The Role of Modeling and Scenario Development in Long-term Strategies
The use of strong analytical frameworks and tools, in most cases models, can facilitate the development of robust and comprehensive long-term strategies. Models, as simplified representations of reality, are key to long-term policy formulation, and implementation in the near-term, consistent with a country’s socioeconomic development priorities. It’s nearly impossible to deal with the reality of developing and implementing these long-term strategies due to all the complexities that come with this. Models are therefore used to simplify these complexities in order to generate some understanding. For example, in establishing a baseline reference for projecting future policies and implementation options in specific sectors, developing monitoring and reporting indicators, models can be very useful.
In examining the role of modeling and scenario development in long-term development planning, it is instructive to bear in mind that these are essentially analytical tools for describing future outlook, using current information/datasets and assumptions. These datasets and assumptions could be incomplete or completely wrong, making models largely imperfect, and thus cannot be heavily relied on in dealing with the complexities of reality. Long-term strategies that depend on modeling results alone can be severely flawed, and should be avoided!
This essay discusses the fundamental roles of modeling as a tool for developing long-term strategies and presents some of the limitations and uncertainties that these tools come with. Within this context, the essay also examines how modeling and research results can inform decision-making processes.
1. Primary functions of modeling as a tool and method for developing long-term strategies
1.1. Linking sectoral and economy-wide models in the context of sustainable development and climate
Sectoral and economy-wide policies are both needed to properly plan and implement sustainable development goals in a changing climate. For example, in the agricultural sector, the effects of policies and broader exchange-rate polices on productivity both need to be considered and analyzed. The impacts of these policies and shocks within sectors and the broader economy can be assessed with modeling tools. Quantitative models have been used by national governments to design policies and measures that form the basis of short-term decision-making processes leading to implementation of long-term development goals.
Typically, economy-wide modeling is required to analyze different development strategies whenever the outcomes of interest (e.g., an SDG target) are influenced significantly not only by the direct effects of single policies but also by indirect effects and policy interactions that feed back into the processes that determine these outcomes. Many studies use a sectoral model (e.g., land use or electricity or transport) and link with economy-wide models to get an understanding of the combined economy-wide performance. For example, the United Nations Environment Programme conducted a modeling exercise to test the hypothesis that investing in the environment delivers positive macroeconomic results (as well as environmental ones), as part of its work on the 2011 Green Economy Report. The team used the Threshold 21 World Model (T21-World), which comprises several sectoral models integrated into a global model. The sectoral models were seen as the core of the modeling exercise supporting the team’s analysis, which enabled evaluation of the effects of investing, in various amounts of GDP, in green economic activities to stimulate economic growth, improve resource efficiency, lower emissions, create decent jobs, and so on.
1.2. Target setting and multidimensional analysis
Modeling can be represented as a magnifying glass that depicts all the details of intended actions targeted at the future. National governments have achieved greater successes in economic and industrial growth, poverty reduction, employment, emission abatement, and natural resource protection by setting concrete short and medium-term targets using multidimensional analysis/modeling. Setting very clear economy-wide and sector-specific targets is an important step toward translating short and medium-term goals into long-term development priorities. Achieving these targets requires developing metrics and indicators that describe the various trajectories that must be pursued toward the long-term goals, and this is done using multidimensional analysis—that is, economic and technical analysis.
Another way to ensure accurate assumptions and enhance the reliability of future predictions is to establish both long- and medium-term economy-wide targets as well as short-term sector-specific targets. Also, long-term targets can ensure strategic direction, while short-term targets can guide concrete actions and achieve immediate benefits.
1.3. Creating storylines that link abstracts with narratives
Developing and using storylines of innovative development pathways and scenarios can be a great way to establish a comprehensible link between the world of modeling (abstract) and the narratives (the world of policymaking). Since models are usually abstractions of the natural system, the uncertainty (i.e., the lack of exact knowledge) of their outcomes needs to be taken into consideration when they are utilized for decision-making. This means that robust decision analysis should be done with both sectoral and economy-wide modeling to give decision-makers as realistic a picture as possible of the current knowledge and its deficiencies by drawing on all the available relevant information. The information could be data, expert opinion, or models and tools with technological detail for specific key sectors and economy-wide options.
2. Limitations of modeling and quantitative projections in long-term strategies
2.1. Models are not always perfect!
Although modeling and scenario development are based on assumptions and thus often imperfect, they are still useful. Modeling outcomes depend heavily on available information and/or data, which may not necessarily be as reliable as they should be. Inasmuch as models can be used to generate scenarios with different possible outcomes and provide a platform for comparison of options, these may only partly reflect reality. Reality is complex; it consists of strongly or weakly related events, and this makes modeling reality and quantitative projection into the future complex.
2.2. The uncertainty and risks of using quantitative and qualitative models as tools for producing long-term development pathways
Quantitative modeling is very useful in analyzing complex interactions that lead to an understanding of future impacts, an understanding that would not be achieved easily with other methods. They are useful in revealing the unintended consequences of actions and/or policy implementation. Qualitative models complement these quantitative results and are needed to set the stage for the complete storyline. Developing both of these modeling approaches and incorporating explicit development metrics that evolve with the development challenges can produce useful future results.
However, these approaches have in them inherent risks and limitations, particularly when used to produce long-term development pathways, if or when they are not well managed.1 First, these models are developed using data and assumptions. These datasets and/or their sources may not be right, or the assumptions made in developing the models could change with time, and once this happens, the strategies developed with the models will not yield the needed impacts or will simply be misleading. Second, decision-makers depend greatly on results produced by planners. There is, however, the risk of providing them with false or “overhyped” results, which may adversely affect decision-making. Given the wide range of possible outcomes resulting from the scenarios or models used, the supposed “line of best fit” may not be the best. Beyond using the “line of best fit” approach, the integration of the wide range of model results into long-term strategies can be further enhanced through model quantification—that is, quantifying the impact of the model risks through analysis of data deficiencies, assessment of estimation uncertainty and potential misuse of the model.
In essence, the quality and reliability of the output—in other words, the strategies—depends on the accuracy of the data and how they are analyzed. This is to say, you get what you put in!
2.3. Limitations of modeling long-term cost estimates
Most, if not all, projections don’t end up going exactly as foreseen. Slight or even major changes are bound to set in along the way, causing the expected outcomes to differ. When it comes to cost estimates, the shorter the term, the more accurate the estimate and vice versa. As such, cost estimates for long-term strategies are one of the biggest limitations associated with modeling. Inflation, devaluation of products or services, decline of the country’s currency resulting from economic crisis, new policies on taxes, increase in fuel prices, and so on are all developments that could affect cost estimations made in the past. The occurrence of any of these directly or indirectly shadows the proposed long-term strategy. To make up for these foreseen but “difficult-to-perfectly-predict” changes, most economic analysts factor in contingencies (i.e., take into account factors and/or policies to mitigate the effects of the occurrences, e.g., provide guarantees, legal and regulatory frameworks, incentives, etc.) in the economic analysis of the proposed long-term strategy. Although it may not completely solve the possibility of unexpected outcomes, this minimizes the effects.
2.4. Unbiased modeling is a useful tool for informing long-term strategies
The bigger issue is whether or not modeling results are biased in ways that could lend unjustifiable support to certain pathways or levels of ambition. The interest of the person performing the analysis or research work undoubtedly affects the type, form, and even number of scenarios he or she may employ. All these variables influence the final outcome conveyed to leaders to help them make crucial decisions.
Appropriate trainings on collection of the right data, participatory approaches, exemplary case studies, and peer review of the analysis can all help bring about unbiased modeling, which in turn supports the development of reliable long-term strategies.
3. Linking research outputs to decision-making
3.1. Collaboration between research and policymaking
Quality and independent research outputs can make a huge difference by providing policy-relevant analyses to inform decision-making processes. When the research community speaks to and collaborates with the different groups of decision-makers and stakeholders to understand their specific needs, this leads to good and relevant research outputs, which help shape good policies.
3.2. Providing independent research results to the decision-making process
Typically, much of the modeling work by researchers dwells on data use, which is crucial since it is at the heart of developing models. Once the data are properly validated, researchers can use them to develop suitable models, which can be utilized by policymakers. There is thus a strong relationship between the production of scientific knowledge and its use in policy formulation and implementation. The interrelations between researchers and decision-makers have been considered a prime factor in analyzing knowledge transfer processes. Research outcomes and the subsequent uptake by policymakers depend largely on certain characteristics of the actors, that is, on the researchers’ behavior and the decision-makers’ receptiveness.
3.3. Highlighting crucial research outputs not considered in decision-making and formulation
Some aspects of research outputs, although crucial, are not often considered in decision-making processes by some national governments, like path dependencies (i.e., linking the present with the past) and hedging strategies in an adverse context. Hedging strategies are risk- management strategies that enable policymakers to navigate complex uncertainties without having to fix interventions in advance. Once the right path dependency is created, combined with the right hedging strategies, policymakers can address issues like whether it is safe to take a decision at any time in the context of existing uncertainties, or whether it is possible, or even more effective, to delay the decision until a reasonable level of certainty is achieved. Researchers can greatly improve the implementation processes through collaboration by raising these issues with decision-makers. There are also issues of global enabling conditions, which can be shaped in each local context.
In short, researchers should improve communication of their work to various kinds of decision-makers to better explain what the analyses can do and how they should or should not be interpreted.
1 Model risk is the potential for unintended consequences of decisions based on incorrect or misused model outputs and reports to occur. Ways to manage model risks and limitations include the following:
- Strengthening governance of data and developing data quality assurance functions to protect the quality, integrity, and consistency of data
- Conducting effective, critical, and independent validation of models
- Pilot testing model results before full deployment