The Clark-Fisher model — developed by Allan Fisher (1939) and Colin Clark (1940) — predicts that economies move through phases of dominant employment: primary in pre-industrial countries, secondary in industrialising countries, tertiary (and now quaternary) in post-industrial countries. Whether this model captures the CENTRAL pattern of economic development is largely a YES, with important qualifications.
Strong evidence FOR the model.
1. Historical UK trajectory. The UK's employment shift over 200 years tracks the Clark-Fisher model almost perfectly. In 1841, ~25% of the workforce was in agriculture. By 1900, agriculture had fallen to ~10%, with manufacturing peaking at ~40% during the Industrial Revolution. By 1970, manufacturing had peaked and started declining; services were rising. Today: primary ~1%, secondary ~18%, tertiary ~80%. This is exactly the predicted path.
2. China 1980-2025. China shows the same pattern at FAST FORWARD. Primary fell from ~70% in 1980 to ~25% today; secondary boomed; tertiary now leads. The pattern matches Clark-Fisher even though China's trajectory was state-directed rather than market-led.
3. Global cross-section today. Looking at countries at different income levels reveals the predicted pattern:
- LICs (Ethiopia, Burkina Faso, Nepal): primary-dominated.
- MICs (China, Indonesia, Brazil, Vietnam): secondary substantial, tertiary growing.
- HICs (UK, USA, Germany, Japan): tertiary + quaternary dominate.
This cross-sectional pattern matches the longitudinal (over-time) path the model predicts.
4. Predictive power. Countries that industrialise (India, Vietnam, Bangladesh in the 2000s-2020s) DO shift workers from agriculture into manufacturing as the model predicts. Then as they mature, services rise. The model's predictions hold empirically.
5. Underlying logic is sound. The model captures a REAL CAUSE-AND-EFFECT mechanism: as productivity rises in lower sectors (mechanised farming, automated factories), labour gets RELEASED into higher sectors that demand it. Mechanisation of UK farming reduced primary employment from 25% to 1% over 200 years; the released workers went into industry then services. This logic is universal.
Qualifications and exceptions.
1. Resource-rich economies skip the pattern. Oil-rich states (Saudi Arabia, UAE, Qatar, Kuwait) have small primary sectors (~5-10%) but never went through a major industrial phase. They jumped from primary extraction to a service-heavy economy underpinned by oil exports. These don't fit Clark-Fisher.
2. Stalled industrialisation / Middle-income trap. Some countries get stuck in MICs without progressing to HIC status. Argentina, Brazil, Malaysia have all hovered at middle-income for decades. The model predicts continued transition; reality shows it can stall.
3. Differential survival of secondary in HICs. Germany retains ~24% secondary; Japan ~25%; whereas UK ~18% and USA ~15%. The model predicts secondary decline in all HICs; the reality is that POLICY + SPECIALISATION matter. Germany's Mittelstand engineering, dual-system apprenticeships, export focus have preserved manufacturing in a way the model doesn't predict.
4. Geographic unevenness within countries. India's national average (~42% primary) hides that Bangalore is QUATERNARY-dominated while rural Bihar is still essentially pre-industrial. National statistics MASK enormous regional variation. China similarly: Shenzhen is post-industrial; Western provinces still primary-heavy.
5. The quaternary sector wasn't in the original model. Fisher + Clark used only three sectors. The quaternary (knowledge economy) emerged from ~1970s and was only added later. The original model didn't predict a high-value knowledge sector that would account for 15-25% of HIC employment.
6. Globalisation changes the rules. The model assumes countries develop their own manufacturing then their own services. But globalisation means modern LICs/MICs can BYPASS some stages — Bangladesh became a major garment manufacturer without going through the heavy industry phase the UK did. Vietnam went from farming to electronics assembly in one generation. India developed a service-export economy (IT) BEFORE traditional industrialisation completed. The model is too LINEAR for a globalised economy.
7. Informal economies undercount the model. Much LIC and MIC employment is INFORMAL — street vendors, casual labour, subsistence farming — that isn't fully captured in official sector statistics. Ethiopia's "65% primary" partly reflects informal urban services that go uncounted.
8. Automation may break the pattern at the top. As AI + robotics develop, even tertiary + quaternary work may shrink — translators replaced by GPT, designers by Midjourney, accountants by automation. The model didn't anticipate this. The next phase may be a Phase 4: shrinking employment overall as machines take over more tasks.
Applied judgement.
The Clark-Fisher model captures the CENTRAL pattern of economic development — the long-run shift from primary to secondary to tertiary + quaternary — and the underlying ECONOMIC LOGIC is sound. It's the best simple framework geographers have for comparing employment structures across countries and centuries.
But the model is a MAP, not the territory. It works as a BIG-PICTURE GUIDE but misses:
- Resource-rich exceptions.
- Stalled transitions.
- Policy-driven preservation of secondary (Germany, Japan).
- Geographic unevenness within countries.
- Globalisation-driven bypass of traditional stages.
- Informal economies.
- The original three-sector framing being too narrow for the quaternary age.
- Automation potentially upending the whole pattern.
The right use of the model is as a STARTING POINT for understanding employment structures — the central tendency is real and important — coupled with awareness that DEVIATIONS reveal the most interesting geographical stories. Germany's preserved secondary tells us about policy; India's geographic unevenness about uneven development; Bangladesh's manufacturing leapfrog about globalisation; Saudi Arabia's bypass about resource dependence.
Conclusion. The Clark-Fisher model captures the central pattern of economic development AND deserves its position as the standard framework taught in 4GE1. But A* candidates should know its limits + use it with awareness, not as a rigid prediction but as a useful simplified map of a complex territory. The model is RIGHT enough often enough to be invaluable; sophisticated geographers know when to look BEYOND it.