Project Presentation • Ironhack DA bootcamp

01 / 10

CO₂ & Economic Growth:�Who's Actually Decoupling?

An honest, data-driven analysis of the Paris Agreement's promise across 189 countries, 7 years, and 3 critical hypotheses.

189

COUNTRIES TESTED

2014–20

POST-PARIS WINDOW

1,323

DATA POINTS ANALYZED

5

INTEGRATED DATASETS

02 / 10

Business Case

As a coalition of environmental researchers we track whether the Paris Agreement has delivered measurable results. Using data from 2014 to 2020, we investigate whether economic growth and carbon emissions are genuinely decoupling or, whether wealthy nations are simply exporting their footprint. We aim to hold regional climate commitments accountable.

Hypothesis 2

Where renewable energy adoption has grown, have CO₂ emissions actually fallen — and where is the gap largest?

Hypothesis 3

Do high CO₂ emissions always reflect development and improved living standards, or can they coexist with energy poverty?

The Decoupling Conflict

Hypothesis 1

Has GDP decoupling from CO₂ emissions? — and does the answer change depending on how we measure emissions?

03 / 10

Our World in Data: Primary CO₂, land use change, CO2 per capita and primary energy consumption.

United Nations Stats: Country-level GDP and population database (2014-2020).

World Bank DataBank: Electricity access, renewable energy production and consumption.

Collect & merge: 5 datasets joined on country + year keys, producing 1,323 country-year observations across 189 nations

Clean & validate: handle missing values (..), removed regional aggregates and standardized country names.

Create features: calculated derived variables like CO2 intensity, CO2 per capita and GDP quartile groups.

The project Pipeline

DATA PIPELINE

03 / 10

ERD schema

To understand if economic growth is truly decoupling, analysis is done on carbon intensity (CO2 emissions per dollar of GDP) through two different perspectives:

04 / 10

co2_int = co2 / GDP

Fossil and industrial emissions only. This is the highly optimistic headline metric commonly reported by developed countries to showcase their green progress.

co2_luc_int = co2_luc / GDP

Includes land-use change and regional deforestation. This exposes whether nations are simply shifting their physical ecological damage to foreign supply chains.

H1: Has decoupling happened across regions?

-> CO₂ intensity of GDP declined across regions

Two Definitions of Decoupling

H1: Has CO₂ intensity declined across regions 2014–2020?

Decoupling metric = CO₂ intensity = co2 / GDP).

• CO₂ intensity declined across most regions between 2015 and 2018.
• Only Europe showing the most consistent downward trend (decoupling sign) throughout the full 2014–2020 window.
• Asia holds the highest intensity levels.

The smaller CO₂ intensity -> more decoupling

H1: Has CO₂ intensity declined across regions 2014 -2020?

Does the answer change with land-use included?

​

• CO2 intensity levels increases for all regions, as expected as land-use captures the hidden emissions.
• The ranking on CO2 intensity does changes, Africa takes the place of Asia as the region with the highest CO2 intensity .

• Africa & South America: clearest evidence of emissions displacement rather than elimination. Fossil and industrial processes can look relatively clean without the land-use change (primarily deforestation).
• Europe stops showing signs of decoupling as its decreasing trend flattens, however it maintains its CO2 intensity levels through time.

Scenario B

Scenario A & B

H1: Has CO₂ intensity declined across regions 2014 vs 2020?

Does the answer change with land-use included?

​

• Scenario A (fossil only): most regions show positive delta = CO₂ intensity increased 2014→2020. Only Europe and North America declined.
• Scenario B (land-use included): all regions show negative delta = CO₂ intensity decreased when land-use is added.
• Comparison is counterintuitive: likely mechanism - land-use CO₂ (mostly deforestation) could peaked before 2014 and declined over the period, which pulls the full-footprint intensity down even as fossil emissions kept rising. So Scenario B could look "better" not because countries cleaned up, but because they cut fewer forests in 2020 than in 2014 — which is a very different thing from actual decoupling.

07 / 10

The Energy Production-Consumption Gap

H2: Renewables -> Driver for CO2 emissions decrease?

Where renewable energy (RE) adoption has grown, have CO₂ emissions actually fallen — and where is the gap largest?

• For Africa, Asia and Oceania: RE share production increases vs. CO2 emissions trend increasing -> RE not driving CO2 decrease.
• North America, Europe, South America: RE share relates better with CO2 emissions decrease, but not consistent pattern. Renewable-to-CO₂ relation is region- and period-specific.
• 'Exporting clean energy' signal [Africa, Asia, Oceania and North America]: the RE production trend differs from RE consumption, gap for most years

08 / 10

Hypothesis 3: Emissions Without Benefit?

THE HYPOTHESIS

Do high CO₂ emissions reflect development and improved living standards or can they coexist with energy poverty, revealing that some countries may be polluting on behalf of others?

WE USED THREE LENSES:

Wealth vs. Access

Does GDP per capita predict electricity access across regions?

Income Groups vs. Emissions

Do richer countries emit more, and do their people benefit more?

The Paradox Quadrant

Which countries emit heavily while their population lacks electricity?

08 / 10

Hypothesis 3: Wealth vs Access

THE HYPOTHESIS

FINDING 1: Richer Regions, Better Access to Electricity

​

Wealth vs. Access

Does GDP per capita predict electricity access across regions?

Income Groups vs. Emissions

The ranking is identical in both charts except for South America. The level of wealth predicts mostly the level of access to electricity.

09 / 10

Wealth categories by percentiles:

Q3 Upper-Middle (75%): they emit the most in absolute terms, suggesting they are in peak industrialisation.

Q4 Richest (100%): richest countries emit the most per person but not the most in total.

Q2 Lower-Middle (50%): follows the same tendency as Q1, lower access to electricity equals lower CO2 emissions.

Q1 Poorest (25%): poorer countries have less electricity access but also emit less in total

Hypothesis 3: Who emits the most but has the least?

FINDING 2: Mid-income countries emit the most in total. Rich countries emit the most per person. Poor countries emit the least and still go without power.

​

08 / 10

FINDING 3: Emissions Without Domestic Benefit

Nigeria emits ~105 Mt CO₂ annually YET, only 56% of its population has electricity access. Its emissions are driven by oil extraction and industrial activity, not by powering its own people.

Hypothesis 3: The Energy Access Paradox

India, Indonesia, Pakistan, and South Africa all appear in the high-emissions/low-access quadrant. India alone has ~230 million people without electricity despite ranking among the world's top 3 total emitters.

08 / 10

High emitters, Low Beneficiaries

Hypothesis 3: The Paradox Quadrant

08 / 10

Hypothesis 3: Progress is real but unequal

FINDING 4: Electricity access improved in every region between 2014 and 2020. But Africa's starting point is so far below the rest that positive growth still leaves it at ~50% by end of period, the gap is not closing fast enough.

All 6 regions improved

Electricity access rose in every region. The direction of travel is positive everywhere.

Africa: +7.7pp gain

Grew from 43% to ~51% over 7 years which is meaningful progress but far from parity with other regions.

Gap vs. Europe: 48pp

Africa ends the period ~48 percentage points below Europe. At current pace, convergence is decades away.

• Our analysis proves that environmental success claims depend entirely on reporting honest metrics and even with the current metrics..

​

• There’s a structural dynamic: a country’s generation of carbon output is decoupled from local welfare delivery. Countries produce renewable energy and export it; countries emit industrial levels of CO₂ without their populations accessing electricity.

​

• Decoupling is partial, fragile, and regionally uneven: H1 establishes that CO₂ intensity declined in some regions over some periods. H2 shows that renewables are a contributing mechanism but not dominant. The Paris Agreement has produced measurable but insufficient progress.

​

10 / 10

Hypothesis 1

Is GDP decoupling from CO₂ emissions ?

Partial decoupling is visible in some regions (Europe), but not sustained or universal.

Hypothesis 1

​

Does the answer change with land-use?

​

It does, Europe stops showing clear signal of decoupling and land-use shows evidence of emissions displacement rather than elimination.

Hypothesis 2

Renewable Penetration as the Mechanism

partially supported. Renewables are a contributing factor but not the sole driver and this view is region-period specific.

Hypothesis 3

The CO2 - welfare disconnect

Wealth does predict access to electricity and the paradox quadrant does exist but with this data we can’t claim wealth displacement from CO2.

Conclusions

10 / 10

DATA ANOMALIES

Africa's trend is outlier-sensitive and requires country-level decomposition before any regional claim holds

CAUSALITY IN HYPOTHESIS

H2 cannot measure renewables' impact without controlling for other variables. H3 cannot confirm export-driven emissions without trade data.

COVID-19 IN 2020

The economic contraction artificially suppresses CO₂. Any decoupling in 2020 requires a growth scenario before it can be attributed to structural change.

Caveats & recommendations

https://static.vecteezy.com/system/resources/previews/065/964/564/large_2x/deforestation-causes-destruction-alongside-lush-green-forest-in-contrasting-landscape-view-photo.jpeg

Source: www.vecteezy.com

Image Sources