+}+{\forall\:}_{2}{nrr}_{t-1}^{-}+{\forall\:}_{3}{fd}_{t-1}+{\forall\:}_{4}{gdpp}_{t-1}+{\forall\:}_{5}{inst}_{t-1}+{\epsilon}_{t}$$

(7)

To create a nonlinear ARDL, we augment Eq. (7) with partial cumulative sums of $\:nrr$. (Alsamara, Mrabet & Hatemi-J 2020; Olaniyi et al. 2023a). This nonlinear model allows economic complexity to respond asymmetrically to changes in natural resource rents, either in short-run and long-run dynamics or both. For both the linear model in Eq. (2) and the nonlinear model in Eq. (7), the bounds testing method of Pesaran et al. (2001) works well for assessing cointegration. For the critical values to remain at high and conservative levels, both positive and negative changes $\:\left({nrr}^{+},{nrr}^{-}\right)$ must be considered as one variable, despite the fact that the number of variables in Eq. (7) is one more than those in Eq. (2) (Shin et al. 2014; Bahmani-Oskooee et al., 2020; Olaniyi et al. 2023a). Empirical analysis of Eq. (7) allows testing the presence of asymmetry in the short and long run. One, if $\:{\gamma\:}_{2}\ne\:{\gamma\:}_{3}$, it implies that both positive and negative changes $\:\left({nrr}^{+}and\ {nrr}^{-}\right)$ take different lag length orders and it signifies asymmetry. This shows that economic complexity responds differently to positive changes in natural resource rents $\:\left({nrr}^{+}\right)$ compared to negative changes $\:\left({nrr}^{-}\right)$. Two, given a specific lag order $\:i$, we confirm short-run asymmetry if the estimates of $\:{c}_{i}$ and $\:{d}_{i}$ are significantly different. It suggests there is a confirmed short-run asymmetry in the sensitivity of economic complexity to changes in natural resource rents. Testing and rejecting the null hypothesis, $\:\sum\:{c}_{i}=\sum\:{d}_{i}$, via the Wald test could also confirm the validity of short-run asymmetry. Also, by testing the null hypothesis, $\:\frac{{\forall\:}_{1}}{-{\forall\:}_{0}}=\frac{{\forall\:}_{2}}{{-\forall\:}_{0}}$, we apply the Wald test to assess the long-run asymmetric impact of natural resource rents on economic complexity.

Furthermore, following the nonlinear ARDL model explained, we extract two cumulative dynamic multipliers $\:\left({m}_{k}^{+}\:\text{a}\text{n}\text{d}\:{m}_{k}^{-}\right)$. This process is helpful in determining the dynamic nonlinear adjustments to a new equilibrium after a shock. Therefore, we evaluate the cumulative dynamic multiplier effects. This evaluation accounts for the impact of a 1% change in $\:{nrr}^{+}$ and $\:{nrr}^{-}$, respectively, on economic complexity as follows:

$$\:{m}_{k}^{+}=\sum\limits_{j=0}^{k}\frac{\partial\:{eci}_{t+j}}{\partial\:{nrr}_{t}^{+}},\:{m}_{k}^{-}=\sum\limits_{j=0}^{k}\frac{\partial\:{eci}_{t+j}}{\partial\:{nrr}_{t}^{-}},\:\:k=0,\:1,\:2,\dots\:$$

By construction, the system follows that as $\:k\to\:\infty\:,\:\:{m}_{k}^{+}\to\:\:\frac{{\forall\:}_{1}}{-{\forall\:}_{0}},\:\text{a}\text{n}\text{d}\:{m}_{k}^{-}\to\:\frac{{\forall\:}_{2}}{{-\forall\:}_{0}}\:\:$

This study further uses a fully modified ordinary least squares (FMOLS) estimator as a robustness check. Several studies have identified the effectiveness of the FMOLS estimator in addressing potential econometric issues such as endogeneity and serial correlation (Pedroni 2001; Sadorsky 2009; Bhattacharya et al. 2016; Rahman et al. 2020; Olaniyi and Oladeji 2021, 2022; Mahendru et al. 2023; Bashir et al. 2023, 2024; Ma et al. 2023). This approach is better equipped to handle endogeneity problems when faced with stationarity properties of higher order, as well as long-run relationships among variables, compared to the generalized method of moments (GMM) (Olaniyi and Odhiambo 2024b). Additionally, long-run estimates are more reliable and provide a more comprehensive understanding of policy perspectives and global views (Olaniyi and Adedokun 2022). The FMOLS estimator is also more suitable for cases with small sample sizes (Rahman et al. 2020; Pedroni 2001; Phillips and Hansen 1990).

The study’s discussion of findings

Exposition of descriptive statistics

This study examines descriptive statistics to unravel the characteristics and behaviour of the data distribution of variables to guide the choice of estimation techniques and analytical methods. In Table 1, this study presents the detailed results of descriptive statistics. The values’ comparisons of the mean and standard deviation of all the variables reveal that actual data on variables such as total natural resource rents $\:\left(nrr\right)$, institutional quality $\:\left(inst\right)$, and financial development $\:\left(fd\right)$ spread out from their mean values. The data distribution of these variables questions the reliability of their average values as a mirror of the actual dataset and their distribution (Olaniyi 2022, 2023; Olaniyi et al. 2023a). Meanwhile, data on the economic complexity index $\:\left(eci\right)$ and natural logarithms of real GDP per capita $\:\left(lgdpp\right)$ cluster around their respective mean values to a reasonable extent. It implies that the two variables have mean values that relatively reflect the features of the actual data spread and distribution. These findings are consistent with the coefficients of variation. The skewness coefficients reveal that only economic complexity and institutional quality skew negatively, while the remaining variables skew positively. The kurtosis coefficients reveal the extent of outliers in the data distribution. The economic complexity index and total natural resource rents are leptokurtic, as the coefficients are marginally above 3. It implies evidence of a fatter or thicker tail with a slim chance of mild outliers in the data distribution. Other variables are platykurtic, highlighting a light tail with the likelihood of insignificant outliers in the data distribution. These findings imply that mesokurtic data distribution does not exist in the dataset. This suggests the existence of outliers, but a severe outlier does not exist in any of the variables. All variables are normally distributed, with the exception of the natural logarithms of real GDP per capita $\:\left(lgdpp\right)$, according to the coefficients of the Jarque-Bera statistic. Descriptive statistical analyses show asymmetric features in the variable data distribution. This signals the necessity of adopting an asymmetric and nonlinear approach to reconcile the pragmatic features of real-world phenomena with socioeconomic realities (Shahbaz et al. 2017; Olaniyi 2019, 2020; Olaniyi and Olayeni 2020; Olaniyi et al. 2023a; Olaniyi and Odhiambo 2023b, 2024a).

Table 1 Descriptive statistics

Asymmetric and nonlinear features in the trend of natural resource rents

The data decomposition of natural resource rents into positive and negative change components clearly exhibits asymmetric and nonlinear features. As shown in Fig. 2, the graphical representations of the negative and positive shock components of Nigeria’s natural resource wealth confirm the asymmetric structures. This supports our earlier argument that the natural resource wealth-economic complexity nexus should take into account asymmetric and nonlinear characteristics. This validates the argument that there might be asymmetric and nonlinear features in the sensitivity of economic complexity to changes in total natural resource rents. This approach reveals novel perspectives to explain resource curses or blessings related to economic complexity.

Unit root test

The ARDL estimator does not require confirming the unit root, but it is essential since the ARDL method fails when any variable becomes stationary at the second order of integration, $\:I\left(2\right)$. Existing research and recent advances in econometrics have questioned the appropriateness of traditional unit root tests. They tend to have low explanatory and predictive power in the presence of asymmetry and nonlinearity in data distribution. This may lead to invalid decisions when there are asymmetric and nonlinear features in the data distribution of variables (Olaniyi et al. 2023a; Otero and Smith 2017; Shahbaz et al. 2020). The data decomposition procedure reveals robust asymmetric and nonlinear features in financial development indicators and natural resource rents. The descriptive statistic also affirms an asymmetric structure in data distribution. Hence, this work uses the Kapetanios and Shin (2008) (henceforth, KSUR) unit root test to account for asymmetric structure and nonlinearity. Otero and Smith (2017) provide the Stata codes to execute the KSUR technique. The approach yields more reliable and robust critical values. Given the study’s argument for asymmetric and nonlinear effects on natural resource wealth on economic complexity, it might not be adequate to examine the variables’ stationarity properties using the linear and symmetric unit root approach (Olaniyi et al. 2023a; Dong et al. 2021).

This method has an advantage over the widely used linear, traditional method. It accounts for nonlinearities and asymmetries. KSUR is an exponential smooth transition autoregressive technique following the theoretical path of global stationarity. This makes use of a Monte Carlo simulation model that takes nonlinearity and asymmetry into account. To guarantee consistency in the outcomes and decisions, this study endogenously determines the lag length using the Akaike Information Criteria (AIC). We do not report the findings of other variants of lag length criteria, such as Schwarz Information Criteria (SIC), a data-dependent technique known as the General-to-Specific (GTS) algorithm of Hall (1994), applying 5% (GTS05) and 10% (GTS05) levels of significance. In Table 2, we present the results of the KSUR unit root tests for both intercept only and that of intercept and trend. The results reveal that financial development (FD) is stationary at level, I(0). Similarly, natural resource rents (nrr) highlight weak stationarity at level, I(0). Other variables become stationary at the first difference, I(1). These findings affirm the mixed order of integration. Collectively, all the variables are stationary at the first difference. These findings validate the adoption of nonlinear ARDL, which performs efficiently when the stationarity properties of variables follow the pattern of mixed order of integration. The findings equally fulfill the condition that the dependent variable, economic complexity, should be integrated of order one, I(1). The findings from the KSUR unit root test highlight that nonlinear ARDL is the perfect match for this study.

Table 2 Nonlinear unit root test (stationary nonlinear ESTAR model)

Cointegration test

After confirming mixed orders of integration using the KSUR unit root test, this study examines the possibility of cointegration among the variables using the ARDL bounds testing. We chose this cointegration test as preferable to the Johansen procedure because it performs better when stationarity properties reveal mixed orders of integration zero and one (Olaniyi and Oladeji 2022; Olaniyi and Adedokun 2022). The cointegration test becomes necessary because the unit root test’s results suggest that the variables behave divergently in the short run (Xu 2017). Table 3 presents the results of linear and nonlinear cointegration. Both nonlinear and linear cointegration tests confirm the existence of a long-run relationship. These decisions are premised on the fact that calculated F-statistics are above the upper and lower bounds’ critical values. This signals a long-term relationship. While the KSUR unit root results confirm distortions and divergence among the variables in the short run, resulting in disequilibrium, The cointegration, on the contrary, implies that the variables converge along long-run dynamics. These results highlight that long-run estimates are more reliable and predictable than short-run estimates. In the sequel to the confirmation of a long-run relationship in the nonlinear and linear models, the results of the long- and short-run dynamics of the error-correction variant of the ARDL estimator to explain the effects of dependent variables on economic complexity in Nigeria are presented and interpreted in the subsequent sections.

Table 3 Linear and nonlinear ARDL bounds test

The role of natural resource rents in economic complexity in Nigeria

The asymmetric effect of natural resource rents on economic complexity

In line with the main study’s focus, we examine the existence of asymmetric features in the natural resource rent-economic complexity nexus in the short and long run. The study uses the dynamic multiplier graph and the Wald coefficient diagnostic test to examine the presence of asymmetry. These results are presented in Table 4; Fig. 3. The results in Table 4 highlight that there are short- and long-run asymmetries in the sensitivity of economic complexity to changes in natural resource wealth. To further validate the robustness of the results, we use the fully modified least squares (FMOLS) estimator and the Wald coefficients diagnostic test. The FMOLS estimator also confirms the existence of long-run asymmetry. This implies that negative and positive changes in natural resource rents have differential effects on economic complexity. Contrary to the assumption of symmetry in existing research, these findings highlight significant evidence of the asymmetric and nonlinear effect of natural resource wealth on economic complexity in Nigeria. These research outcomes invalidate the restrictive suppositions of linearity and symmetry in the existing studies. All previously published works that investigated the effect of natural resource wealth on economic complexity assumed symmetry and linearity. All existing studies neglect asymmetry and nonlinearity. These might have led to overestimation of their models, exaggeration of findings, and impaired policy implications. These are inconsistent with real-world phenomena and socioeconomic fundamentals surrounding natural resource rent management and its implications for economic complexity. Different from existing studies’ practices and perspectives, the confirmation of asymmetry implies that there should be policy variations explaining why cumulative increases in resource income and cumulative decreases in resource wealth have different effects on economic complexity. In natural resource wealth dynamics and management, the study’s innovative idea, which exhibits asymmetric phenomena, might provide novel perspectives to unravel hidden shrewd practices, rent-seeking activities, opportunistic behaviour, and racketeering. It also offers fresh insights into how to allocate resource wealth to enhance knowledge-based productive capacities and technological innovation to produce sophisticated and internationally competitive exports.

Table 4 Tests for asymmetries in the short and long-run nonlinear ARDL and FMOLS

Fig. 3

Nonlinear ARDL Dynamic Multiplier Graph

To further validate the robustness of the asymmetric effect, we use a nonlinear ARDL dynamic multiplier graph as presented in Fig. 3. The dynamic multiplier graph attests to clear evidence of an asymmetric effect of natural resource rents on economic complexity, both in the short and long run. The asymmetry plot line confirms asymmetric structure, as it deviates from the line of no asymmetry. To further confirm asymmetry, the broken red line of asymmetry persists within the confidence interval (CI) line of 95%. Nigeria’s multiplier graphs of negative and positive changes in natural resource rents over the short and long run show consistent and significant differences. This highlights the fact that economic complexity in Nigeria is asymmetrically sensitive to negative and positive changes in natural resource rents. It implies that positive and negative shock components in natural resource rents have varying impacts on economic complexity in Nigeria. This finding supports the case for asymmetric structure and nonlinearity in the sensitivity of economic complexity to changes in natural resource wealth.

Long-and short-run dynamics of nonlinear and linear estimates

This subsection provides a detailed explanation of linear and nonlinear estimates for the error-correction variant of ARDL (autoregressive distributed lag) and a fully modified least squares (FMOLS) estimator as a robustness check. The ARDL estimator distinguishes between short-run estimations and long-run dynamics, whereas FMOLS accounts for the model’s endogeneity and serial correlation. We provide both short-run and long-run dynamics to distinguish between narrow views and global perspectives for unique and socioeconomically driven policy consequences. Tables 5 and 6 present the results of error-correction variants of linear and nonlinear ARDL. The coefficients of error correction terms (ect) in both nonlinear and linear ARDL are significant and negative. These coefficients are less than 1. The coefficients are correctly signed and within conventional ranges. These estimates reaffirm the existence of a long-run relationship among the variables in support of the linear and nonlinear ARDL bounds testing presented and explained earlier (Banerjee et al. 1998; Olaniyi and Oladeji 2022; Olaniyi and Adedokun 2022; Olaniyi et al. 2023a). These findings attest that variables are liable to converge after a series of distortions, shocks, divergences, and deviations on the pathways to long-run dynamism. These coefficients of error correction terms $\:\left(ect\right)$ give the percentage of divergences that are liable to be restored to equilibrium over a period of time, namely a year. These speeds of adjustment to equilibrium are 86.9 and 73.4% for nonlinear and linear ARDL models, respectively.

Table 5 Short and long-run estimates of linear ARDL (23341)

Table 6 Short and long-run estimates of nonlinear ARDL (244320)

The coefficients of lagged economic complexity (eci) in both linear and nonlinear ARDL are significant and positive at the conventional level. This implies that previous performance in economic complexity is a positive impulse that stimulate current upgrades in economic complexity in Nigeria. These results are consistent in both linear and nonlinear error-correction version of ARDL as reported in Tables 5 and 6. The implication is that efforts in previous years to promote economic complexity through concerted moves to build up R&D activities, technological innovation, and knowledge-based productive capacities result in enhancing current improvements in economic complexity. These findings specifically suggest that advancements in EC in the previous period serve as a motivator to facilitate an upgrade in EC in the current year (Chu 2020). It also implies that there should be persistent efforts and resolute determination to create policies and schemes that will improve EC from one period to another. It suggests that it will take persistence and tenacity for government agencies, parastatals, and Nigerians to develop instincts and commit to transitioning to sophisticated manufacturing quality and competitive exports. It also underlines that present EC advancements are catalysts for future EC improvements. Therefore, consistent efforts are necessary to enhance EC in Nigeria. These findings align with the studies of Gnangnon (2021); Nguyen and Su 2021a, b; Olaniyi and Odhiambo 2023a); Kamguia et al. 2022), 2023); Ndoya and Bakouan (2023); Atasoy (2021); Chu (2020).

The Wald diagnostic test shows that short-run coefficients of natural resource rents in the linear error-correction version of ARDL are combinedly positive but insignificant. On the contrary, the natural resource wealth coefficient is significant in the long run but negative at a conventional level of significance. It appears that natural resource rents do not have the immediate and sustained potency to enhance Nigeria’s technological innovation and knowledge-based productivity, thereby impairing its ability to produce high-quality and sophisticated exports that will compete globally. Meanwhile, natural resource wealth undermines Nigeria’s long-term economic complexity. This suggests that an increase in resource income causes stakeholders and policymakers to neglect the Nigerian economy’s diversification into productive and innovative sectors. It also breeds complacency, kills innovative tendencies and knowledge-based productive capacities, and discourages investment in economic complexity-enhancing components, such as R&D, technology and innovation, high-powered infrastructure, human capital development, quality education, entrepreneurial prospects, and other initiatives to build up a sophisticated production base in the long run. This implies that persistent flows of income from natural resource abundance cause stakeholders, politicians, and policymakers to lose focus on the need to leverage resource wealth to build up initiatives and innovation to develop sophisticated manufacturing and production bases to produce a wide range of competitive exports. The implication is that the natural resource curse phenomenon holds true regarding economic complexity in Nigeria. These findings confirm the prevalent outcomes of existing research, such as (Avom et al. 2022; Ajide 2022; Arpaci-Ayhan 2023; Kamguia et al. 2023; Oumbé et al. 2023; Hoang et al. 2023; Njangang and Nvuh-Njoya 2023; Inoua 2023; Mayer et al. 2023; Chu 2023; Avom and Ndoya 2024; Ketu et al. 2022; Yalta and Yalta 2021; Owjimehr and Jamshidi 2024; Ketu and Ningaye 2024; Nguea 2024; Valentine et al. 2024). Meanwhile, these studies assume the restricted suppositions of symmetry and linearity, which have low power of explaining real-world phenomena and socioeconomic realities in the natural resource wealth-economic complexity.

Hence, an error-correction version of the nonlinear ARDL is used in this study to examine nonlinearities and asymmetry between natural resource rent and economic complexity in the long-run and the short-run. A combined short-run coefficient of natural resource rents indicates that negative and positive changes in natural resource wealth have differential effects on economic complexity. All the short-run coefficients of the positive change component (positive shocks) in natural resource wealth are statistically significant. The sum of the coefficients is positive (0.011), and the Wald test also indicates they are jointly significant. This implies that cumulative increases in natural resource rents in Nigeria provide stimuli and incentives that promote economic complexity in the short run. These could be in the form of supporting investment in initiatives that stimulate technological innovation and knowledge-based productive capacities to produce sophisticated and globally competitive exports. These findings suggest that more increases in natural resource wealth could be used to fund immediate investments in human capital, education, R&D, modern innovation and technology, high-quality infrastructure, entrepreneurial inclinations, and other initiatives aimed at improving Nigeria’s economic complexity. Therefore, cumulative increases in resource income (positive shocks) promote Nigerian economic complexity, creating a short-run resource blessing syndrome in the nexus. Unlike the positive shock, the cumulative negative shock component in natural resource rents has negative coefficients, except for lag three. The sum of these coefficients is negative (-0.036), and the Wald test confirms their joint significance.

This suggests that cumulative dips in natural resource income cause immediate adverse effects on Nigeria’s economic complexity. This highlights that decreases in natural resource rents cause disincentives that reduce resource income allocation to activities and initiatives that enhance economic complexity. These research outcomes serve as strong signals to stakeholders, government agencies, and policymakers on the negative implications of reducing resource income allocated to economic complexity-enhancing schemes in Nigeria. It also highlights that the country expends resource incomes on activities and traditional methods of production at the expense of sophisticated technological innovation and knowledge-based productive capacities. The short-run differential impacts of positive and negative shocks in natural resource rents on economic complexity affirm the existence of asymmetry and nonlinearity. It also allows more flexible, deep, and insightful policies to address real-world and socioeconomic realities. It also explains the resource curse syndrome regarding economic complexity in unique ways that are more pragmatic, rather than oversimplifying approaches of symmetry and linearity in the existing studies. This innovatively reveals that cumulative increases in natural resource income do not reflect resource curses regarding economic complexity in the short run, but cumulative decreases result in resource curses.

Unlike short-run dynamics, the long-run component of the error-correction variant of the ARDL estimator indicates that total natural resource rents reduce economic complexity. These results are consistent for both linear and nonlinear ARDL variants. These findings provide robust evidence of the resource curse phenomenon regarding economic complexity upgrades in Nigeria. It implies that the long-run prospects of allocating resource incomes to promote investment in all the necessary ingredients that enhance economic complexity upgrades are non-existent. Natural resource abundance and associated income streams breed complacency. It makes stakeholders, politicians, and government parastatals blind to the need to use resource income to fund long-term diversification into creative and innovative industries. It also discourages the government and technocrats from investing extensively in human capital, entrepreneurial opportunities, technological innovation, high-quality infrastructure, R&D, and other efforts that may drive long-term and sustainable increases in Nigeria’s economic complexity. It is clear that resource revenues are being used to fund outmoded production processes, which undermines the ability of resource incomes to stimulate long-term and sustainable economic complexity advances. Nigeria’s natural resource rents encourage actions and initiatives that undermine long-term chances for increasing economic complexity.

An error-correction version of nonlinear ARDL also confirms these results. The long-run estimates of positive and negative shock components of natural resource wealth are statistically significant but negative. The FMOLS (fully modified least squares) estimator also confirms these findings (See Table 7). These research outcomes indicate that cumulative increases and decreases in natural resource rents mitigate upgrades in economic complexity in the long run. This is clear evidence that the resource curse phenomenon holds true regarding economic complexity in Nigeria. It implies that natural resource incomes are channelled to unproductive and less innovative activities. Hence, it reduces technological innovation and knowledge-based productive capabilities in the long run, thereby reducing economic complexity in Nigeria. Both positive and negative change components in natural resource rents (cumulative increases and decreases) impede long-run prospects and initiatives to promote economic complexity in Nigeria. Natural resource wealth is a long-term factor that constitutes a drag on economic complexity in Nigeria. Natural resource wealth, therefore, acts as a drag on the nation’s efforts to shift to an advanced manufacturing process in order to deepen technological innovation and knowledge-based productive capabilities that will eventually result in the production of high-quality and globally competitive export varieties. Nigeria’s management of its natural resource wealth requires a complete overhaul and scrutiny in order to unravel the innate shrewdness, rent-seeking, opportunism, and sharp practices in the system. This helps to explain why resource income set.

Table 7 The results of FMOLS estimates

aside for fostering the development of technological innovation and knowledge-based productive capabilities embedded in the nation’s productive structure reduces economic complexity. This study introduces a new dimension to enhance the understanding on the sensitivity of economic complexity to changes in natural resource rents. This study’s findings justify the need to account for nonlinearity and asymmetry in the sensitivity of economic complexity to natural resource wealth changes. It also increases the quality and flexibility of policy perspectives and insights. Having analysed and discussed this study’s main variables, we shift our attention to discussing other determinants of economic complexity and how they affect Nigeria’s economic complexity.

The Wald test coefficients’ diagnostic show that the coefficients of income measured by real GDP per capita $\:\left(lgdpp\right)$ are combinedly positive and significant in both the linear and nonlinear models in the short-run. The sums of the coefficients are positive (3.654 and 3.097, respectively). The long-run estimates also confirm these results in the linear and nonlinear models. These findings support the theoretical expectation, which regard income as a catalyst to develop and propel the moves that enhance an increase in technological innovation and knowledge-based productive capabilities implanted in the productive base to manufacture sophisticated and internationally competitive exports. It demonstrates that rising income levels help manufacturers become more technically innovative and knowledge-intensive, enabling them to manufacture high-quality and sophisticated products. Higher incomes further cause consumers’ preferences and tastes to change in favour of a wider range of more sophisticated goods and services (Olaniyi and Odhiambo 2023a). These research outcomes are consistent with the findings of the studies of Ajide 2022); Yalta and Yalta (2021); Arpaci-Ayhan (2023), Hoang et al. (2023), and Avom and Ndoya (2024). Conversely, it contradicts the research outcomes of Ndoya and Bakouan (2023); Olaniyi and Odhiambo 2023a); Chu (2020), which show that income mitigates economic complexity in African and high-income countries. From a policy perspective, these findings imply that Nigeria’s increased income could promote investments in R&D, high-tech infrastructure, human capital, technical innovation, entrepreneurial drive, and other initiatives that might promote economic complexity.

In the short run, financial development coefficients produce contradictory findings. The nonlinear ARDL has a negative and significant sum of coefficients, whereas the linear version of the ARDL’s error-correction has a positive and significant sum. Because it is in line with the study’s objective of drawing policy conclusions, this study places more emphasis on the nonlinear model when making decisions. In all the long-run analyses (linear and nonlinear ARDL, FMOLS), the coefficients of financial development are significantly negative. These findings indicate that the operations of Nigeria’s financial systems constitute impediments that reduce economic complexity. The findings contradict the theoretical exposition, which emphasizes the importance of an efficient and well-performing financial system. Financial systems provide finance and financial expertise to support the innovative and strategic efforts of firms and other organizations to deepen their knowledge-based productive capabilities. These help to propel investments in human capital, R&D, entrepreneurial inclinations, high-tech infrastructure, quality education, and other supporting initiatives to stimulate an upgrade in technological innovation and knowledge-based productive capacities to produce a wide range of sophisticated and globally competitive exports. The situation in Nigeria shows that financial systems have been promoting and channelling credit facilities to less innovative enterprises and firms producing unsophisticated and less competitive products for export. The findings indicate that Nigeria’s financial systems appear to encourage unproductive and less innovative enterprises at the expense of economic complexity-enhancing components. In line with the conclusions of this study, we suggest redirecting and pruning the activities of operators and stakeholders in Nigeria’s financial system. These could help to correct the lapses in the system and instigate support for economic complexity upgrades.

Nigeria’s overall financial development and operations appear to be disincentivizing the creation of supporting systems that encourage innovative and strategic initiatives to increase economic complexity. According to the research, Nigeria’s financial sector plays a key role in the country’s low economic complexity. The Nigerian government needs to work with financial systems’ regulatory agencies to create incentives and establish the required institutional structure, which could facilitate financial institutions and markets to support economic complexity-enhancing initiatives and programmes. These findings are consistent with the study outcomes of Ndoya et al. (2024), which show that financial development has a reducing effect on economic complexity. Conversely, it contradicts the findings of the studies, such as Olaniyi and Odhiambo 2023a); Avom and Ndoya (2024); Kamguia et al. (2023); Ndoya and Bakouan (2023); Ha (2023); Njangang and Nvuh-Njoya (2023); Ajide (2022), Neagu et al. (2022), Nguyen and Su (2021b); Atasoy (2021), Njangang et al. (2021), and Chu (2020). These studies establish that financial development is instrumental in raising the needed capital and providing financial know-how to support upgrades in economic complexity.

Institutional quality is another crucial determinant of economic complexity. In the nonlinear analyses (nonlinear ARDL and FMOLS), the short- and long-run coefficients of institutional quality (as measured by control of corruption) are insignificant. The linear ARDL’s long-run estimate validates this outcome. In the short run of the linear ARDL, the coefficient is statistically significant but negative. This study gives priority to the empirical results of the nonlinear analysis, in accordance with the earlier confirmation of asymmetry and nonlinearity. These results suggest that regulatory frameworks and institutional structures do not actively contribute to the economic complexity of Nigeria. The inactive role of institutional structures suggests the nonchalant attitudes of Nigerian governments and relevant agencies towards the development of technological innovation and knowledge-based productive capabilities to build up the country’s productive structure. Poor institutional performance may be the cause of Nigeria’s low placement in the world’s economic complexity rankings. As shown in Fig. 1, the average statistics indicate that Nigeria has the least sophisticated economy and productive structure in Africa. This finding might have a link to Nigeria’s appalling record in combating corruption. The country has an average score of 1.59 on a 0–6 scale. The findings highlight the inability of institutional quality to provide the necessary impetus and stimulus that could stimulate the initiatives and components that promote economic complexity upgrades in Nigeria. These findings have a variety of policy consequences. It implies that the regulatory frameworks and auxiliary institutions that direct the augmentation of investment in initiatives and activities aimed at enhancing technological innovation and knowledge-based productive capacities in Nigeria’s manufacturing process are ineffective.

These institutional deficiencies may make it more difficult to encourage firms and manufacturing companies to employ advanced technology, innovation, and knowledge-intensive manufacturing processes. It suggests that efforts to invest in R&D, entrepreneurial opportunities, high-tech infrastructure, human capital, high-quality education, technology, innovation, and other areas that are precursors to upgrades in economic complexity are not effectively encouraged by Nigerian institutions. This finding is inconsistent with the studies of Ajide (2022); Gnangnon (2021); Hoang and Chu (2023), Kamguia et al. (2023), Lapatinas and Litina (2019), Ndoya et al. (2024), Nguyen et al. (2023), Nguyen and Su (2021b), Njangang and Nvuh-Njoya (2023), and Vu (2022), which establish that institutional quality is a crucial component of EC that offers incentives and impetus for the generation of new knowledge of production, the development of productive capacities, R&D investments, the expansion of the patent market, inventive activity, and the improvement of human capital. It also stands at variance with the research outcomes of Zhang et al. (2023), which indicates a diminishing effect of institutional quality on economic complexity.

Diagnostic evaluations of nonlinear and linear error-correction version ARDL estimates

Following the procedures in existing research (Mohanty 2019; Olaniyi and Oladeji 2022; Olaniyi and Adedokun 2022; Olaniyi et al. 2023a; Ntembe et al. 2018), this study examines reliability and authenticity using a diverse range of diagnostic tests. These tests help to validate the robustness and appropriateness of estimates of nonlinear and linear error-correction variants of the ARDL estimation technique as outlined in Table 6 and 5. Table 8 presents a summary of the relevant diagnostic tests. The Jarque-Bera statistics in the two models indicate that error terms follow the assumption of a normal distribution. Ramsey Reset tests also attest that the linear and nonlinear models are free from specification errors. Thus, the model specifications follow scientific procedures, and they capture the correct and relevant variables. The Breusch-Godfrey serial correlation LM test confirms the absence of serial correlation. Similarly, the Breusch-Pagan-Godfrey test for heteroskedasticity affirms that there is no severe presence of outliers in the linear and nonlinear models. All four tests confirm the reliability and robustness of linear and nonlinear estimates.

Table 8 Diagnostic test for nonlinear and linear ARDL estimates

We use the cumulative sum (CUMSUM) and cumulative sum of squares (CUSUMSQ) tests to assess the stability of parameters. These tests help improve the robustness and reliability of the linear and nonlinear estimates from error-correction versions of ARDL estimators. The mean of the error terms and the variance are the foundation for the conclusions of these two tests, respectively. Figures 4 and 5 present the CUMSUM and CUMSUMSQ graphical representations for the linear and nonlinear models, respectively. The results of these tests confirm the stability of the short- and long-run estimates in the error-correction ARDL variants. Following Bahmani-Oskooee and Ng (2002) and Olaniyi et al. (2023a), the tests verify that the parameters are stable because the stability lines remain within the critical bounds.

Fig. 4

Stability test (CUSUM and CUSUMSQ) for the linear ARDL (2 3 3 4 1)

Fig. 5

Stability test (CUSUM and CUSUMSQ) for nonlinear ARDL (2 4 4 3 2 0)

The practical novelties and policy implications of the study

This section highlights the study’s practical novelties and contributions to expand the knowledge space to explain the main features of the nexus between natural resource wealth and economic complexity. This section also delves into the policy implications of the findings for resource-rich African countries. By doing so, this study offers potential solutions to address the interconnectedness of economic complexity, institutional quality, and natural resource wealth, taking cognizance of the socioeconomic fundamentals. The first phase discusses the new insights this study adds to advance the frontier of knowledge, while the second phase focuses on the policy implications of empirical outcomes for government, stakeholders, and policymakers’ consumption. The study’s novelties are outlined below: First, unlike earlier research, this study innovatively argues for and establishes asymmetric structures and nonlinearity in the sensitivity of economic complexity to changes in natural resource wealth. Second, this study contributes to global discussion by examining the differential impacts of positive and negative shock components in natural resource rents on economic complexity. Third, this study introduces an innovative idea that helps to explain resource curse syndrome regarding economic complexity within the context of asymmetry and nonlinearity. Fourth, this study is the first to adopt a nonlinear ARDL estimator to investigate the asymmetric effects of natural resource rents on economic complexity. Fifth, this study is the first to examine the determinants of economic complexity in Nigeria.

This study provides macroeconomic policy pathways for how Nigeria can leverage its natural resource abundance to improve its economic complexity. The study’s empirical findings and analysis serve as the foundation for policy suggestions. One, it would take tenacity and persistent efforts on the part of government agencies, parastatals, stakeholders, and Nigerians to design policies and initiatives to improve Nigeria’s economic complexity from one period to another. The study’s findings underline that present EC advancements are catalysts for future EC improvements. Two, there should be total overhauling, monitoring, and pruning of the portion of natural resource wealth earmarked for financing and creating incentives that enhance technological innovation and knowledge-based productive capacities ingrained in Nigerian manufacturing and productive bases. The study’s findings suggest that overall income from natural resources hinders Nigeria’s long-term prospects for improving economic complexity. This implies that the money earned from resources is being spent on unproductive activities, killing innovation and diversification, and dampening Nigeria’s motivation to enhance economic complexity. Thus, a thorough scrutiny of the regulatory framework that guides the administration of natural resource wealth becomes important.

Three, The Nigerian government should channel more of the cumulative increases in components of natural resource wealth to promote immediate (short-run) initiatives such as investment in the development of non-resource-based sectors and exports, patent markets, technological innovation, knowledge-based productive capacities, entrepreneurial inclinations, R&D, human capital, high-tech infrastructure, quality education, and others that are instrumental to increasing Nigeria’s economic complexity. This policy suggestion is based on the study’s findings, which indicate that the short-run positive shock component of natural resource wealth has a positive impact on economic complexity. Four, the management of Nigeria’s natural resource wealth needs thorough evaluation, scrutiny, and overhauling to uncover the system’s inherent shrewdness, rent-seeking, opportunism, and manipulative practices. This helps to explain why cumulative positive and negative changes’ components of resource income set aside for promoting the development of knowledge-based productive capabilities and technological innovation reduce economic complexity. Both the cumulative increases and decreases in natural resource wealth have the potential to hinder productive initiatives and hinder the adoption of innovative manufacturing methodologies in Nigeria. This, in turn, hampers the development of technological innovation and knowledge-based production capacities in the country. Therefore, it is crucial to consider this policy suggestion, as it is vital for the long-term development of economic complexity in Nigeria.

The study’s summary and conclusion

Scholars from all over the world have continued to show interest in economic complexity as a more comprehensive indicator of economic advancement. The nascent literature continues to expand and demonstrates the multidimensionality of the factors that determine economic complexity. A body of research shows that natural resource wealth can serve as a potential source of the capital required to diversify into economically viable and creative industries and to make investments in the knowledge-based technological capabilities that promote economic complexity, such as technology, innovation, entrepreneurial tendencies, R&D, high-tech infrastructure, sophisticated manufacturing abilities, quality education, and human capital. Meanwhile, in the worldwide ranking of economic complexity, most resource-rich countries score appallingly because they are entangled in a web of low technological innovation and weak knowledge-based productive capacities that prevent them from producing sophisticated and globally competitive exports. This phenomenon calls into question the effective channelization and allocation of natural resource wealth to promote initiatives that enhance economic complexity. The abundance of natural resource endowments breeds complacency and disincentivizes diversification and investment in components that stimulate economic complexity upgrades. Within the context of resource curse theory regarding economic complexity, scholars have continued to investigate the sensitivity of economic complexity to changes in natural resource wealth. The findings of existing studies predominantly confirm a diminishing effect of natural resource wealth on economic complexity.

Meanwhile, all existing studies assume linearity and symmetry in the sensitivity of economic complexity to changes in natural resource wealth. These oversimplifications do not align with the realities of the real world and socioeconomics. A recent advance in empirical research and econometrics confirms the existence of nonlinear and asymmetric features in financial data dynamics, including natural resource rents. This study is different from previous research because it explores how changes in natural resource wealth impact economic complexity, taking into account the nonlinearity and asymmetry commonly found in natural resource rents. This is important because natural resource prices are volatile, natural resource rent data is nonlinearly and asymmetrically distributed, and resource wealth fluctuates due to international market events. This study uses Nigeria’s annual dataset for the period 1984–2021. To accomplish the goal of the study, we use a nonlinear autoregressive distributed lag (ARDL) estimator.

Unlike previous studies, the study’s findings reveal robust evidence of short- and long-run asymmetry and nonlinearity in the sensitivity of economic complexity to changes in natural resource rents in Nigeria. The short-run positive and negative change components of natural resource rents have asymmetric and differential impacts on economic complexity. Specifically, we find that positive and negative changes in resource rents have different and asymmetric impacts on economic complexity in the short term. In the short term, cumulative increases in resource wealth (positive shocks) immediately stimulate and incentivize diversification and investment in initiatives that enhance technological innovation and knowledge-based productive capacities for manufacturing competitive and sophisticated exports. This, in turn, promotes economic complexity. Therefore, it is clear that cumulative increases in resource wealth positively contribute to Nigeria’s economic complexity in the short term. On the contrary, the cumulative decreases (negative shocks) in natural resource wealth disincentivize the initiatives that could produce short-run economic complexity-enhancing benefits in Nigeria. It implies that cumulation decreases in natural resource rents constitute curses to Nigeria’s economic complexity upgrades in the short run. The long-run findings indicate that both positive and negative shock components in natural resource rents impair long-term economic complexity in Nigeria. These findings highlight that natural resource rents breed complacency and disincentivize innovative moves to promote long-run prospects of economic complexity upgrades. These indicate that Nigeria’s natural resource endowments are catalysts that discourage investing in entrepreneurial initiatives, R&D, technology and innovation, high-tech infrastructure, human capital, quality education, and knowledge-induced productive capacities, all of which are economic complexity enhancers. These findings suggest that Nigeria’s economic complexity is a resource curse. Thus, Nigeria’s natural resource abundance inhibits long-term economic complexity upgrades.

These findings provide some fascinating, insightful, and thought-provoking facts on the roles of natural resource wealth in driving Nigeria’s technological innovation and knowledge-based productive capacities ingrained in manufacturing structures. Thus, this study presents the following policy highlights on how to leverage natural resource endowments to drive Nigeria’s economic complexity upgrades: Nigeria’s government, stakeholders, and policymakers should consider directing a larger portion of the increased income from natural resources towards initiatives that can enhance the country’s economic complexity in the short term. Two, Nigeria should discourage the persistent decline in the allocation of resource wealth towards promoting investment in key factors that stimulate an increase in economic complexity. This policy perspective is crucial. The short-run reductions in resource income allocated for the promotion of current technical innovation and knowledge-based productive capacities could lead to a deterioration of Nigeria’s existing economic complexity. Three, Nigeria’s management of natural resource wealth needs to be reviewed and overhauled in order to stimulate long-term and sustained economic complexity upgrades. The study’s findings reveal that both the positive and negative components of natural resource rents distort and mitigate long-run economic complexity in Nigeria. The Nigerian government should devise mechanisms and apparatuses for pruning and reviewing the management of natural resources. This would help detect and uncover sharp practices, corruption, racketeering, and shrewdness that divert resource incomes to unproductive and less innovative initiatives that hinder Nigeria’s long-term economic complexity upgrade. Fourth, the constant flow of income from natural resources in Nigeria breeds complacency, discourages diversification of the economy, and deters innovation. As a result of these socioeconomic ills, Nigeria’s long-term economic complexity deteriorates. Hence, these lapses should be prevented to spur an effective utilization of natural resource wealth to promote long-term economic complexity in Nigeria.

The study’s limitations and suggestions for future research

This study has successfully introduced a new dimension to global scholarship by providing new insights into the role of natural resource endowments in economic complexity. The study introduces an argument for and establishes an asymmetry and nonlinearity in economic complexity’s sensitivity to changes in natural resource wealth. However, critical examinations of existing knowledge and empirical exploration of the natural resource wealth-economic complexity nexus reveal some constraints. These constraints could give novel directions and insights to complement the study’s findings, enrich policy perspectives, and provide suggestions for future research. One, this study focuses specifically on Nigeria’s dataset. This restricts the policy’s relevance to Nigeria. Thus, scholars in subsequent research efforts should consider the cases of other countries to increase the global acceptance of our arguments and findings. Two, although we account for potential reverse causality bias and endogeneity concerns using a fully modified least squares (FMOLS) estimator, subsequent research endeavours should consider the possibility of bi-directional causality between natural resource wealth and economic complexity within the context of asymmetry and nonlinearity. Three, this study recommends that future research endeavours should extend to a panel analysis. The panel study should employ estimation techniques that take into account the asymmetric structures, nonlinearity, endogeneity, cross-sectional dependence, heterogeneity, and policy variations across quantiles that exist in the relationship between natural resource wealth and economic complexity. Four, we suggest that future research efforts should explore how unregulated corruption moderates the sensitivity of economic complexity to changes in natural resource rents. This study observes that a high level of uncontrolled corruption in resource-rich countries might be a bane that weakens the ability of natural resource wealth to accelerate economic complexity. These identified aspects should be the focus of future research endeavours to keep the research cycle revolving. These pitfalls do not diminish the study’s novelties and contributions to the global discussion. We raise these constraints to complement our research outcomes and enhance the global values of the novelties in the knowledge space at the nexus between natural resource wealth and economic complexity.