Three Essays on Nonparametric Hypothesis Testing and Stochastic Frontier Analysis

Green innovation (GI) plays a pivotal role in advancing sustainable transformation. To implement the concept of green development, China is vigorously advancing green innovation adoption (GIA). Although prior research has largely focused on the manufacturing sector, little is known about how ownership structures shape GIA in the construction industry, nor about its impact on cost. To address this gap, this study, grounded in the Technology–Organisation–Environment (TOE) framework, investigates the extent to which technological, organisational, and environmental factors influence the GIA in the construction sector and how GIA contributes to cost reduction, as well as how these effects differ between state-owned and private firms. Data were collected from 277 construction enterprises, and analysed using Partial Least Squares Structural Equation Modelling (PLS-SEM). Our findings show that GIA significantly reduces costs in both state-owned enterprises (SOEs) and private firms, while market pressure exerts no obvious influence on GIA. Notably, organisational support is found to negatively affect green process innovation in SOEs, suggesting that such initiatives may be more symbolic than substantive. This study could serve as a reference for businesses and governments, and contribute to China’s new development philosophy of “innovative, coordinated, green, open, and shared.”

Background: Efficient utilization of agricultural inputs is fundamental to improving farm productivity and economic performance in crop production systems. In input-intensive crops such as cabbage, the profitability of cultivation depends not only on the level of input use but also on their optimal allocation. Imbalances in the use of key inputs, particularly fertilizers and plant protection chemicals, can lead to diminishing returns and increased production costs, thereby reducing overall efficiency. Such inefficiencies are of particular concern in off-season vegetable production, where higher input intensity and market-driven incentives amplify the economic consequences of suboptimal resource use. Aims: The study was undertaken to analyse allocative and technical efficiency and their implications for resource use in off-season cabbage farming. Study Design: A farm-level study based on primary data collected through simple random sampling with proportional allocation. Methodology: The study was conducted using data from 80 off-season cabbage farmers. Cobb–Douglas production function was employed to estimate allocative efficiency, while stochastic frontier analysis (SFA) was used to estimate technical efficiency. A Tobit regression model was applied to identify the influence of socio-economic factors on technical efficiency. Results: The results revealed significant allocative inefficiency across farm categories. Seed was underutilized (2.63), while labor was overutilized (-2.79) while fertilizer was underutilized in marginal farms (4.51). In small farms, seed (-3.84) and labor (-4.07) were overutilized, whereas fertilizer (2.65) was underutilized. In the stochastic frontier model, γ = 0.9988 indicated that the composite error variance was largely attributable to the inefficiency component relative to random noise, while σ² = 0.00123 reflected low overall dispersion in deviations from the frontier. The mean technical efficiency was 97.5 per cent, indicating that farms operated close to the production frontier with limited variation. Tobit results showed that age, farm size, and labor availability were statistically insignificant, implying a homogeneous efficiency structure across farms. Conclusion: The results suggest that the primary scope for improving performance in off-season cabbage production lies in enhancing allocative efficiency through rationalization of fertilizer and plant protection chemical use, given the relatively high and homogeneous levels of technical efficiency across farms.

Group-specific stochastic production frontier models with parametric specifications

The stochastic frontier model was first proposed in the context of production function estimation to account for the effect of technical inefficiency. The inefficiency causes actual output to fall below the potential level (that is, the production frontier) and also raises production cost above the minimum level (that is, the cost frontier). Recent applications of the model are found in many fields of study including labour, finance, and economic growth. In these applications, the observed outcome (of wages, investment, and so on) is modelled as being deviating from a frontier level in one direction owing to factors such as information asymmetry.

This thesis examines factors affecting aggregate total factor productivity (TFP) growth in developing countries. While the first two essays are devoted to exploring trade-embedded foreign R&D spillovers, the last two essays are more concerned about firm dynamics and misallocation of resources across firms. Although foreign R&D spillovers have been studied intensively since the emergence of the innovation-led growth models in the early 1990s, little research has been conducted on the impact of Free Trade Agreements (FTAs) on North-South diffusion processes. Also, few studies have incorporated domestic R&D and trade liberalisation to examine North-South R&D spillovers at industry level. The first two essays fill these gaps. The last two essays contribute to the literature by partly addressing the long-lasting puzzle of why productivity dispersion in developing countries is large and persistent, even within narrowly-defined industries. In particular, these essays explore the role of firm dynamics, including entry and exit, and reallocation of resources across firms in reducing TFP gap in the manufacturing sector of a transitional economy, namely Vietnam. Given the recent proliferation of FTAs, the first essay investigates the impact of FTAs on North-South R&D spillovers. Using panel data covering 56 developing countries and 15 OECD countries over 1980-2008, the results show that North-South R&D spillovers are sizeable. FTAs have, however, a negative impact on this diffusion process. This finding may be attributed to decreases in the variety of goods imported, and the switching of imports from more efficient non-member countries to less efficient member countries. The second essay employs the industry-level data in Vietnamese manufacturing over 2000-2009 to address the foreign technology diffusion from 16 OECD countries. The empirical results indicate that the domestic own industry R&D improves TFP growth in the sector, but in a much smaller scale than the foreign R&D counterparts. The foreign R&D spillovers themselves differ in diffusion channels, as Vietnamese manufacturing’s TFP seems to benefit more from the foreign R&D embedded in the inter-industrial relation through the Input-Output tables than that embedded in import from the same industries abroad. Despite experiencing several major trade liberalisations in the manufacturing sector over the study period, the sectoral TFP has seen little impact from these reforms. The third essay studies the impact of market deregulation on the aggregate TFP of Vietnamese manufacturing. Over 2000-2008, Vietnam experienced a big market deregulation which induces a massive entry of private firms amid privatisation of state-owned enterprises (SOEs). The results reveal that during this period firms’ entry and exit makes up around a half of aggregate TFP growth. Private firms account for most of this TFP improvement as they exhibit a “growth” effect over the state entrants and a “level” effect over the foreign counterparts. Despite huge entry of private firms, SOEs face lower hazard rates than the private counterparts, even accounting for productivity differentials. This may be attributed to SOEs’ receipt of favourable treatment from the government that prevents them from exposure to the same degree of market experimentation as experienced by the private firms. The last essay examines whether firm dynamics improves allocative efficiency in Vietnamese manufacturing over 2000-2008. Given that this is a period of unprecedented growth in domestic credit supply, the essay also investigates the impact of the credit policy on capital misallocation across firms in the sector. It finds that firms’ entry and exit contributes significantly to the allocative efficiency improvement in Vietnamese manufacturing, particularly in the second half of the study period. The across-firm TFP dispersion within 4-digit ISIC industries is large and persistent because SOEs disproportionately absorb a lion’s share of credit. In that context, providing more credit to the SOEs, relatively to their private counterparts, would yield an anti-capital distortion reducing effect rather than the desired capital distortion reducing effect of commercial and subsidised credit.

Rice is one of the most important crops in terms of harvested area and food security both globally and for Colombia. Improvement of technical efficiency levels in rice production in order to close yield gaps in a context in which rice demand increases, natural resources are depleted, and where there are growing expectations about both climate changes and trade agreements is likely the most important challenge that farmers confront. This research assessed the main management factors that limit both rice crop productivity and the likely drivers of non-optimal technical efficiency levels (a proxy for yield gaps). This study focused on both upland and irrigated direct seeding systems across a variety of environments in Colombia. Stochastic frontier models were used to integrate microeconomic theory and empirical regression analysis in conjunction with a large commercial rice production database developed by the Colombian rice growers’ federation (Fedearroz). A large variation was found in technical efficiency (from 40 to 95%) levels for both upland and irrigated systems, and major differences were obtained in the limiting factors of the two systems (e.g. seed availability, variety type, market accessibility, fertilizer type, and use rate). This suggests both substantial and varied opportunities for improvements in current technical efficiency levels. Across systems, the correct choice of variety was identified as a common key factorfor maximizing yield for a particular environment. For upland systems, optimal choices were F174 and F2000, whereas for irrigated rice F473 was found to produce the highest yield. Additionally, numerical analysis suggests a yield impact of ca. 0.18% for each 1% increase in the nitrogen application rate for upland systems. For irrigated rice, phosphorous rather than nitrogen application rates were found to be more important. Since our analysis is based on farm-scale commercial production data, we argue that once our results are brought to consensus with local extension agents, technicians and agronomists, then management recommendations for closing yield gaps can be used to improve rice productivity.

Observed and Unobserved Heterogeneity in Stochastic Frontier Models: An Application to the Electricity Distribution Industry

Type II failure and specification testing in the Stochastic Frontier Model

An attempt has been made to investigate the online bank specific deposits efficiency using stochastic frontier technique and it intends to determine various factors affecting the efficiency level of banks for the period 2001 - 2007. We used a panel of 20 banks divided into four groups namely, NBs (National Banks), ISBs (Islamic Banks), FBs (Foreign Banks), and PBs (Private Banks) in Bangladesh. In this study, a comparison was made with the efficiency scores of banks group-wise, year-wise and individually. This study showed that the estimated year-wise average online banks deposits efficiency was 0.738 while group-wise average deposits efficiency was 0.777. At the bank group level, Nationalized Commercial Banks (NBs) and Islamic Banks (ISBS) were more efficient by 90.9 and 86.8% respectively, followed by Private Banks (PBs) which had 63.4% and Foreign Banks (FBs), 62.7%. We observed that the foreign banks were less efficient in producing deposits. However, the overall deposits efficiency of all bank groups steadily increased over time except in 2007. The most efficient bank was found to be Islami Bank Ltd. and the most inefficient bank was Pubali Bank with efficiency scores of 0.96 and 0.52, respectively. Key words: Deposits efficiency, stochastic frontier analysis, banking industry.

State-owned enterprises (SOEs) play an important role in Emerging Europe's economies, notably in the energy and transport sectors. Based on a new firm-level dataset, this paper reviews the SOE landscape, assesses SOE performance across countries and vis-a-vis private firms, and evaluates recent SOE governance reform experience in 11 Emerging European countries, as well as Sweden as a benchmark. Profitability and efficiency of resource allocation of SOEs lag those of private firms in most sectors, with substantial cross-country variation. Poor SOE performance raises three main risks: large and risky contingent liabilities could stretch public finances; sizeable state ownership of banks coupled with poor governance could threaten financial stability; and negative productivity spillovers could affect the economy at large. SOE governance frameworks are partly weak and should be strengthened along three lines: fleshing out a consistent ownership policy; giving teeth to financial oversight; and making SOE boards more professional.

This paper investigates the how effective protection and firm ownership affected firm productivity in Vietnam during 2005-2010. In labour-intensive industries and industries with intermediate labour intensity, the level of effective protection in an industry had a significantly negative effect on firm productivity. Multinational enterprise (MNE) joint ventures (JVs) and state-owned enterprises (SOEs) had consistently higher productivity than private firms, with productivity usually being highest in JVs. Wholly-foreign MNEs (WOs) also had significantly higher productivity than private firms in 2005-2007, but lower productivity than JVs or SOEs, and in 2008-2010, WO-private differentials were insignificant. In capital-intensive industries, the pattern of productivity differentials (highest in JVs, followed by SOEs, WOs, and private firms) was similar in the earlier period, but not in the latter period or when all years were included in the sample. The level of effective protection also did not have a significant, independent effect on firm productivity in capital-intensive industries.

Despite the critical role that state‐owned enterprises (SOEs) can play in the green transition, relatively little is known about governments' environmental behavior as foreign shareholders. This study examines the effect of state ownership on domestic and foreign environmental performance. We argue that the incentives associated with state ownership differ depending on whether SOEs operate domestically or abroad. To test this argument, we use random‐effects and logit models on a panel dataset containing information on 96 multinational oil and gas companies from 26 countries, along with their annual domestic and foreign CO 2 emissions for 2013 and 2020. The results indicate that, in domestic operations, SOEs pollute less than private firms in countries with lower levels of economic development, whereas the opposite pattern emerges in more developed economies. Second, SOEs with majority state ownership are more likely to reduce domestic emissions than private firms, regardless of the home country's level of development. Finally, we found that emissions under state and private ownership do not differ when operating abroad.

Currently local government debt is a major concern for all sectors of society, but there is still a debate about the relationship between local government debt and social welfare. This paper addresses the question of the impact of local debt and social welfare in China, using a case study of data from Zhejiang and Anhui from 2014-2019 to qualitatively analyze the impact of local debt and social welfare in four dimensions: consumption, infrastructure, education, and health. The results found that: first, the growth of local debt has an inverted U-shaped impact on residents' welfare, which is manifested in the fact that before the size of local debt reaches the critical value, the increase of local debt can enhance social welfare, but after exceeding the critical value, the increase of local debt makes social welfare decline; second, in terms of sub-welfare index factors, there are different correlation degrees between different factors and the indicators of local debt, and thus have different characteristics of development at different stages are characterized by different developments.

ABSTRACTMixed‐ownership reform (the Reform) is crucial for the construction of a modern market economy. Through a proprietary data set of 32 provinces in China, we find that government intelligent transformation (GIT) can attract private firms acquiring local state‐owned enterprises (SOEs), and such link is stronger among target firms in central and western regions, and acquirers with lower competitive positions and higher levels of digital transformation. Additionally, we also find this effect exists both in local and nonlocal private firms. Further tests suggest that GIT can regulate government behaviors, reduce relational transactions, and enhance local policy environment quality and the level of local market legal governance, thereby promoting private firms to participate in the Reform. Moreover, in the provinces with higher levels of GIT, there is a greater likelihood that private firms will gain control rights over SOEs and obtain more resources following the Reform. Our findings may have policy implications for the world's largest emerging market.

The purpose of this paper is to analyse the efficiency of the use of foreign exchange reserves (FER) by 21 nations over 5 years across the globe. The study follows an analytical research design to analyse the efficiency levels in the use of foreign exchange reserves using stochastic frontier analysis (SFA). Three models under SFA, that is, Cobb–Douglas production frontier for panel data with half‐normal distributions, translog production frontier with panel data assuming truncated normal distributions and total technical efficiency effects model with panel data, have been used to study efficiency levels. The research uses the nominal value of exports and foreign direct investment as inputs and foreign exchange reserves as output. The results indicate that out of 21 nations, developing countries are found to be more efficient in comparison to developed countries in the use of FER.