Foreign Expertise and State Capacity

Abstract

Technical assistance (TA) is widely used to accelerate development in low-income countries, often with explicit expectations that foreign expertise will be transferred to local institutions. Outcomes vary sharply: some governments convert imported know-how into durable bureaucratic capability; others deliver projects through parallel donor structures that collapse when funding ends (Berg et al., 1993; Cox & Norrington-Davies, 2020).

This report synthesizes published evidence to identify Context–Mechanism–Outcome (CMO) configurations under which TA contributes to sustainable development and genuine knowledge transfer (Pawson & Tilley, 1997; Wong et al., 2013). Using critical realism and realist synthesis, seven cases — Botswana, Rwanda, Senegal, Nigeria, Liberia, Afghanistan, and Sierra Leone — are coded into a conditions matrix. The central finding is that institutional location and enforceable transfer matter more than the volume of advice: embedding experts inside line ministries with succession planning correlates with durable capacity and retained knowledge; parallel, off-budget delivery correlates with dependency and knowledge loss when experts depart.

Introduction

Background to the problem

Technical advisors, embedded consultants, and management contracts have been standard development tools for decades. Donors and recipient governments routinely frame TA as a vehicle for building local capability — not merely for importing skills on a temporary basis. Yet the historical record is uneven (Andrews et al., 2013; Berg et al., 1993). Existing literature documents failure modes: capability traps, isomorphic mimicry, and project implementation units that hollow out civil services (Pritchett et al., 2013). It also documents success. What remains under-synthesized is the set of conditions under which TA produces sustainable development outcomes and transfers knowledge that persists after experts leave.

The stakes are structural. When TA substitutes for the state rather than building it, short-term delivery indicators can look strong while domestic capacity weakens. When TA is embedded, selective, and tied to transfer, imported expertise can become institutional memory rather than a recurring rental (Evans, 1995).

Research question and approach

The research question is deliberately conditional: under what conditions does technical assistance contribute to sustainable development and durable knowledge transfer in low-income countries? The question does not ask whether TA "works on average." It asks when, through which mechanisms, and under which contextual constraints expertise is likely to strengthen domestic capability rather than create dependency.

I synthesize secondary sources through a realist synthesis methodology grounded in critical realism (Pawson & Tilley, 1997; Wong et al., 2013). Seven cases are purposively selected and coded into a conditions matrix; CMO configurations are inferred from cross-case patterns. The contribution is an applied programme theory: conditional pathways linking TA design to sustainable development and knowledge transfer — generated from existing scholarship, not a new grand theory of state-building.

Literature Review

Technical assistance and the limits of transfer

Long-standing reviews caution that TA often creates useful but isolated "islands of capacity" that do not survive project closure (Berg et al., 1993; Cox & Norrington-Davies, 2020). Knowledge transfer is frequently promised in contract language but weak in practice. Modality matters analytically: expertise delivered inside public institutions, with named counterparts and handover obligations, differs fundamentally from expertise routed through parallel structures where consultants report to donors rather than to line ministries.

Embedded autonomy and capable bureaucracy

Evans's (1995) embedded autonomy explains why some states convert external resources into development. Capable bureaucracies are coherent and meritocratic yet connected enough to society to pursue shared goals — "embedded in a concrete set of social ties that binds the state to society" (Evans, 1995, p. 12). Evans and Rauch (1999) link Weberian features — merit recruitment, predictable careers — to stronger performance. For TA, the crucial question is whether foreign experts strengthen that bureaucratic core or bypass it.

Capability traps, mimicry, and PDIA

Pritchett et al. (2013) describe capability traps and isomorphic mimicry: states may adopt the visible form of capable institutions without building the underlying ability to implement and learn. Donors see reform; transfer stalls. Andrews et al. (2013) propose Problem-Driven Iterative Adaptation (PDIA): locally defined problems, experimentation, feedback, and broad engagement instead of imported blueprints. PDIA aligns with transfer-friendly TA when local actors own the problem frame and experts support iterative learning rather than deliver fixed solutions.

Sequencing and institutional absorption

Fukuyama (2004) frames state-building as a sequencing problem. External actors transfer resources, people, and technology more easily than institutional habits and tacit know-how — donors "know how to transfer resources, people, and technology," but "well-functioning public institutions require certain habits of mind, and operate in complex ways that resist being moved" (Fukuyama, 2004, p. 17). Inherited institutional strength, fragility, and political commitment shape whether knowledge sticks — though case evidence shows weak inheritance can be partly compensated by committed embedding (Acemoglu et al., 2003; Booth & Golooba-Mutebi, 2012).

Case-level evidence in the literature

Botswana and Rwanda appear among the clearest durable-transfer cases: embedded advisors, selective use, systematic counterpart training, and institutional locations that keep knowledge inside line ministries (Acemoglu et al., 2003; Sebudubudu & Lotshwao, 2009; Chemouni, 2019; Redifer et al., 2020). Afghanistan and Sierra Leone illustrate parallel delivery: off-budget aid, contractors, and project units that substitute for the state and shed capability when experts exit (Bizhan, 2018, 2023; DFID, 2008; World Bank Group, 2019b). Senegal and Nigeria show partial embedding — training cycles and local-content gains without secure transfer or retention (Blundo, 2006; Cowater International, 2026; Ovadia, 2013; Ovadia & Wolf, 2018). Liberia embeds experts inside ministries and achieves fiscal stabilization, yet succession and handover fall short (Gilpin & Hyu, 2008; World Bank Group, 2019a).

Read across cases, a pattern emerges: TA with embedding and transfer associates with sustainable development; TA through parallel structures associates with dependency and knowledge loss; partial embedding produces partial retention.

The synthesis gap

Much case literature explains individual countries in isolation. Realist synthesis (Pawson & Tilley, 1997; Wong et al., 2013) offers a mid-range approach: infer recurring mechanisms and contexts and express them as conditional propositions about when knowledge transfer succeeds. That is the gap this report addresses.

Methodology

The study follows critical realist philosophy and realist synthesis (Greenhalgh et al., 2011; Rycroft-Malone et al., 2010). Published secondary sources — peer-reviewed articles and selected donor or agency grey literature — are coded into a conditions matrix covering embedding, knowledge transfer, parallel structures, inherited institutions, fragility, and political commitment. From cross-case comparison, CMO configurations are inferred and assessed for confidence (high, medium, low) based on consistency and source quality.

Twenty-eight sources inform the synthesis. Search covered Scopus and EBSCO-indexed journals, publisher and repository verification, and agency portals (World Bank, IMF, DFID/FCDO, and others). Illustrative strings combined "technical assistance" or "foreign expert*" with "state capacity," "knowledge transfer," or "institutional capacity," plus theory-specific and case-specific terms.

Findings and Discussion

CMO-1: Embedding with enforceable transfer → sustainable development

Context. Botswana maintained expatriate officers inside the state to train counterparts rather than prematurely localizing posts without qualified staff; revenues flowed through the bureaucratic core (Acemoglu et al., 2003; Sebudubudu & Lotshwao, 2009). Rwanda, after institutional devastation, embedded advisors selectively inside ministries with strong local ownership and explicit transfer mandates (Booth & Golooba-Mutebi, 2012; Chemouni, 2019; Redifer et al., 2020).

Mechanism. Expertise placed in line institutions, paired with counterpart development, selective deployment, and succession planning, builds capability and retains knowledge inside the permanent state.

Outcome. Sustainable development outcomes — domestic staff who can perform functions independently after expert exit. Confidence: high.

CMO-2: Partial embedding → incomplete transfer

Context. Senegal's constrained administration (Blundo, 2006; Cowater International, 2026) and Nigeria's oil-sector local-content regime (Ovadia, 2013; Ovadia & Wolf, 2018).

Mechanism. Embedding and transfer were incomplete. Senegal trained large numbers of officials but institutionalization of know-how remained uncertain; Nigeria increased national participation while skill transfer proved hard to measure and vulnerable to capture.

Outcome. Mixed results — higher local presence without guaranteed retention of knowledge or durable capability. Participation is not the same as transfer. Confidence: medium.

CMO-3: Parallel structures → dependency and knowledge loss

Context. Afghanistan's aid dependence (Bizhan, 2018, 2023) and Sierra Leone's post-conflict gap-filling (DFID, 2008; World Bank Group, 2019b).

Mechanism. Off-budget delivery, contractors, and project units substitute for the state. Knowledge sits in parallel systems; when funding stops, capability evaporates. Donor-hired local assistants and PIUs can draw talent away from civil services rather than strengthen them.

Outcome. Short-term project success on weak institutional bases; collapse or dependency when experts depart. Confidence: high.

Liberia contrast: embedding without succession

Liberia embedded international experts inside ministries with co-signatory authority and documented fiscal-governance gains (Gilpin & Hyu, 2008; World Bank Group, 2019a) — yet transfer and succession expectations did not fully materialize. Outputs improved; durable handover did not. This case tightens the programme theory: sustainable development requires embedding plus enforceable knowledge transfer, not placement alone.

Resource rents as modifier (CMO-4)

Botswana suggests resource rents can support capacity building when routed through a meritocratic bureaucratic core that retains institutional knowledge (Acemoglu et al., 2003). Nigeria warns when rents bypass that core (Ovadia, 2013). Rents modify how the embedding-and-transfer mechanism fires; they do not replace it. Confidence: medium.

Conditions for sustainable development and knowledge transfer

Synthesizing across configurations, TA enables sustainable development and durable knowledge transfer when the following conditions align:

• Institutional location: Experts work inside line ministries and national authorities, not parallel donor units or off-budget structures (Evans, 1995).
• Enforceable transfer: Contracts specify named local counterparts, training milestones, documented handover, and time-bound expatriate phase-out.
• Political commitment: A protected meritocratic core channels TA toward institution-building rather than cosmetic reform (Booth & Golooba-Mutebi, 2012).
• Capability metrics: Success is measured by functions retained and knowledge applied after exit — not by project outputs alone (Pritchett et al., 2013).
• Selective use: Expertise is deployed where absorption capacity exists or is being deliberately built, avoiding overload of fragile systems (Fukuyama, 2004).

Strong inherited institutions help but are not strictly necessary. Botswana and Rwanda achieved durable transfer without high pre-existing bureaucratic scores; committed embedding compensated for weak starting points (Acemoglu et al., 2003; Booth & Golooba-Mutebi, 2012).

Limitations

The synthesis relies on secondary sources; CMO inference is only as strong as the underlying case literature. Transfer is often poorly measured in primary studies, which limits precision. Conditional propositions require context-specific testing before operational use.

Conclusion

Technical assistance is neither inherently building nor inherently destructive. Sustainable development and knowledge transfer depend on design: where advisors sit, what they must hand over, how retention is measured, and whether the state they enter is strengthened or substituted (Andrews et al., 2013; Berg et al., 1993).

Durable cases share embedding, selective use, counterpart development, and political will to protect institutional memory. Failure cases share parallel structures, off-budget delivery, and metrics that reward visible outputs over retained capability. For donors and recipient governments alike, the architecture of the TA contract — institutional location, transfer clauses, exit timelines, and capability indicators — is as decisive as the technical credentials of the advisor.

The programme theory offers a conditional map: when these conditions hold, imported expertise is more likely to become sustainable domestic capacity; when they do not, TA will continue to rent knowledge rather than transfer it.

References

1. Acemoglu, D., Johnson, S., & Robinson, J. A. (2003). An African success story: Botswana. In D. Rodrik (Ed.), In search of prosperity: Analytic narratives on economic growth (pp. 80–119). Princeton University Press.
2. Andrews, M., Pritchett, L., & Woolcock, M. (2013). Escaping capability traps through problem driven iterative adaptation (PDIA). World Development, 51, 234–244.
3. Berg, E., et al. (1993). Rethinking technical cooperation: Reforms for capacity building in Africa. UNDP.
4. Bizhan, N. (2018). Aid paradoxes in Afghanistan: Building and undermining the state. Asian Survey, 58(2), 297–318.
5. Bizhan, N. (2023). The political economy of state-building in Afghanistan. UNU-WIDER.
6. Blundo, G. (2006). Decentralisation, participatory governance and corruption in Senegal. In R. Hodess & T. Wolfe (Eds.), Global corruption report 2006 (pp. 142–148). Pluto Press.
7. Booth, D., & Golooba-Mutebi, F. (2012). Developmental patrimonialism? The case of Rwanda. African Affairs, 111(444), 379–403.
8. Chemouni, B. (2019). The rise of the economic technocracy in Rwanda. ESID, University of Manchester.
9. Cowater International. (2026). Technical assistance project to support the implementation of the Emerging Senegal Plan (AT-PSE). https://www.cowater.com/project/technical-assistance-project-to-support-the-implementation-of-the-emerging-senegal-plan-at-pse-project/
10. Cox, M., & Norrington-Davies, G. (2020). Technical assistance: New thinking on an old problem.
11. DFID. (2008). Evaluation of DFID country programmes: Sierra Leone (Evaluation Report EV690). https://assets.publishing.service.gov.uk/media/5a789b7240f0b63247698bac/ev690.pdf
12. Evans, P. B. (1995). Embedded autonomy: States and industrial transformation. Princeton University Press.
13. Evans, P., & Rauch, J. E. (1999). Bureaucracy and growth: A cross-national analysis of the effects of "Weberian" state structures on economic growth. American Sociological Review, 64(5), 748–765. https://doi.org/10.2307/2657374
14. Fukuyama, F. (2004). The imperative of state-building. Journal of Democracy, 15(2), 17–31. https://doi.org/10.1353/jod.2004.0026
15. Gilpin, R., & Hyu, E. (2008). Is Liberia's governance and economic management assistance program a necessary intrusion? USIP Peace Brief.
16. Greenhalgh, T., Wong, G., Westhorp, G., & Pawson, R. (2011). Protocol—realist and meta-narrative evidence synthesis: Evolving standards (RAMESES). BMC Medical Research Methodology, 11(1), 115. https://doi.org/10.1186/1471-2288-11-115
17. Ovadia, J. S. (2013). The Nigerian "one percent" and the management of national oil wealth through Nigerian content. Science & Society, 77(3), 315–341. https://doi.org/10.1521/siso.2013.77.3.315
18. Ovadia, J. S., & Wolf, T. (2018). Studying the developmental state: Theory and method in research on industrial policy and state-led development in Africa. Third World Quarterly, 39(6), 1056–1076. https://doi.org/10.1080/01436597.2017.1368382
19. Pawson, R., & Tilley, N. (1997). Realistic evaluation. Sage.
20. Pritchett, L., Woolcock, M., & Andrews, M. (2013). Looking like a state: Techniques of persistent failure in state capability for implementation. Journal of Development Studies, 49(1), 1–18. https://doi.org/10.1080/00220388.2012.709614
21. Redifer, L., Alper, E., Newiak, M., Gursoy, T., Meads, N., Kwalingana, S., & Thomas, A. (2020). The development path less traveled. IMF Departmental Paper 20/010. https://doi.org/10.5089/9781513551371.087
22. Rycroft-Malone, J., Fontenla, M., Bick, D., & Seers, K. (2010). A realistic evaluation: The case of protocol-based care. Implementation Science, 5(1), 38. https://doi.org/10.1186/1748-5908-5-38
23. Sebudubudu, D., & Lotshwao, K. (2009). Managing resources and the democratic order: The experience of Botswana (Occasional Paper No. 31). South African Institute of International Affairs.
24. Wong, G., Greenhalgh, T., Westhorp, G., Buckingham, J., & Pawson, R. (2013). RAMESES publication standards: Realist syntheses. BMC Medicine, 11(1), 21. https://doi.org/10.1186/1741-7015-11-21
25. World Bank Group. (2019a). Building financial management capacity in fragile and conflict-affected states: The case of Liberia. https://documents1.worldbank.org/curated/en/845771555484260672/pdf/Building-Financial-Management-Capacity-in-Fragile-and-Confliict-Affected-States-The-Case-of-Liberia.pdf
26. World Bank Group. (2019b). Sierra Leone—Public sector pay and performance project. http://documents.worldbank.org/curated/en/767651554150478426