Capital Investment and Carbon Emissions in Somalia: What a 2024 Study Reveals

A digest of Nur, Adan, Ahmed, Gutale, Ali & Dalmar (2024), International Journal of Energy Economics and Policy
Since the Federal Republic of Somalia was established in 2012, investment in ports, roads, airports, hospitals and private enterprise has picked up. The authors report that gross fixed capital formation — spending on long-lived assets such as buildings, infrastructure and equipment — grew from about $134 million in 1991 to roughly $920 million in 2019. Environmental economists often expect such build-outs to push carbon dioxide emissions upward. This study tests that expectation for Somalia using national data for 1991–2019. Its results are more nuanced than its own summary suggests, making it a useful read for anyone working on climate policy, reconstruction or energy transitions in low-income, fragile economies.

The study covers 1991–2019, giving 29 annual observations, and uses 3 estimators: ARDL, FMOLS and DOLS. The long-run GCF elasticity in the ARDL model is −0.105, and each 1% rise in the renewables share is linked to a fall in CO₂ of about 12%.

Article at a Glance
The article is "Investigating the Effect of Gross Capital Formation on Carbon Emissions in Somalia," by Abdirahman Mohamed Nur, Ahmed Hassan Adan, Ahmed Dahir Ahmed, Ali Abdukadir Ali Gutale, Ali Yassin Sheikh Ali and Mohamed Saney Dalmar. It was published in the International Journal of Energy Economics and Policy, Vol. 14, No. 4, pp. 631–641 (EconJournals; open access, CC BY 4.0), in 2024 (received 15 January; accepted 27 May). Its research area is energy and environmental economics, and its study context is Somalia, using national annual data for 1991–2019. The method is quantitative time series: ARDL bounds testing, with FMOLS/DOLS robustness checks and Granger causality. The DOI is https://doi.org/10.32479/ijeep.15788.

The Research Problem
Evidence on whether capital formation raises emissions is divided. Among the studies the authors review, some find emission-increasing effects (China, Malaysia, the Balkans, ASEAN), others find no significant link (Egypt, a 44-country African panel), and one finds that it depends on the policy period (India).
For Somalia, the authors argue, the question had not been examined directly — even though investment is central to reconstruction and environmental pressure is acute. The paper cites evidence that about 82% of household energy comes from charcoal and firewood.

Research Objective
The central aim is to estimate how gross capital formation (GCF) relates to CO₂ emissions in Somalia in the long and short run, and to draw policy lessons. No formal research questions are stated; the tested hypothesis is the standard one of no long-run relationship. The model also examines:

• whether GDP and GDP² trace an Environmental Kuznets Curve (emissions rising, then falling, with income);

• renewable energy's share of final energy use;

• urbanization, measured as urban population.

How the Study Was Conducted
A quantitative, single-country time-series design using 29 annual observations (1991–2019) from the OIC's SESRIC database and the World Bank's World Development Indicators. CO₂ emissions are the dependent variable; GCF, GDP, GDP², renewable energy share and urban population are explanatory. All variables are in logarithms, so coefficients read as approximate percentage responses (elasticities).

The analysis had six stages. In stage 01, Stationarity, ADF and Phillips–Perron tests showed a mix of I(0) and I(1) series, none beyond first difference. In stage 02, Cointegration, the ARDL bounds test gave F = 13.489, above the 1% upper bound (4.68), so a long-run relationship exists. In stage 03, Estimation, ARDL long- and short-run models were fitted, with an error-correction term of −0.750, meaning about 75% of any gap closes within a year. In stage 04, Diagnostics, normality, heteroskedasticity, serial-correlation and ARCH tests passed, and CUSUM plots were stable. In stage 05, Robustness, the long-run results were re-estimated with FMOLS and DOLS. In stage 06, Direction, pairwise Granger causality tests were run, with lags chosen by the Hannan–Quinn criterion.

Key Findings
The long-run coefficients below are as reported in Tables 6 and 7 of the article, across the ARDL, FMOLS and DOLS estimators. Significance is marked at 10% (), 5% () and 1% (). Gross capital formation has a coefficient of −0.105 in ARDL (significant at 5%), −0.111 in FMOLS (1%) and −0.151 in DOLS (1%). GDP has 1.752 in ARDL (not significant), 2.382 in FMOLS (10%) and 4.245 in DOLS (1%). GDP² has −0.040 in ARDL (not significant), −0.053 in FMOLS (10%) and −0.096 in DOLS (1%). Renewable energy share has −11.977 in ARDL, −11.508 in FMOLS and −13.383 in DOLS, all significant at 1%. Urbanization has 0.142 in ARDL (5%), −0.021 in FMOLS (not significant) and 0.095 in DOLS (5%).

1 Capital formation shows a small negative association with emissions.
In the ARDL model, a 1% rise in GCF is associated with about a 0.105% fall in long-run CO₂; FMOLS and DOLS give similar values. The short-run effect is also negative (−0.063), and GCF Granger-causes CO₂ at the 5% level, not the reverse.
READING NOTE
The abstract and conclusion say GCF has no significant influence on emissions, yet the paper's tables report small but statistically significant negative coefficients. The article does not reconcile the two; readers should consult Tables 6–8 directly.
2 Renewable energy has the strongest association.
A 1% rise in renewables' share of final energy is linked to roughly a 12% fall in long-run CO₂ in all three estimators, and renewables Granger-cause CO₂ at 1%. The share varies only between about 88% and 95%, which partly explains the large elasticity.
3 Urbanization is linked to higher emissions in most models.
ARDL (+0.142) and DOLS (+0.095) show a positive long-run link; FMOLS does not, and Granger tests find no causal ordering.
4 Evidence for an Environmental Kuznets Curve is mixed.
GDP and GDP² carry the expected signs but are insignificant in the ARDL model, so no EKC is confirmed there. The authors read the significant FMOLS and DOLS results as supportive.

What the Study Contributes
The contribution is primarily empirical: country-level time-series evidence for Somalia, a setting with little prior work on this question, estimated with three long-run estimators rather than one. It does not propose new theory or methods. The authors recommend an environmental framework that looks beyond investment volumes — energy diversification, technology transfer, sustainable practices and climate resilience — backed by monitoring, public awareness and international collaboration.

Important Limitations
IDENTIFIED BY THE AUTHORS

• Data end in 2019, missing recent changes.

• Model complexity raises overfitting and interpretation risks.

• Focusing on GCF may omit other drivers; data are incomplete.

• A linear specification may oversimplify; geopolitical shocks are hard to isolate.
  SKILFUL CAUTIONS ARISING FROM THE STUDY DESIGN

• 29 observations is a small sample for a six-variable model with lags.

• Granger tests show predictive precedence, not cause and effect.

• Aggregate GCF cannot separate road building from, say, machinery imports.

• National CO₂ series usually exclude land-use emissions such as deforestation, which the paper flags as a major pressure.

• Some units in the variable and descriptive tables look inconsistent; replication should start from the source data.

Why This Research Matters
Somalia is planning large infrastructure outlays, including a transport investment plan of about $1.1 billion over ten years cited in the paper. Evidence on how investment relates to emissions is relevant to planners, development partners and climate negotiators. For students, the article is an open-access worked example of the ARDL workflow in a data-scarce country.

Skilful Research Insight
Editorial commentary — not a finding of the study
Three lessons stand out. First, read the results tables alongside the abstract: here the headline and the coefficients point in different directions, and the numbers deserve the last word. Second, match the emissions measure to the mechanism. If charcoal and deforestation dominate Somalia's environmental pressures, future work could pair fossil CO₂ with land-use emissions or broader indicators; note too that the World Bank's renewable-energy indicator counts solid biofuels such as charcoal and firewood as renewable. Third, disaggregating investment (public infrastructure, housing, equipment), testing non-linear models as the authors propose, and extending the data beyond 2019 would sharpen the policy message.

Read the Original Research
The original article is "Investigating the Effect of Gross Capital Formation on Carbon Emissions in Somalia," published in the International Journal of Energy Economics and Policy, 14(4), 631–641 (2024), by Abdirahman Mohamed Nur, Ahmed Hassan Adan, Ahmed Dahir Ahmed, Ali Abdukadir Ali Gutale, Ali Yassin Sheikh Ali and Mohamed Saney Dalmar. Its DOI is https://doi.org/10.32479/ijeep.15788, and the publisher/journal page is EconJournals, www.econjournals.com (open access, CC BY 4.0). Skilful summarises; all findings belong to the original authors.

SEO Information
The SEO title is "Gross Capital Formation and CO₂ Emissions in Somalia: 2024 Study Summary." The meta description is "A Skilful digest of a 2024 IJEEP study using ARDL, FMOLS and DOLS to examine how capital formation and renewables relate to CO₂ in Somalia." The URL slug is /gross-capital-formation-carbon-emissions-somalia. The primary keyword is gross capital formation and carbon emissions. The secondary keywords are Somalia CO₂ emissions; ARDL bounds test; environmental Kuznets curve; renewable energy and emissions; FMOLS and DOLS.