Causal Inference in Econometrics

We develop a method to validate the use of Markov Switching models in modelling time series subject to structural changes. Particularly, we consider multivariate autoregressive models subject to Markov Switching and derive close-form formulae for the spectral density of such models, based on their autocovariance functions and stable representations. Within this framework, we check the capability of the model to capture the relative importance of high- and low-frequency variability of the series. Applications to U.S. macroeconomic and financial data illustrate the behaviour at different frequencies.

Objective of this paper is to highlight the rapid assessment (in a few minutes) of fractionally differenced standard long memory time series in terms of parameter estimation and optimal lag order assessment. Initially, theoretical aspects of standard fractionally differenced processes with long memory and related state space modelling will be discussed. An efficient mechanism based on theory to estimate parameters and detect optimal lag order in minimizing processing speed and turnaround time is introduced subsequently. The methodology is extended using an available result in literature to present rapid results of an optimal fractionally differenced standard long memory model. Finally, the technique is applied to a couple of real data applications illustrating it’s feasibility and importance.

The stars appear to be aligned for a sustained effort to improve information to rural development policy makers about the impact space has on the opportunities for development of the ubiquitous smallholder households in rural areas of Southeast Asian countries. The influences of spatially heterogeneous resource constraints on farming activities, distance to markets and institutions, and spatial interaction among smallholders can now be better accounted for in modelling work as a result of improvements in analytical methodologies, the growing availability of so-called ‘big data’ and access to spatially defined information in panel data sets. The scope for taking advantage of these advances is demonstrated with two examples from a Southeast Asian country, the Philippines: spillovers and neighbourhood effects in impact studies and the development of sophisticated spatial stochastic frontier models to measure and decompose productivity growth on smallholdings.

Popular yield curve models include affine term structure models. These models are usually based on a fixed set of parameters which is calibrated to the actual financial market conditions. Under changing market conditions also parametrization changes. We discuss how parameters need to be updated with changing market conditions so that the re-calibration meets the premise of being free of arbitrage. We demonstrate this (consistent) re-calibration on the example of the Hull–White extended discrete-time Vasiček model, but this concept applies to a wide range of related term structure models.

Trade duration and daily range data often exhibit asymmetric shape with long right tail. In analysing the dynamics of these positively valued time series under autoregressive conditional duration (ACD) models, the choice of the conditional distribution for innovations has posed challenges. A suitably chosen distribution, which is capable of capturing unique characteristics inherent in these data, particularly the heavy tailedness, is proved to be very useful. This paper introduces a new extension to the class of Weibull distributions, which is shown to perform better than the existing Weibull distribution in ACD and CARR modelling. By incorporating an additional shape parameter, the Weibull distribution is extended to the extended Weibull (EW) distribution to enhance its flexibility in the tails. An MCMC based sampling scheme under a Bayesian framework is employed for statistical inference and its performance is demonstrated in a simulation experiment. Empirical application is based on trade duration and daily range data from the Australian Securities Exchange (ASX). The performance of EW distribution, in terms of model fit, is assessed in comparison to two other frequently used error distributions, the exponential and Weibull distributions.

In many real-life situations, when there is a need for a spending cut, this cut is performed in an across-the-board way, so that each budget item is decreased by the same percentage. Such cuts are ubiquitous, they happen on all levels, from the US budget to the university budget cuts on the college and departmental levels. The main reason for the ubiquity of such cuts is that they are perceived as fair and, at the same time, economically reasonable. In this paper, we perform a quantitative analysis of this problem and show that, contrary to the widely spread positive opinion about across-the-board cuts, these cuts are, on average, very inefficient.

In many situations, we have an (approximately) linear dependence between several quantities: $$y\approx c+\sum \limits _{i=1}^n a_i\cdot x_i$$y≈c+∑i=1nai·xi. The variance $$v=\sigma ^2$$v=σ2 of the corresponding approximation error $$\varepsilon =y-\left( c+\sum \limits _{i=1}^n a_i\cdot x_i\right) $$ε=y-c+∑i=1nai·xi often depends on the values of the quantities $$x_1,\ldots ,x_n$$x1,…,xn: $$v=v(x_1,\ldots ,x_n)$$v=v(x1,…,xn); the function describing this dependence is known as the skedactic function. Empirically, two classes of skedactic functions are most successful: multiplicative functions $$v=c\cdot \prod \limits _{i=1}^n |x_i|^{\gamma _i}$$v=c·∏i=1n|xi|γi and exponential functions $$v=\exp \left( \alpha +\sum \limits _{i=1}^n \gamma _i\cdot x_i\right) $$v=expα+∑i=1nγi·xi. In this paper, we use natural invariance ideas to provide a possible theoretical explanation for this empirical success; we explain why in some situations multiplicative skedactic functions work better and in some exponential ones. We also come up with a general class of invariant skedactic function that includes both multiplicative and exponential functions as particular cases.

Out of many possible families of probability distributions, some families turned out to be most efficient in practical situations. Why these particular families and not others? To explain this empirical success, we formulate the general problem of selecting a distribution with the largest possible utility under appropriate constraints. We then show that if we select the utility functional and the constraints which are invariant under natural symmetries—shift and scaling corresponding to changing the starting point and the measuring unit for describing the corresponding quantity x— then the resulting optimal families of probability distributions indeed include most of the empirically successful families. Thus, we get a symmetry-based explanation for their empirical success.

In this paper, the class of multivariate extended skew normal distributions is introduced. The properties of this class of distributions, such as, the moment generating function, probability density function, and independence are discussed. Based on this class of distributions, the extended noncentral skew chi-square distribution is defined and its properties are investigated. Also the necessary and sufficient conditions, under which a quadratic form of the model has an extended noncentral skew chi-square distribution, are obtained. For illustration of our main results, several examples are given.

In this paper, the multiple directional dependence between response variable and covariates using non-exchangeable copulas based regression is introduced. The general measure for the multiple directional dependence in the joint behavior is provided. Several multivariate non-exchangeable copula families including skew normal copula, and the generalized Farlie-Gumbel-Morgenstern copula models are investigated. For the illustration of main results, several examples are given.

In this paper, the characterization of the joint distribution of random set vector by the belief function is investigated. A routine of calculating the bivariate coarsening at random model of finite random sets is obtained. In the context of reliable computations with imprecise data, we show that the maximum likelihood estimators of parameters in CAR model are consistent. Several examples are given to illustrate our results.

When applying probabilistic models to support decision making processes, the users have to strictly distinguish whether the impact of their decision changes the considered situation or not. In the former case it means that they are planing to make an intervention, and its respective impact cannot be estimated from a usual stochastic model but one has to use a causal model. The present paper thoroughly explains the difference between conditioning, which can be computed from both usual stochastic model and a causal model, and computing the effect of intervention, which can only be computed from a causal model. In the paper a new type of causal models, so called compositional causal models are introduced. Its great advantage is that both conditioning and the result of intervention are computed in very similar ways in these models. On an example, the paper illustrates that like in Pearl’s causal networks, also in the described compositional models one can consider models with hidden variables.

Timely Corporate failure prediction is a major issue in today’s economy especially considering the financial crisis that has affected the World Economy in the last decade. Any prediction technique must be reliable (good recognition rate, sensitivity and specificity), robust and able to give predictions with a sufficient time lag to allow for corrective actions. In this paper we have considered the case of Small-Medium Enterprises (SMEs) in Italy during the 2008 crisis, introducing a non-parametric classification algorithm to predict corporate failure based on financial indicators up to 8 years in advance.

Among the investment assets, gold is historically been thought as a powerful inflation hedge to many households in Malaysia. This paper examines and compares the hedging properties of gold against both consumer and energy inflation risks in Malaysia. Using the monthly domestic gold price, we test the long-run and short-run relationships between gold return and consumer inflation as well as energy inflation. We find that gold investment in Malaysia is a good hedge against consumer inflation and energy inflation in the long run but not for the short run. We also could not find any evidence of short-run causality between gold return and both consumer and energy inflations.

This paper presents an integrated methodology to find out key factors that affects people choose for different types of clinic and hospital department. The requirements of the methodology not only consider factors before, during and after treatment, but also identified clinic, dental clinic, aesthetic clinic, dental department in hospital, department of family medicine in hospital, and department of orthopedics. Although there are multiple and contradictory objectives to be considered respectively, grey relational analysis (GRA) can sort out key factors to each clinic/department and be the decision maker.

Thailand is the first rank cassava exporter in the world. The cassava export quantity from Thailand influences cassava trading in international market. Therefore, Thailand’s cassava export forecasting is important for stakeholders who make decision based on the future cassava export. There are two main types of cassava export which are cassava starch and cassava chip. This paper focuses on the cassava starch, which is around 60 % of the total cassava export value, including three following products: native starch, modified starch and sago. The cassava starch export time series from January 2001 to December 2013 are used to predict the cassava starch export in 2014. The objectives of this paper are to develop ARIMA models and the artificial neural network (ANN) models for forecasting cassava starch export from Thailand, and to compare accuracy of the ANN models to the ARIMA models as benchmarking models. MSE, MAE and MAPE are used as accuracy measures. After various scenarios of experiments are conducted, the results show that ANN models overcome the ARIMA models for all three cassava starch exports. Hence, the ANN models have capability to forecast the cassava starch exports with high accuracy which is better than well-known statistical forecasting method such as the ARIMA models. Moreover, our finding would give motivation for further study in developing forecasting models with other types of ANN models and hybrid models for the cassava export.

Many empirical works used risk modeling under the assumption of Gaussian distribution to investigate the market risk. The Gaussian assumption may not be appropriate for risk estimation techniques in some situations. In this study, we used the extreme value theory (EVT) to examine more precisely the tail distribution of market risk and incorporate high dimensional copulas to explore the dependence between stock markets. We gathered data of stock markets from Asean countries (Thailand, Singapore, Malaysia, Indonesia and the Philippines) to simulate the portfolio analysis during and post subprime crisis. The results found that D-vine copula GARCH-EVT model can simulate the efficient frontier of portfolios greater than other models. Furthermore, we also found the positive dependence for the overall markets.

This study simultaneously determines the factors affecting each pairing of health behaviors such as alcohol-consumption and physical activity, tobacco-consumption and physical activity, and alcohol-consumption and tobacco-consumption. The measure of dependence between these pairs was quantified using the copula approach. The Copula-based Ordered Probit Model was used to control any common unobserved factors that might affect the random errors related to each pair of health behaviors. The results is more efficient parameter estimates, in terms of lower standard errors, in comparison with separate estimations. Moreover, understanding the dependencies between ordinal choices for each pair of health behaviors gives useful information for designing more efficient health care programs.

The researchers reinvestigate the effect of alcohol consumption on hypertension from observational data, taken from the Thai National Health Examination Survey. In the observed samples, the treatment assignment is not ignorable, thus using treatment as a dummy variable in the statistical model will lead to the bias estimation of treatment effects. Factors affecting self-selection (drink/not drink) may cause the dummy variable of treatment to be correlated with random errors in the outcome model, which leads to the biased parameters estimation. We propose to use copula-based endogenous switching regression for ordinal outcomes as the more appropriate model for treatment effect estimation. The new results should give us more a accurate and reliable treatment effect for causal inference.

We applied the vine copulas, which can measure the dependence structure of uncertainty in portfolio investments. C-vine and D-vine copulas based on capital asset pricing models were used to exhibit portfolio risk structure in the content of asset allocation. With this approach, we employed the Monte Carlo simulation and the empirical results of C-vine and D-vine copulas to determine the expected shortfall of an optimally weighted portfolio. Furthermore, we used the condition Value-at-Risk (CVaR) model with the assumption of C-vine and D-vine joint distribution to gain the maximum returns in portfolios.

Copulas have become one of the most significant new tools to measure nonlinear dependence structure and tail dependence. Combining time-varying copulas and VAR model with kernel density function, this paper proposes a new method, called the time-varying copula-based VAR model, to analyze the transmission and co-movement of rice export prices between Thailand and Vietnam. The time-varying BB1 and BB7 copulas are proposed to measure asymmetric tail dependences. The main findings of this study reveal that there exists obvious co-movement between rice export prices of Thailand and Vietnam, and the time-varying BB7 copula has a better performance than others. In addition, the price transmission between the two markets is bi-directional, and the Vietnamese price is more suitable as price leader in terms of the results of impulse response functions.

This paper aims to model volatility and correlation dynamics in spot price returns of gold and silver, and examines the corresponding market risk management implications. VaR (value at risk) and ES (expected shortfall) are used to analyze the market risk associated with investments in gold and silver. Many GARCH family models are employed to describe the volatility. This work applied the copula based-GARCH model in the estimation of a portfolio VaR and ES composed of gold and silver spot prices. The empirical results exhibit that the NAGARCH and the TGARCH families performed better than other GARCH family members in describing the volatility of gold and silver returns, respectively. Furthermore, the time-varying T copula has the most appropriate performance in capturing the dependence structure between gold and silver returns. The out-of-sample forecast performance indicates that the time-varying T copula-based GARCH model can measure the VaR and ES with the accurate estimates of gold and silver.

This paper aims to find the optimal Global Healthcare Portfolios at different levels of risks and returns to obtain the efficient frontier. The risks are measured by expected shortfall. The dependency of selected stocks in portfolios cannot be ignored. The high-dimension copula-models are used to capture the dependency parameters of the selected stocks. Five largest market capitalization stocks in the global healthcare sector are selected for this analysis. According to the Akaike Information Criterion (AIC), the empirical results show that t-copula is better fitted between the t- and the Gaussian copulas. Based on the t-copula, the result of this study which is the efficient frontier of the global healthcare portfolios is finally shown in Table 4 for related decision makers.

This study aims to analyze the Morgan Stanley Capital International (MSCI) world return and volatility of the healthcare price index using daily time series data. Since the data of MSCI healthcare returns cannot be described by linear models, the residual CUSUM GARCH(1,1) model is applied in this paper. The CUSUM test is used to estimate the optimal change point. The findings of this paper are (1) the estimated point is at day 1,201 of the entire daily data set of 4,209 observations; (2) if the change point is not taken into consideration, the estimated parameters of GARCH(1,1) become $$\hat{\gamma }_1+\hat{\beta }_1 \approx 1$$γ^1+β^1≈1, i.e., we encounter the “IGARCH effect”, which leads to an infinite variance for a model. The contribution of this paper is the recommendation for the analysis of the change point as the necessary condition, rather than jumping into using the whole data set to estimate all parameters of the model without testing nonlinearity, especially for financial time series data.

This paper introduces the indicator circuit with incremental clustering (ICIC) and shows that the ICIC works better than the indicator circuit with reference points (ICRP) for the evaluation of the telecommunications companies’ performance presented in Suriya Int. J. Intell. Technol. Appl. Stat. vol 8, pp 103–112 (2015) [4]. Moreover, it also extends the ICIC to detect high-yield stocks in the Stock Exchange of Thailand. It classifies 134 stocks by 6 indicators; E/P ratio (the inverse of P/E ratio), BV/P ratio (the inverse of P/BV ratio), return on equity (ROE), growth of the E/P ratio, dividend growth, and ROE growth with the data at the end of 2013. It justifies the performance of the model by the yield of the stock measured at the peak price of each stock during April 1st, 2014 to March 31st, 2015. The buying date is the first trading day on the second quarter of 2014, when most of the 2013 financial statements have already been announced. Surprisingly, the method detects the low-yield stocks instead of the high-yield ones. Therefore, it acts like a warning signal to investors to avoid the low-yields.

The main objective of this study is to evaluate some alternatives to estimate tourism arrivals under the presence of structural changes in the sample size. Several specification of Self-exciting threshold autoregressive (SETAR) model and Smooth transition autoregressive (STAR) model, especially Logistic STAR (LSTAR) are estimated. Once the parameters are estimated, a one period out of sample forecasting is performed to evaluate the forecasting efficiency of the best specifications. The finding from the study is that the STAR model beats SETAR model slightly, and these two groups of models have forecast proficiency at least in the tourism field.

This study found the evidences of the dependence between the volatility of stock price index returns and the volatility of exchange rate returns measured against US Dollar and Japanese Yen, and the independence between the volatility of stock price index returns and the volatility of exchange rate returns measured against Euro, in both Thailand and Singapore, under the operation of QE programs. It also found that all bivariate copula of the volatility of stock price index returns—the volatility of Thai Baht/US Dollar exchange rate returns, and the volatility of stock price index returns—the volatility of Thai Baht/Japanese Yen of Thailand, had a degree of dependence greater than that of Singapore. This can be explained that the QE programs can affect capital flows to Thailand and Singapore, and also may have different effects on the volatility of each exchange rate returns and the volatility of stock price index returns, of the individual country. This information can be useful for policy makers and investors so that they can directly focus on avoiding adverse implications from the operation of QE programs, in terms of the risks incurred from the volatility of exchange rate returns and the volatility of stock price index returns.

This paper introduced the seemingly unrelated regression (SUR) model based on Copula to improve a linear regression system since the conventional SUR model has a strong assumption of normally distributed residuals. The Copula density functions were incorporated into the likelihood to relax the restriction of the marginal distribution. The real dataset of Thai rice was used for an application comparing the conventional SUR model estimated by GLS and the Copula-based SUR model. The result indicated that the Copula-based SUR model performed slightly better than the conventional SUR. In addition, the estimated results showed that Gaussian Copula was the most appropriate function for being the linkage between the marginal distributions. Moreover, the marginal distributions also were tested, and the result showed that a normal distribution and student-t distribution were the best fit for the marginal distributions of demand and supply equations, respectively.

This study aimed to analyze price transmission in the Thai rice market using the MS-BVECM. We focused on the data set related to Thailands rice price, including Thai white rice price, Thai parboiled rice, Thai paddy price, and World rice price collected from M1/2004 to M3/2014. We estimated the model with two regimes; namely high market price regime and low market price regime. The estimated results showed that there existed some short-run relationships between these rice prices in both regimes. Unlike the long-run, there existed only one long-run relationship (one cointegrating equation) in the high market price regime expressed in the Thai white rice equation. Meanwhile, Thai paddy price has the long-run relationship and short-run adjustment dynamics in the low market price regime. In addition, we found that India’s non-basmati rice exports and the paddy price guaranteed at 15,000 THB per ton are two main reasons which caused the switching between these two regimes.

Using annual data for the period 1953–2013 or a shorter period 1977–2013, this paper investigates the relationship among money, output and prices in Thailand. The empirical results, obtained by three techniques, namely the Engle–Granger cointegration approach, Johansens cointegation approach and the autoregressive distributed lag (ARDL) bounds approach, suggest the presence of a cointegral-causal relation among money, output and prices, especially for the shorter sample period. Predictably, short-run causal relations also exist between money growth and inflation and between money growth and output growth. The empirical results obtained by a structural vector autoregression (SVAR) model are confirmatory, showing that the accumulated impulse responses of output and prices to monetary shocks are positive and significant. The overall results are consistent with classical monetary theory that money matters insofar its impacts on output and prices are concerned.

China’s capital market is still an emerging market. The weaknesses of a transition market are speculative behaviors of investors and heavy government intervention. The recent observed Chinese stock market has experienced an extraordinary rise while the real economics is going down. Behavioral economics tells us that public comments can profoundly affect individual behavior and decision-making. In this information era, the public can be the “collective intelligence” in social media. This study investigates what is the role of Chinese government play in Sina weibo, which is the biggest microblog in China. Are the public posts, which also contain government messages, correlated or a causal relationship of economic indicators? Here we figure out whether measurements of collective public comments derived from large-scale Sina weibo posts are correlated to the value of the Shanghai Composite over time. A Granger causality analysis was finally used to detect the causal relationship between government role in social media and recent Chinese stock market price. The results show that the positive government opinion is the Granger cause of the Chinese stock market.

In this study, we measure the ability to produce and provide services of telecommunication firms in Thailand. We propose a methodology to estimate firm technical efficiency using the Stochastic Frontier Analysis (SFA) with copulas to capture both the dependence of the stochastic and inefficiency error components and the dependence of the errors across time. Allowing the dependence between the stochastic error and the aggregate error prevents the estimation of the technical efficiency from being biased. Moreover, allowing the time dependence for the aggregate error improves the efficiency of the model. The results show that the average technical efficiency is 0.54, which indicates a room for improvement in the industry.

Macroeconomic factors affecting the growth rate of foreign direct investment (FDI) in AEC member countries were investigated using panel quantile regression to investigate the effects of foreign direct investment (FDI) of ASEAN Economic Community (AEC) member countries. Yearly data covering the period of 2001 to 2012 for nine countries except Myanmar were used for the estimation. As the data of Myanmar are limited, this study covered only Singapore, Philippines, Brunei, Cambodia, Indonesia, Laos, Malaysia, Vietnam and Thailand. The independent variables include the growth rate of gross domestic product, the growth rate of exchange rate ratio and inflation. The findings of this study show that the growth rate of gross domestic product affects the growth rate of foreign direct investment (FDI) positively and statistically significant at all levels of quantile except the quantile at 0.75.

Thailand’s performance on poverty reduction is obviously impressive because it has already achieved the first goal about halving poverty rate of the Millennium Development Goals promoted by the United Nations Development Programme. The main reason for this success is an outstanding economic growth which is widely accepted to be an efficient tool in eliminating of poverty in many developing countries. This paper attempts to quantitatively estimate the relationship between per capita income and poverty rate in Thailand at national and provincial level using the panel data between 2006 and 2013, and then to suggest the proper policies to accelerate the progress on poverty.

After the economic crisis in 2007, the United States enter to the economic recession. Thus the central banks (Fed) purposed an unconventional policy and launch various programs in order to restore the weak economic. However, it also generated a spillover effects toward Emerging countries through capital flow. Therefore, the paper aims to provide a new empirical finding by examining the effect of quantitative easing (QE) policy of the United States on Thailand, Indonesian, and the Philippine, Singapore, and Malaysian financial markets (ASEAN-5). In this study, the ASEAN-5 financial markets, comprising the exchange rate market, the stock market, and the bond market are considered. To measure the effect of QE on those markets, we employed the Markov-switching VAR model to study the transmission mechanisms of QE shocks between periods of expansion in the QE program and QE tapering. Moreover, we restrict the structure of the model in order to identify the determinant of the structural change. This paper finds that ASEAN-5 financial markets receive the effect form QE. The treasury securities purchase program seems to generate a larger effect to the ASEAN-5 financial market than other programs. Moreover, the test of best MS-VAR specification, provide the result that MSH(2)-VAR(1) is the best specification model for the exchange rate market and the stock market, while MSIH(2)-VAR(1) is the best specification model for bond markets. This indicates that QE was not the factor leading the ASEAN-5 financial markets switch from one regime to another regime.

We analyze the dependence relationship between the Thai currency and international currencies after the introduction of quantitative easing (QE). The daily currency exchange rates of Thailand, European countries, Great Britain, Japan, Indonesia, the Philippines, Singapore, and Malaysia during 2009–2014 are applied in this study. We proposed a Markov-switching dynamic copula approach to test the co-movement between the exchange rates and the Thai Baht. The results show that there is a dependence relationship between the Thai Baht and the other currencies except in the case of the Great British Pound. Additionally, we also found that a high dependence regime has higher volatility than a low dependence regime.

In this paper, we analyze the dependency between the Thailand, Indonesia, and the Philippine (TIP) stock markets and gold markets using dynamic copula with the Markov-switching model with 2 regimes, namely high dependence and low dependence regimes, and extend the obtained correlation to measure the market risk. We are particularly interested in examining whether or not gold serves as a hedge in the TIP stock markets. Using daily data from January 2008 to November 2014, we find that the Gaussian copula identifies a long period of high dependence of TIPGOLD returns (market downturn) which coincides with the European debt crisis. However, if we do not take gold into account, the dependence between the TIP returns is lower in both regimes, thereby leading to a higher value at risk (VaR) and expected shortfall (ES). Therefore, gold can serve as a hedging, or a safe haven, for TIP stock markets during market downturns and upturns. Additionally, the Kupiec unconditional coverage and the Christoffersen conditional coverage test are conducted for VaR and ES backtesting. The results reveal that the Gaussian Markov-switching dynamic copula is the appropriate model to estimate a dynamic VaR and ES.

This paper investigated the influence of various factors-product knowledge and perceived risk, benefit, and value-on consumers’ willingness to purchase telehomecare products (e.g., the RFID smart care watch). Of the 500 returned questionnaires, 388 were valid (77.60 %). Structural equation modeling (SEM) was used as the main research method. The results showed that the perceived risk of telehomecare products by consumers had a significant impact on perceived value, but it had no significant impact on willingness to purchase. In contrast, product knowledge significantly influenced the perceived benefit of the product and consumers’ willingness to buy.

This paper applies the $$\beta $$β-convergence regression model in order to assess convergence of total factor productivity among Vietnamese provinces for manufacturing industries. Specifically, we express this model in the form of a Spatial Durbin Model (SDM), which allows us to take into account the presence of omitted variables that can be spatially correlated and correlated with the initial level of productivity. We calculate the annual total factor productivity (TFP) of 63 Vietnamese provinces and 6 manufacturing industries, using the results of the structural estimation of a value-added production function from firm data over the period from 2000 to 2012. The regression of growth rates of TFP over this period on the initial levels of productivity using SDM shows that there is convergence in most industries, i.e. the gap between lower-productivity and higher-productivity provinces decreases. These results also show the importance of modeling the indirect effect of the initial level of productivity of a province on its TFP growth rate, through its effect on neighboring provinces. The inclusion of these indirect effects is made possible by SDM and increases the speed of convergence for most considered manufacturing industries, except for metal and machinery, and transportation and telecommunication.

The purpose of this study is to identify the possible causal links between rural bank development and poverty in provincial area of Indonesia. This study uses panel data for 27 provinces in Indonesia over the period of 2000–2013. The panel co-integration and Granger causality tests are applied to investigate the relationship between rural bank assets and regional poverty. The results of co-integration tests show that there is no long run relationship between rural bank assets and provincial poverty. Moreover, the Granger causality tests show that there is no evidence of causality relationship between provincial poverty and rural bank assets. This means rural banks still have no significant contribution to reducing provincial poverty as intended by the government.