Semiparametric Bayesian estimation of mixed count regression models

This paper develops semiparametric Bayesian estimation approach for Poisson regression models with unobserved heterogeneity of unknown density. This approach is computationally efficient and allows automatic adaptation of the approximating density to data during estimation. Simulations show the estimator performs well.     

Simulation-based inference in dynamic panel probit models: An application to health

This paper considers the determinants of a binary indicator for the existence of functional limitations using seven waves (1991–1997) of the British Household Panel Survey (BHPS). The focal point of our analysis is the contributions of state dependence, heterogeneity and serial correlation in explaining the dynamics of health. To investigate these issues we apply static and dynamic panel probit models with flexible error structures. To estimate the models we use maximum simulated likelihood (MSL) with antithetic acceleration and implement a recently proposed test for the existence of asymptotic bias. The dynamic models show strong positive state dependence. Data from the British Household Panel Survey (BHPS) were supplied by the ESRC Data Archive. Neither the original collectors of the data nor the Archive bear any responsibility for the analysis or interpretations presented here. Funding was provided by the ESRC award no: R000238169-Simulation-based econometric approaches to investigating the interaction of lifestyle and health. The authors would like to thank William Greene for valuable comments on an earlier draft of the paper, Roberto Leon Gonzalez for valuable programming advice, and participants at the iHEA Third International conference, York, 22–25 July 2001 and York Seminars in Health Econometrics (YSHE) for their comments.     

An analysis of East Asian currency area: Bayesian dynamic factor model approach

There has recently been an increasing interest in the establishment of a common currency area in East Asia in the aftermath of the East Asian financial crisis. In this article I examine the desirability and feasibility of forming a currency area in the region by checking the symmetry of shocks as an important criterion of the theory of Optimum Currency Area. I employ a dynamic factor model to decompose aggregate output into world, regional and country‐specific components and estimate the model using a Gibbs sampling simulation. Persistent properties of those components are examined and variance decomposition analysis is performed to investigate the role of each component in output variance. The European Monetary Union, with the successful launch of the euro, is the natural benchmark for comparison. Based on variance analysis, it is found that East Asian countries, on average, are less plausible candidates for a currency area than European counterparts. However, a subgroup of countries in East Asia is as qualified as those in Europe. Given the ongoing integration in East Asia, it is not premature to prepare for such a currency area in this region.     

Bayesian model comparison by Markov chain simulation: Illustration using stock market data

The paper illustrates the computation of marginal likelihoods and Bayes factors when Markov Chain Monte Carlo has been used to produce draws from a model’s posterior distribution. The method is based on Raftery (1996) and does not require that Gibbs sampling is used or conditional posterior distributions are available in closed form. Models used include a normal finite mixture, a GARCH and a Student t -model as alternative models for the Standard and Poor’s stock returns.     

Bayesian estimation of cointegrating thresholds in the term structure of interest rates

Threshold Error Correction Models are used to analyse the term structure of interest Rates. The paper develops and uses a generalisation of existing models that encompasses both the Band and Equilibrium threshold models of [Balke and Fomby ((1997) Threshold cointegration. Int Econ Rev 38(3):627–645)] and estimates this model using a Bayesian approach. Evidence is found for threshold effects in pairs of longer rates but not in pairs of short rates. The Band threshold model is supported in preference to the Equilibrium model.

The choice of foreign market entry mode: An analysis of the dynamic probit model

Based on the official survey data of Taiwanese FDI in China, this study investigates the transaction cost determinants of foreign market entry mode choice. In contrast to the existing literature, this paper constructs a panel data set and makes a dynamic probit analysis on the mode choice between wholly-owned subsidiaries and joint ventures. Empirical results indicate that state dependence between current and past modes plays an important role in determining entry mode choice. However, the impact of state dependence will be reduced significantly if the mode adopted in the previous period is associated with poor performance of the parent firm. There is strong linkage between mode performance and future mode choice. Evidence also shows that the hypothesized influences of transaction cost factors on entry mode choice gain support in the dynamic choice process.     

Cost performance of Brazilian soccer clubs: A Bayesian varying efficiency distribution model

This paper analyzes the cost efficiency of Brazilian first league soccer clubs using a Bayesian Varying Efficiency Distribution (VED) model. We confirm that the model fits the data well with all coefficients correctly signed and in line with the theoretical requirements. From the efficiency results, it was clear that the Brazilian soccer league operates at a lower performance in comparison to other international soccer leagues. Factors which contributed to this finding as well as other policy implications are provided.     

Research & development and growth: A Bayesian model averaging analysis

We examine the effect of research and development (R&D) on long-term economic growth using the Bayesian model averaging (BMA) to deal rigorously with model uncertainty. Previous empirical studies, which applied BMA, investigated the effect of dozens of regressors on long-term growth, but they did not examine the effect of R&D due to data unavailability. We extend these studies by proposing to capture the investment in R&D by the number of Nobel prizes in science. Using our indicator, the estimates show that R&D exerts a positive effect on long-term growth. This result is robust to many different parameter and model prior structures as well as to alternative definitions of R&D indicator.     

Empirical tests on the asset pricing model with liquidity risk: An unobserved components approach

The aim of this paper is to test empirically the conditional liquidity-adjusted capital asset pricing model (L-CAPM) developed by Acharya and Pedersen (2005). Accordingly, we propose to estimate the L-CAPM using unobserved components methodology, which allows us to take into account the main stylized facts characterizing liquidity. Based on a sample of firms listed on the NASDAQ, our empirical analysis reveals several findings. Firstly, we show that liquidity is time-varying and exhibits strong seasonality. Secondly, we highlight the impact of the liquidity level premium on asset prices. Thirdly, we show that the most important liquidity risk is related to the covariance between portfolio illiquidity and market returns. Fourthly, we observe a negative relationship between portfolio returns and market illiquidity. Fifthly, we find that liquidity risk and illiquidity level are not always positively correlated.     

中国核心通货膨胀的估计——基于贝叶斯Gibbs Sampler状态空间模型

观测到的通货膨胀可以分解为趋势成分和暂时成分,其中的趋势成分即核心通货膨胀。准确度量核心通货膨胀对宏观经济政策的制定和经济形势的判断有重要意义。本文首先建立了估计核心通货膨胀的状态空间模型,然后将贝叶斯Gibbs Sampler方法应用于估计该状态空间模型的参数,以克服卡尔曼滤波的缺陷。本文估计了1991-2010年的核心CPI,结果表明估计的核心CPI很好地反映了货币政策的变化,同CPI相比,核心CPI有较小的波动性并且与货币供给增长率具有更强的相关性。由此得出结论,当前只要食品价格不出现持续的大幅上涨,同时继续保持稳健的货币政策,中国就不会出现严重的通货膨胀,即使个别食品价格上涨造成CPI上涨也必然是不可持续的。     

The existence of R&D spillovers: A cost function estimation with random coefficients

This is a study of the effects of R&D spillovers on the cost and production structures of Finnish manufacturing firms. Confidential data on firms is used to estimate a translog cost function system with random coefficients. Although the results suggest that intra-industry spillovers are present in Finnish manufacturing, the findings regarding inter-industry spillovers are inconclusive. The variable cost reduction associated with spillovers is positive, but relatively low. Spillovers reduce the demand for labor but increase the demand for materials. Spillovers also reduce the willingness to pay for capital inputs.     

Model uncertainty and asset return predictability: an application of Bayesian model averaging

We investigate model uncertainty associated with predictive regressions employed in asset return forecasting research. We use simple combination and Bayesian model averaging (BMA) techniques to compare the performance of these forecasting approaches in short-vs. long-run horizons of S&P500 monthly excess returns. Simple averaging involves an equally-weighted averaging of the forecasts from alternative combinations of factors used in the predictive regressions, whereas BMA involves computing the predictive probability that each model is the true model and uses these predictive probabilities as weights in combing the forecasts from different models. From a given set of multiple factors, we evaluate all possible pricing models to the extent, which they describe the data as dictated by the posterior model probabilities. We find that, while simple averaging compares quite favorably to forecasts derived from a random walk model with drift (using a 10-year out-of-sample iterative period), BMA outperforms simple averaging in longer compared to shorter forecast horizons. Moreover, we find further evidence of the latter when the predictive Bayesian model includes shorter, rather than longer lags of the predictive factors. An interesting outcome of this study tends to illustrate the power of BMA in suppressing model uncertainty through model as well as parameter shrinkage, especially when applied to longer predictive horizons.     

Tobin <Emphasis Type="Italic">q</Emphasis>: Forecast performance for hierarchical Bayes,shrinkage, heterogeneous and homogeneous panel data estimators

This paper reconsiders the Tobin q investment model studied by Hsiao et al. (1999) using a panel of 337 U.S. firms over the period 1982–1998. It contrasts the out-of-sample forecasts performance of hierarchical Bayes, shrinkage, as well as heterogeneous and homogeneous panel data estimators. We are very grateful to Cheng Hsiao and A. Kamil Tahmiscioglu for providing us with the data set and computer code for the Hierarchical Bayes estimator as well as useful suggestions and comments.     

Adoption of conservation tillage in Australian cropping regions: An application of duration analysis

The diffusion of agricultural innovations is rarely rapid. As a result, adoption decisions over time are very likely to be influenced by changes in generic factors such as the price of complementary inputs and environmental conditions. Despite this, the vast majority of studies aiming to explain innovation adoption are limited to cross-sectional data and analysis techniques that cannot accommodate time-dependent variables. This study departs from these attempts by using a duration analysis technique to investigate the significance of time-dependent economic and environmental variables, along with cross-sectional variables such as technology-specific perceptions, on the adoption of soil-conserving cropping practices by grain growers in southern Australia over the period 1983–2003. Of particular interest is the possible trade-off faced by farmers where adoption of no-tillage cropping technology can lead to greater herbicide reliance and subsequently unsustainable weed management due to high risks of herbicide resistance. Results show that factors affecting the cost-effectiveness of herbicides are important in the adoption of the erosion-reducing cropping systems. Several factors relating to the availability and use of technical information are also shown to be influential. The duration analysis approach allowed changes in time-dependent variables, including the fall in the price of the herbicide glyphosate, to be identified as determinants of the timing of no-till adoption.     

An application of maximum entropy estimation: the demand for meat in the United Kingdom

In many econometric studies of demand relationships the design matrix is frequently subject to severe collinearity. In this paper the Generalized Maximum Entropy methodology is introduced and used to estimate a set of demand relationships. The ability of Generalized Maximum Entropy to estimate economic relationships that are typically subject to a high degree of collinearity among the explanatory variables, thus potentially causing traditional methods of estimation to be unreliable, is explained. The results derived by this alternative method of estimation, for a UK meat demand data set, are analysed and examined. The potential for this emerging estimation methodology is discussed.     

Currency devaluation with flexible wages: An application of the Cebula model

The results of Salop (1974) and Purvis (1979), which have received wide acceptance among economists, show that a currency devaluation will definitely depress output if money wages are freely flexible. This note reconsiders the Salop-Purvis proposition by adopting the Cebula (1976) specification of the consumption and investment functions, and finds the proposition may not be valid under this amended framework.     

Pooling surveys in the estimation of income and price elasticities: An application to Tunisian households

This paper presents a methodological extension of Deaton's (1990) model for estimating price elasticities, by pooling Tunisian data from several surveys to improve the inter-cluster variability of unit values which is one of the key elements used in the derivation of these elasticities. Since the surveys cover a relatively long period, possible structural changes in consumption behaviour occurring over time are accounted for by postulating that certain response coefficients of the basic model vary from one survey to the other. The own price and cross price elasticities calculated using appropriate estimates of the extended model are satisfactory both from the economic point of view of their sign and the statistical point of view of their significance and superior to those obtained using a single survey. First version received: April 2000/Final version received: June 2001 RID="*" ID="*"  The authors would like to thank Angus Deaton and anonymous referees for useful comments and suggestions.     

The democratization process: An empirical appraisal of the role of political protest

This paper analyses the role of peaceful and violent protest in the democratization process. We interpret the democratization process as a sequence of phases so as to allow citizens' and elites' preferences for democracy to vary according to the particular phase that a country is experiencing. By doing so we jointly model the probability of protest and of moving through different phases of democracy taking into account time-constant and time-varying unobserved heterogeneity. In particular, we develop a multivariate finite mixture model that introduces a latent variable to capture unobservable factors. On a sample of 171 countries from 1971 to 2010, we provide evidence that protest matters in all phases of democratization, especially peaceful citizens’ demonstrations. On the contrary, violent protest has ambiguous consequences, as it favours initial democratic transitions but at the cost of threatening democratic consolidation. We also find that, after conditioning for economic and institutional controls, there is evidence of time-varying unobserved heterogeneity.     

The money-age distribution: Empirical facts and the limits of three monetary models

The money-age distribution is hump-shaped for the US post-war economy. There is no clear-cut relation between the variation of money holdings within generations and age. Furthermore, money is found to be only weakly correlated with both income and wealth. We analyze three motives for money demand in an overlapping generations setup in order to explain these observations: (1) money-in-the-utility, (2) an economy with costly credit service, and (3) limited-participation. All three models are consistent with the hump-shaped relation between average money holdings and age, yet they predict a much closer association between money holdings, income, wealth, and age than we find in the data. Only the limited-participation model partly replicates the low bivariate correlation between money and income as well as between money and interest-bearing assets. None of the three models satisfactorily explains these stylized facts.