System Dynamics Modeling of HWTS Diffusion Process in Developing Countries

Transcription

1 System Dynamics Modeling of HWTS Diffusion Process in Developing Countries International Symposium on Household Water Management and Fourth Annual Meeting of the International Network to Promote HWTS 2-5 June, 2008 Accra, Ghana Tommy Ka Kit Ngai PhD student Centre for Sustainable Dev. University of Cambridge England

2 Problem Background Although household water treatment and safe storage (HWTS) systems have many advantages, usage remains low (Clasen, 2006; Murcott, 2006; Luby et al., 2008) Estimated 5-10 million regular users worldwide Low level of sustained use Few small scale successful initiatives, but not yet reached scale Research Objective: What are the critical factors, barriers, and solutions to the introduction, adoption, scale-up, and sustained use of HWTS practices in less developed countries?

3 Existing HWTS Research Literature review of 100+ HWTS related papers, reports Existing research typically views from 3 perspectives (1) Technical: Key assumption: improved technical, health impact, and/or social characteristics of HWTS can give confidence to implementers, policymakers, and end-users to promote, support, and use HWTS Typical outcome: evaluation and improvement of technology (e.g. Clasen et al., 2005; Brown, 2007)

4 Existing HWTS Research (2) Behaviourial: Key assumption: users behaviour is rational and predictable. Manipulation of behaviourial factors can encourage HWTS adoption Typical outcome: identify and evaluate different social, psychological, and/or communication techniques to influence behaviour (e.g. Moser et al., 2005; Kincaid and Figueroa, 2005)

5 Existing HWTS Research (3) Business Marketing: Key assumption: strategies to market and sell consumer products in western society can be effectively applied to sell HWTS in less developed countries Typical outcome: evaluate the strengths and weaknesses of supply scheme, bench-marking best business practices (e.g. Carpenter, 2003; Frey et al., 2006; POUZN, 2007)

6 Current Knowledge Gaps Many excellent studies, diverse views, many factors identified to affect HWTS dissemination But are these all the critical factors? How are the factors inter-related? Essentially a innovation diffusion process, wealth of knowledge available yet largely unapplied to HWTS

7 Innovations Diffusion Process Diffusion of successful innovations generally follows a S-shape curve (Rogers and Shoemaker, 1971) Implications: 1. Some people adopt first, some adopt later, some not at all 1. Adoption rate is slow at first, accelerates after critical mass 1. To scale-up HWTS, want to quickly reach critical mass, steepen the curve, raise the saturation point, but how?

8 Innovations Diffusion Process Depending on the context, diffusion process is generally governed by one or more of the these 5 perspectives: Communications (Rogers, 2003): Adopting a new innovation is risky Diffusion occurs when perceived risk is gradually reduced e.g. observable tech performance, reputable promoters, peer pressure Economic History (Rosenberg, 1972): Potential adopters are rational economic agents Diffusion performance cost ratio gradually improves

9 Innovations Diffusion Process Development (Harrison, 1994): Potential adopters requires resources for adoption Diffusion people gradually gain purchasing power Market Infrastructure (Brown, 1981): Availability to innovation is nonuniform Diffusion innovation is gradually made available to different regions Entrepreneurs (Miller and Garnsey, 2000): Adoption can only occur when supply meets demand Diffusion entrepreneurs bring buyers and sellers together

10 Where are we now? Diffusion theories suggested additional perspectives and critical factors HWTS dissemination is complicated. 8 perspectives, 100+ factors, many recommendations Rather than digging deeper, should step back, understand the big picture and inter-relationships Want to identify leveraged policy recommendations Need holistic and integrative tool to systemically analyze, operationalize what we already know

11 System Dynamics Modeling Complex systems analytical tool Flexible, can incorporate soft variables Emphasis on non-linearity, linkages, feedbacks, tradeoff, and system behaviour Successfully applied in public and private consulting industries worldwide e.g. US Navy ship building, General Motors car leasing, shrimp aquaculture sustainable management

12 A Simple Model on HWTS Adoption and Sustained Use Initial population Advertising Word of mouth Drop off = 10,000 people = 2%/month = 1/10 of current adopter ratio = 10%/month

13 Hypothetical Model Example 1 Peaks 2200 month 20

14 Hypothetical Model Example 2 Peaks 3800 month 25 Drop-off fraction reduced to 5%/month

15 Model Insights By reducing drop-off fraction, (can be through better tech design, users follow-up, supply availability, etc): Main anticipated effect (blue line): Less drop-off more current adopters Unanticipated effect (red line): More current adopters stronger word-of-mouth quicker adoption Potential policy recommendation: Reduce drop off rate has leveraged effect

16 Model Insights BUT for this overly simplistic model, Total discontinuance behaviour for non-trivial parameter Need to change structure to change model behaviour Changing structure has MORE leverage than changing drop-off rate Nevertheless, The reality is complicated. Hundreds of linkages and sideeffects are incomprehensible for human mind. Need modeling tool, or risk producing counter-intuitive outcomes

17 Next Steps 1. Building a full model for household chlorination in Kathmandu city, Nepal Identify critical factors using qualitative research technique Integrate factors, considering inter-relationships and feedbacks Calibrate to historical adoption and sales records Simulate and analyze for leveraged policies 1. Repeat for ceramic filter in northern Ghana and biosand filter in southern India 1. Compare and contrast models

18 Expected Outcome 1. New approach to understand big picture of HWTS dissemination 1. Virtually simulate the effect of various strategies, identify leveraged policies, without need for actual experimentation (which can be difficult)

19 Conclusions Many critical factors influencing HWTS implementation, from different perspectives Lack a holistic and systematic approach, incorporating the views from various perspectives System dynamics modeling is a novel and promising tool