When evaluating economic development, public health, or environmental sustainability across different regions, a central question emerges: Are lagging regions catching up to the leaders, or are the gaps between them widening? To answer this, researchers rely on two distinct but complementary mathematical frameworks: $\beta$-convergence and $\sigma$-convergence. Together, these tools allow us to diagnose whether a system is naturally moving toward equity or if structural barriers require policy intervention.
Beta Convergence: The Catch-Up Mechanism
The intuition behind $\beta$-convergence is rooted in the concept of diminishing returns. A region starting from a lower baseline often has more room for rapid growth because it can adopt existing technologies, benefit from targeted infrastructure investments, or experience higher marginal returns on capital.
Mathematically, $\beta$-convergence is tested by regressing the growth rate of a specific metric over a period against its initial baseline level:
$$\ln\left(\frac{Y_{i,t}}{Y_{i,0}}\right) = \alpha + \beta \ln(Y_{i,0}) + \varepsilon_i$$
Where $Y_{i,0}$ is the initial value of the metric for region $i$, and the left side of the equation represents its growth rate over time. If the coefficient $\beta$ is negative and statistically significant, it proves that entities with lower initial values are growing faster than their well-off counterparts—the definitive signature of a "catch-up" effect.
This framework is typically applied in two ways:
- Absolute Convergence: This assumes all regions are heading toward the exact same steady state. For instance, Di Wang et al. used this approach in PER CAPITA INCOME CONVERGENCE AMONG PROVINCES IN CHINA to demonstrate how national economic stimulus packages accelerated income catch-up across provinces, allowing poorer regions to grow faster than wealthier ones.
- Conditional Convergence: This acknowledges that different regions have unique structural constraints, such as population density or economic capacity. As explained by Wei Tang et al. in Dynamic Evolution and Convergence Analysis of the Ecological Efficiency of China’s Fisheries, conditional convergence controls for these exogenous variables to see if regions are catching up to their own localized potential. Similarly, Bo-Wen An et al. demonstrated in Assessing green innovation efficiency and spatial characteristics of China’s high tech industry based on the three stage undesirable SBM model that filtering out external environmental "noise" reveals a much clearer picture of lagging regions overcoming structural disparities through targeted support.
Sigma Convergence: Measuring the Overall Spread
While $\beta$-convergence tells us if the underdogs are running faster, it does not guarantee that the overall field is getting closer together. For example, a lagging region might leapfrog a few places in the rankings without reducing the total variance of the entire system, or new outliers might emerge at the top. To measure whether the actual inequality within a system is shrinking, researchers turn to $\sigma$-convergence.
As highlighted by Jay Pan et al. in Disparity and Convergence: Chinese Provincial Government Health Expenditures, $\sigma$-convergence focuses on dispersion-based analysis. Instead of tracking individual growth trajectories, it tracks statistical indicators of the entire population over time, such as the coefficient of variation ($CV$) or the standard deviation of logarithmic values:
$$\sigma_t = \sqrt{\frac{1}{N}\sum_{i=1}^N \left(\ln Y_{i,t} - \bar{\ln Y_t}\right)^2}$$
A declining trend in $\sigma_t$ over time signifies $\sigma$-convergence—meaning the overall gap between the highest and lowest performing regions is actively narrowing. Wei Tang et al. leveraged this metric to show how the performance gap in fishery ecological efficiency stabilized and shrank over a fifteen-year period, providing clear evidence of a reduction in systemic inequality.
The Spatial Dimension: Geography Matters
Traditional convergence models treat regions as isolated data points. However, geography dictates that regions are deeply interconnected through trade, migration, and technology transfer. A booming province creates a "spillover effect" that naturally pulls its neighbors upward, meaning that a region's growth is rarely independent of its location.
To capture this reality, modern researchers integrate spatial econometrics into convergence analysis. Spatial $\beta$-convergence models modify the standard regression by adding a spatial weight matrix ($W$), which mathematically weights the influence of neighboring regions' growth rates and characteristics.
Guixiang Zhang applied this spatial lens in Dynamic Evolution, Regional Differences, and Spatial Spillover Effects of Urban Ecological Welfare Performance in China from the Perspective of Ecological Value, proving that urban welfare performance is heavily dependent on spatial dependencies. Furthermore, Zhouzhou Lin et al. combined both $\sigma$ and spatial $\beta$ models in Measurement, Regional Disparities, and Spatial Convergence in the Symbiotic Level of China's Digital Innovation Ecosystem to map how digital ecosystems evolve across geographic clusters.
This spatial perspective helps explain why convergence happens: as shown by Yue Zhang et al. in Research on the Realization Mechanism of Promoting Provincial Common Prosperity in China through Urban-Rural Integration Development, breaking down institutional barriers between areas allows development benefits to spill over across borders, gradually eroding historical regional inequalities despite different starting points.
Key Foundational Work
PER CAPITA INCOME CONVERGENCE AMONG PROVINCES IN CHINA (2020) — contributed 3 concept elements to this direction.
Disparity and Convergence: Chinese Provincial Government Health Expenditures (2013) — contributed 3 concept elements to this direction.