C13 - Estimation: GeneralReturn

Modeling policy effects in the context of high-dimensional data requires a balanced consideration of omitted interaction bias and overfitting problems. This paper investigates the role of machine learning algorithms in stabilizing estimates and demonstrates the possible regularization bias caused by common LASSO methods. To overcome the three problems simultaneously, post-double selection is used to screen for the interaction terms that need to be included in the model, and the variance estimates are expanded to measure the uncertainty of the interaction effects and marginal effects. Monte Carlo simulations analyze the main factors affecting conditional and non-linear relationships: covariance and sample size. The results of empirical examples show that different model settings and estimation methods can lead to observable differences in the conclusion of treatment effect heterogeneity, and in general, post-double selection has better performance than other estimation methods.

This article deals with an analysis of the relationship between the Brent oil price and the US dollar price. This paper analyses the development of the intensity and direction of dependence between the nominal effective exchange rate of the US dollar and the price of Brent oil and other commodities, such as industrial metals, agricultural commodities, gold (including jewellery and platinum) in the period from January 1994 to April 2018. The next section tests the hypothesis that there is a short-term relationship between the effective US dollar exchange rate and the oil price. The last part of the article deals with the question whether there is a long-term relationship between these variables: Brent oil price, effective exchange rate of the US dollar, industrial production index in OECD countries, three-month treasury bill, US ending stocks of crude oil, US percent utilization of refinery operable capacity and the price of ethanol.

Estimation of claims reserves, which should be held by the insurer so as to be able to meet expected future claims arising from policies currently in force and policies written in the past, presents an important task for insurance companies to predict their liabilities. A common approach to the reser-ving problem is based on generalized linear models (GLM). In this article, the application of genera-lized linear mixed models (GLMM) - an extension of the GLM - for estimation of the loss reserves is shown. Since the GLMM allows incorporating a random effect instead of several fixed effects corresponding to the accident years as in case of the GLM, volatility of the prediction is reduced. This allows more flexible risk valuation, which is a crucial element of risk management and capital allocation practices of non-life insurers. A real data example together with diagnostics for the model selection are provided as an illustration of the potential benefits of the presented approach.

This paper investigates whether UIRP principle holds and what is predominant driving force, which influences exchange rate movement - economic fundamentals or short-term speculative behaviour. Analysis covers seven East European transition countries and empirical data comprise weekly time series ranging from first week in January 2003 to last week in December 2013. The research method is Component-GARCH in Mean Model, which decomposes temporary and permanent element of volatility. The mean and variance equations have been adjusted for the structural breaks' presence in order to improve estimated parameters. The results suggested that UIRP principle does not hold in any country. After structural breaks inclusion, we have found that the permanent effect is significant in determination of exchange rate dynamics in five countries, but it does not apply for the transition effect. However, further outliers' purification revealed that only in Serbia short-term transition component plays an important role.

The paper examines the hysteresis hypothesis in unemployment in the case of eight selected countries in transition, using the Kalman filter and testing whether the NAIRU time series are stationary. The empirical results show that the hysteresis effect is confirmed for the majority of the countries. Testing the influence of the inflation growth rate on the decline in the NAIRU and vice versa, performed using the panel regression with fixed effect, confirmed that the increase in inflation leads to decline in the NAIRU. The conclusion also suggests the existence of the impact of actual unemployment rate on the NAIRU, which may be affected by the change in aggregate demand.

Using time series data, the paper empirically analyzes the effects of trade liberalization on exports, imports and trade balance in Pakistan for the period 1981/82 to 2007/08. It concludes that trade liberalization stimulates both exports and imports with the effect being greater on latter than on former thereby worsening the trade balance. Other variables, that is, real exchange rate, domestic and foreign incomes, terms of trade and foreign exchange market distortions affect exports, imports and trade balance in the theoretically expected directions. Inclusion of interaction terms indicate that liberalization stimulates both price and income elasticities of exports, imports and trade balance. These results have important policy implications for degree of trade liberalization in Pakistan.

This work studies implicitly weighted robust statistical methods suitable for econometric problems. We study robust estimation mainly for the context of heteroscedasticity or high dimension, which are up-to-date topics of current econometrics. We describe a modification of linear regression resistant to heteroscedasticity and study its computational aspects. For a robust version of the instrumental variables estimator we propose an asymptotic test of heteroscedasticity. Further we describe robust statistical methods for dimension reduction and classification analysis. We propose the robust quadratic classification analysis based on a new minimum weighted covariance determinant (MWCD) estimator. In general the robust methods based on down-weighting less reliable observations are resistant to outlying values (outliers) and insensitive to the assumption of Gaussian normal distribution of the data. The methods are illustrated on econometric data examples.

This article presents a meta-regression analysis of 32 studies on frontier efficiency measurement in banking, focusing on the sensitivity of the reported estimates to the methodological design. Our findings suggest that study design is crucial for the resulting scores. The differences between the scores estimated using parametric and non-parametric approaches arise when the Fourierflexible functional form is used since this functional form yields lower scores. Generally, the higher the number of observations, the higher is the average estimated efficiency. The removal of scale effects using equity capital increases the profit efficiency but it is insignificant for other scores. The efficiency analysis should distinguish the commercial banking from other bank types because the former tends to deliver lower efficiency scores.

In the paper the discrete choice models are discussed and applied based on empirical data. The main goal of the paper is to find out whether the various discrete choice models provide the analyst with the robust and reliable estimates of values of natural goods or values of changed quality of such goods. Our results indicate that the yielded value estimates using the choice experiment method and discrete choice models are applicable in expert support of decision-making on allocation of public resources to such goods, because they are stable and robust. These issues are investigated for the marginal willingness of Czech visitors to the Mácha Lake beaches to pay for the water quality and beach characteristics.

We investigate valuation of derivatives with payoff deined as a nonlinear though close to linear function of tradable underlying assets. Interest rate derivatives involving Libor or swap rates in arrears, i.e. rates paid at wrong time, are a typical example. It is generally tempting to replace the future unknown interest rates with the forward rates. We show rigorously that indeed this is not possible in the case of Libor or swap rates in arrears. We introduce formally the notion of linear plain vanilla derivatives as those that can be replicated by a inite set of elementary operations and show that derivatives involving the rates in arrears are not (linear) plain vanilla. We also study the issue of valuation of such derivatives. Beside the popular convexity adjustment formula, we develop an improved two or more variable adjustment formula applicable in particular on swap rates in arrears. Finally, we get a precise fully analytical formula based on the usual assumption of log-normality of the relevant tradable underlying assets applicable to a wide class of convexity related derivatives. We illustrate the techniques and different results on a case study of a real life controversial exotic swap.

The paper investigates some features of the corruption in the public administration in a country, which is in the process of integrating into the new structures of the European Union. We estimated the parameters of a regression model that analyses factors like political system pressure, administration transparency, and service quality provided by civil servants using data sets for a representative civil servants sample. Using the regression model and ANOVA we came to the conclusion, that the perception of the corruption has significant differences at the level of the four types of public administration institutions analysed in this paper. The main result is that the corruption level in the public administration is negatively influenced by the stability of the remuneration system, by the lack of transparency and by the political system pressure. On the other hand, church, media, and schools have a positive influence on the reduction of the corruption level.

This paper investigates dependence structures on selected world stock markets. Firstly, a non-parametric univariate measure of a persistence concerning capital markets efficiency is derived and computed. Secondly, we focus on computing of a non-parametric multivariate measure of the persistence indicating an ability of the price mechanisms to hold capital market efficiency under interaction of shocks.

This paper empirically examines the determinants of economic growth and convergence in transitive economies of Central and Eastern Europe in the 1990s. While the cross-section regression suggests the absence of a significant convergence across the EU15 and other transitive economies, the Visegrad four (Slovakia, the Czech Republic, Hungary and Poland) dummy being positive and significant indicates that this group of countries has done relatively better than the other group of transitive economies. Moreover, the results indicate that there was an income per capita convergence within Visegrad countries. Switching to a panel data approach, and controlling for macroeconomic stability, financial development, human and physical capital accumulations and other policy variables, the results seem to suggest that there was a conditional convergence across EU15 and transitive economies in the 1990s.

This paper finds that poverty (the savings gap), income instability, and external factors that include debt service payments and capital flight to be the main causes of overseas borrowing by developing countries in the 1980s and 1990s. As far as the empirical strategy is concerned, the application of a panel data approach seems to be highly preferred, as it allows to control time-specific events that are linked to overseas borrowing, particularly given the rapid changes in the global macroeconomic environment in the past years. Moreover, this strategy helps to produce a more robust explanation by allowing to incorporate country-specific factors as developing countries themselves are heterogeneous in terms of their colonial heritages, geopolitical and strategic significance, and creditworthiness, all affecting the level of indebtedness and the potential bargaining power to manage the subsequent debt crisis.

The article presents an empirically tested assessment of the Bank of Slovenia (BS), national central bank, adjustment policy to surges in capital flows during the last decade. Exchange rate appreciation, undeveloped banking sector, immoderate money market oscillation, unstable economic trends (all phenomena that can also be found in other transition countries) are just some of the detrimental effects that can be provoked by surges in capital flows if the national economy is faced with some fundamental sectoral deficiencies. Empirical results indicated that BS quite successfully mitigated listed effects of excessive foreign currency inflows during the last decade. With the suitable combination of direct and indirect adjustment methods, BS succeeded in preventing, still vulnerable Slovenian economy from a major form of financial crisis and stronger nominal tolar appreciation (this was not the case in some other countries like Hungary, Poland, Czech Republic, Croatia) although there was some real appreciation.