What would change if we could track poverty in real time and had accurate, up-to-date poverty estimates at our fingertips? The poverty rate is a crucial indicator for monitoring the development of our countries and for the design and evaluation of public policies. However, official poverty rates are derived from household surveys that often have limited frequency, problems associated with fieldwork limitations, and substantial delays in processing and publication.
To address this challenge, nowcasting techniques offer an innovative way to forecast poverty. This approach empowers countries with timely and reliable data, facilitating well-informed policy decisions based on up-to-date indicators. This could support policy decision-making in countries, enabling a timely response that improves the quality of life for hundreds of thousands of people in the region.
Thanks to the support of the Inter- American Development Bank (IDB), Central America, Panama, and the Dominican Republic now have a new tool that utilizes nowcasting techniques to forecast poverty.
Diverse economies, varied poverty: Understanding the region’s unique challenges.
In countries with low-income levels or subject to substantial changes in their social situation, such as economic crises, closely and promptly tracking poverty becomes increasingly important. In the Central America, Panama, and Dominican Republic region, an examination of income levels reveals a diverse range of development stages.
For instance, the region encompasses both one of the highest GDP per capita figures in Latin America and the Caribbean, and includes a nation with one of the lowest, highlighting the vast economic diversity present within the region. Figure (1) captures the dynamic nature of poverty and extreme poverty across these countries, illustrating diverse trends and levels. For example, El Salvador, Guatemala, and Honduras consistently show the highest percentages of poverty and extreme poverty within the region.
Moreover, the composition of households’ income also plays a crucial role in understanding poverty dynamics. For most of the countries in the region, labor income is the primary source of earnings, highlighting its critical role in poverty estimations, especially amidst significant fluctuations in employment.
Figure 1: Evolution of poverty in Central America
Innovating poverty estimations: A comprehensive nowcasting strategy
In this region, the challenge lies in estimating poverty rates for each country, considering the unique characteristics of its labor market while ensuring replicability. In this context, nowcasting techniques based on microsimulations models emerge as a promising solution, offering an innovative approach to estimate poverty and adapt to the specific economic and social dynamics of each country reliably and efficiently, to replicate between countries and across time.
The Inter-American Development Bank (IDB) through the Country Department of Central America, Haiti, Mexico, Panama, and the Dominican Republic has been working to develop a new nowcasting approach to forecast poverty.
In the first step, micro-simulation techniques are used on data from official household surveys to forecast individual and household characteristics. In the second step, these micro projections are integrated within a labor income generation model, leveraging on macro projections of key labor market variables. This integrated model is then used to predict the poverty rate. Overall, this comprehensive methodology aims to provide a more accurate and dynamic forecast of poverty, incorporating both individual and macroeconomic considerations.
Beyond traditional models: Groundbreaking insights into poverty dynamics
The model was applied to the Central America, Panama, and the Dominican Republic region and demonstrated robustness and accuracy. We tested the results against historical data from 2000 to 2020. We meticulously compared the model’s forecasts with the observed poverty change (figure 3), and the model aligns closely with actual poverty rate changes over the two decades and surpasses the predictive accuracy in the literature. Moreover, it outperforms the fit of other methods that solely rely on direct imputations from GPD from households’ income.
Figure 3: Observed and estimated changes in poverty rates
Our model also demonstrates superior performance in simulating labor markets shocks on poverty rates, such as those experienced in 2020, as the microsimulations allow to capture the effect of unemployment on poverty.
Such precision underscores the model’s utility in periods of economic volatility, providing a more accurate lens to view and anticipate the distributional consequences of labor market shocks. Nevertheless, the model can be further enhanced by integrating simulations of non-labor income sources.
Replicability and institutional support for nowcasting poverty
One of the main advantages of our approach is its ease of replication across various countries and time periods, unlike many other microsimulation techniques. This model is more complex to build with respect to the existent literature and yet it is easy to replicate across regions and countries. It uses widely available and standard macroeconomic inputs and can be replicated in a practically automated manner.
With this purpose in mind, the IDB offers this tool to regional authorities to enhance accessibility to poverty predictions for our nations, as we did with our nowcast economic activity tool.
Looking ahead: The future of poverty prediction and policy making
In conclusion, the innovative microsimulation model developed for nowcasting poverty rates in Central America, Panama, and the Dominican Republic region represents a significant step forward in obtaining accurate and timely poverty data. By modeling the labor markets’ decisions and translating them into macroeconomic figures, this model enhances our understanding of poverty trends and provides key data for monitoring and evaluating public policies. The model’s success in outperforming traditional methodologies while being replicable for a large set of countries highlights its useful applications.
This is a powerful tool for governments and policymakers, providing them with reliable data for prompt decision-making to reduce poverty in the region.
Do you want to learn more about Nowcasting poverty? Download our publication.
Thanks for sharing this post, very interesting subject. I wonder how the proposed technique differs from the poverty mapping methodology developed by Elbers and Lanjouw back in the 2000’s. This methodology also builds upon the income data gathered on household surveys and the household characteristics compiled in Census data. Then, it models an income level estimator using information from other sources. I wonder what the main differences are and how the nowcasting adds to that conversation.
Many thanks!