Paper

Impact Assessment Methodologies for Microfinance: A Review

What is required to improve impact assessment studies?

This document reviews the literature and practice on impact assessment (IA) of microfinance projects.
Notes that:

  • Impact assessments can be viewed as laying on a continuum between 'proving impacts' and 'improving practice.' While many different groups have interests in IA there is currently great pressure on donor agencies to prove impact;
  • Different conceptual frameworks provide the foundations for IAs. These cover a model of the impact chain, the identification of units (levels) of assessment and specification of the types of impact to be assessed;
  • The range of methods used for IA have increased over the last decade - sample surveys, rapid appraisal, participant observation, case study and participatory learning and action. Each method has particular strengths and weaknesses. The choice of methods should be based on objectives, costs and feasibility. Mixed methods are preferable;
  • Attribution and fungibility are key problems for impact assessment studies. The ways in which scientific, humanities-based and participatory methods deal with these problems are examined. Argues that the problem of fungibility is not as great as many writers suggest;
  • The key to selecting an appropriate methodology rests on achieving a fit between IA objectives, program contexts and the availability of resources. Four types of methodology are identified - monitoring, simple, moderate and complex - and their key characteristics are examined.

Presents a number of recommendations for IA practice:

  • More emphasis should be placed on internal impact monitoring within MFIs, rather than on impact assessment;
  • The idea of methodological rigour must be applied to all designs and methods;
  • The main need in IA is to improve the quality of simple IA designs;
  • Monitoring MFI performance is not sufficient - IAs must look at impacts on livelihoods;
  • The Household Economic Portfolio Model (HEPM) and other multi-level approaches, are to be recommended for complex IA studies;
  • Problems of attribution can be overcome by careful design;
  • Many IAs have 'low impact' because they do not address the issue of dissemination sufficiently;
  • Donor activity has done little to strengthen the human resources or institutions available for IA in non-OECD countries and this must be corrected;
  • Improving IA standards requires major efforts to raise the capacities of an 'army' of impact assessors who are careful trained and have a code of practice.

[Author's abstract]

About this Publication

By Hulme, D.
Published