📈 economics

Sectoral Productivity and Structural Transformation in Transition Economies: Panel Evidence from China and the Former Soviet Union

This paper analyzes panel data from China and former Soviet Union countries to demonstrate that while their distinct initial industrial structures led to divergent de-industrialization patterns (inverse U-shaped vs. U-shaped), sector-specific productivity growth, rather than inter-sectoral productivity gaps, serves as the primary driver of structural transformation in both regions.

Original authors: Taehyun Ryu

Published 2026-08-13

📖 4 min read☕ Coffee break read

CC BY 4.0

Original authors: Taehyun Ryu

Original paper licensed under CC BY 4.0 (https://creativecommons.org/licenses/by/4.0/). ✨ This is an AI-generated explanation of the paper below. It is not written or endorsed by the authors. For technical accuracy, refer to the original paper. Read full disclaimer

Imagine the economy of a country as a giant, bustling kitchen. In this kitchen, there are three main stations: the garden (where food is grown), the stove (where raw ingredients are cooked into meals), and the dining room (where people chat, get served, and enjoy the experience). For a long time, economists believed that as a kitchen gets richer and more successful, the workers naturally move from the garden to the stove, and eventually, as the stove gets super-efficient, most workers end up in the dining room. This idea is called "structural transformation." It's the story of how a society shifts from farming to making things, and finally to providing services. But what happens if a kitchen was forced by a strict manager to have way too many chefs and not enough gardeners, or vice versa? Does the usual story still work? This is the big question researchers ask when looking at countries that are switching from strict, government-run kitchens to free-market ones. They want to know if the "rules of the kitchen" change when the starting point is weirdly unbalanced.

This paper, written by Taehyun Ryu, dives into two massive, real-life experiments to answer that question: China and the countries of the former Soviet Union (FSU). Think of these two groups as two different kitchens that were forced to change their recipes at the same time. In the late 1970s, China started its transition with a kitchen that was mostly a garden; they had very few chefs and almost no dining room. They were "under-industrialized." On the other hand, when the Soviet Union broke up in the early 1990s, its countries were the opposite: they were "over-industrialized." They had been forced to build massive factories and cook huge meals, even if no one was hungry for them, leaving their gardens and dining rooms neglected.

The author uses data from 31 Chinese provinces and 15 former Soviet countries to see how their kitchens evolved over time. The study finds that the "rules of the kitchen" are actually quite flexible. In China, as the provinces got richer, the share of workers and money in the "stove" (manufacturing) went up and then came down, creating a classic hill shape. This fits the traditional story: they built up their factories, and then naturally shifted toward services as they got wealthier. However, the former Soviet countries did something totally different. Because they started with too many factories, their "stove" share actually dropped sharply at first (a process called de-industrialization) and then slowly started to rise again as they got richer, creating a "U" shape. It's like a kitchen that had to fire half its chefs just to get back to a normal size before it could start growing again.

The most surprising discovery, however, is why these shifts happened. The paper suggests it wasn't just about the difference in how fast one station grew compared to another (a theory known as "Baumol's cost disease," which says services are slower to improve than factories). Instead, the main driver was how fast each station improved on its own. When a specific station (like the stove) got better at cooking (higher productivity), it actually needed fewer workers to do the same amount of work, even though it produced more food. This freed up workers to move to other stations. The study shows that in both China and the Soviet countries, the speed of improvement within each specific sector was the real engine driving the changes in who worked where and how much money each part of the economy made.

So, what does this tell us? It suggests that if a country wants to change its economic structure, simply trying to force workers into a specific industry might not work. Instead, the key is to make each part of the economy better at what it does. When a sector gets more productive, the structure of the whole economy naturally adjusts. For countries that started with a weird imbalance—like having too many factories or too few—the path to a healthy economy looks different than for countries that started from scratch, but the ultimate goal is the same: a kitchen where the workers and the food are in the right places, driven by the speed of improvement in every corner of the room.

Technical Summary: Sectoral Productivity and Structural Transformation in Transition Economies

Problem Statement
This study investigates the drivers of structural transformation in economies transitioning from planned to market systems, specifically comparing 31 Chinese provinces (1978–2018) and 15 Former Soviet Union (FSU) countries (1991–2018). The central problem addresses the "structural distortion" inherent in centrally planned economies, where industrial structures were driven by policy objectives (e.g., heavy industry prioritization) rather than market demand or comparative advantage. This resulted in two divergent starting points: China began with an underdeveloped, agrarian base, while FSU countries began with an artificially "over-industrialized" structure. The paper seeks to determine how these initial conditions, combined with sectoral productivity growth and income levels, shaped the trajectory of structural change (shifts in output and employment shares) during the transition period. It challenges standard de-industrialization models, such as Baumol's cost disease, which rely on fixed inter-sectoral productivity gaps, by testing whether sector-specific productivity growth rates are the primary drivers of structural reallocation.

Methodology
The research utilizes a panel data approach covering two distinct regions with different transition histories.

Data Sources: The study employs secondary macroeconomic data from CEIC (Chinese provinces), World Bank WDI, IMF World Economic Outlook, Penn World Tables 9.0, and EBRD Transition Indicators.
Variables:
- Dependent Variables: Sectoral output shares (GDP or Value-Added) and employment shares for primary, secondary, and tertiary sectors.
- Independent Variables: Sector-specific productivity indices (calculated as output per worker relative to the previous year), GDP per capita (and its squared term to capture non-linearities), and a vector of control variables (investment, FDI, trade openness, human capital, unemployment, inflation, and transition indices).
Theoretical Framework: The paper adapts the three-sector hypothesis and de-industrialization models (Fisher, Clark; Rowthorn & Wells; Baumol). It modifies standard models to allow for variable sectoral productivity growth rates ( $\lambda_a, \lambda_i, \lambda_s$ ) rather than assuming a fixed gap where service sector productivity lags.
Estimation Techniques: To address heteroskedasticity, serial correlation, and the compositional interdependence of sectoral shares, the study employs:
1. Feasible Generalized Least Squares (FGLS): Specifically with panel-specific first-order autocorrelation (PSAR1) to handle time-series properties.
2. Fixed Effects (FE): To control for unobserved time-invariant heterogeneity across provinces/countries.
3. Seemingly Unrelated Regressions (SUR): To account for the correlation between error terms of the three sectoral equations.
  Robustness checks include the inclusion of additional macroeconomic and policy control variables.

Key Results
The empirical analysis yields distinct patterns for output shares versus employment shares and between the two regions:

Sectoral Productivity as a Driver:
- Output Share: Higher productivity growth within a specific sector is positively associated with that sector's share of total output.
- Employment Share: Higher productivity growth within a sector is negatively associated with that sector's share of employment. This suggests that productivity gains allow sectors to maintain or grow output with fewer workers, driving labor reallocation.
- Implication: This finding departs from Baumol's cost disease (which attributes structural change to relative productivity levels or gaps) and supports the view that the growth rate of sectoral productivity is the primary mechanism for structural change.
Income and Structural Change (The Shape of Transformation):
- Chinese Provinces: Exhibit an inverse U-shaped relationship between GDP per capita and the secondary sector's output/employment share. This aligns with the standard three-sector hypothesis, where industrialization peaks and then declines as the economy matures.
- FSU Countries: Exhibit a U-shaped relationship between GDP per capita and the secondary sector's output share. This reflects the "over-industrialization" of the starting point; as transition progressed, the secondary sector share initially collapsed (de-industrialization) before stabilizing or rising as the structure corrected toward market equilibrium.
- Employment: Both regions show an inverse U-shaped relationship between income and secondary sector employment share, indicating that manufacturing employment peaks at lower income levels in these transition contexts.
Robustness: These findings hold across FGLS, FE, and SUR estimations and remain consistent when controlling for investment, trade, human capital, and transition-specific policy indicators.

Key Contributions

Re-evaluating Structural Drivers: The paper provides panel evidence that sector-specific productivity growth (rather than static inter-sectoral gaps) is the dominant driver of structural transformation in transition economies.
Initial Conditions Matter: It demonstrates that the trajectory of structural change is heavily contingent on the initial industrial structure. The "over-industrialized" starting point of FSU countries led to a U-shaped adjustment path, contrasting with the standard inverse U-shape observed in China's development from an agrarian base.
Refining De-industrialization Theory: The study contributes to the literature on de-industrialization by showing that in transition economies, the process is often a correction of policy-induced distortions rather than solely a result of technological progress or globalization. It highlights that structural change can be growth-reducing if workers relocate to less productive sectors, a phenomenon observed in the FSU adjustment.

Significance and Claims
The paper claims to offer a systematic panel-data account of the sectoral productivity mechanisms behind atypical structural trajectories in transition economies. It argues that industrial policies targeting sectoral productivity growth are more likely to yield durable structural change than policies targeting sector shares directly. The study concludes that understanding structural transformation requires accounting for initial distortions; the "correction" of these distortions (as seen in the FSU) follows a different mathematical and economic path than organic industrialization (as seen in China). The findings support the emerging literature suggesting that structural transformation is a consequence of productivity reallocation and within-sector growth, rather than a policy target in itself. The paper explicitly limits its scope to the transition period ending in 2018, noting that post-2018 geopolitical and pandemic disruptions represent a structurally distinct episode outside the transition framework.

Technical Summary: Sectoral Productivity and Structural Transformation in Transition Economies

More like this