The conduct of effective economic policy faces a multiplicity of macroeconomic challenges, which requires a wide scope of theoretical and empirical analyses. With a focus on the European Union, this doctoral dissertation consists of two parts which make empirical and methodological contributions to the literature on forecasting real economic activity and on the analysis of business cycles in a boom-bust framework in the light of the EMU enlargement. In the first part, we tackle the problem of publication lags and analyse the role of the information flow in computing short-term forecasts up to one quarter ahead for the euro area GDP and its main components. A huge dataset of monthly indicators is used to estimate simple bridge equations. The individual forecasts are then pooled, using different weighting schemes. To take into consideration the release calendar of each indicator, six forecasts are compiled successively during the quarter. We find that the sequencing of information determines the weight allocated to each block of indicators, especially when the first month of hard data becomes available. This conclusion extends the findings of the recent literature. Moreover, when combining forecasts, two weighting schemes are found to outperform the equal weighting scheme in almost all cases. In the second part, we focus on the potential accession of the new EU Member States in Central and Eastern Europe to the euro area. In contrast to the discussion of Optimum Currency Areas, we follow a non-standard approach for the discussion on abandonment of national currencies the boom-bust theory. We analyse whether evidence for boom-bust cycles is given and draw conclusions whether these countries should join the EMU in the near future. Using a broad range of data sets and empirical methods we document credit market imperfections, comprising asymmetric financing opportunities across sectors, excess foreign currency liabilities and contract enforceability problems both at macro and micro level. Furthermore, we depart from the standard analysis of comovements of business cycles among countries and rather consider long-run and short-run comovements across sectors. While the results differ across countries, we find evidence for credit market imperfections in Central and Eastern Europe and different sectoral reactions to shocks. This gives favour for the assessment of the potential euro accession using this supplementary, non-standard approach.

我國自民國68年成立外匯市場以來，積極的推動經濟國際化與自由化，由於台灣對外經貿依存度相當的高，國際貿易是我國經濟發展的趨動力，而匯率扮演著經貿活動關鍵的角色，因而對匯率走勢的預測與掌握，乃成為管理外匯風險的首要工作。 　　影響匯率的因素相當的複雜，其常受到政府政策的干預，再者，匯率未來的走勢往往容易受到預期心理的影響，眾多的影響因素往往使得對匯率預測的困難程度提高。有鑑於此，本文試圖從貨幣學派結構模型著手--包括價格充分調整模型與實質利率差模型，討論貨幣學派結構模型與匯率資料是否配適良好，然而，若未考慮變數的恆定性與否，而進行迴歸分析，將會造成「假性迴歸」的錯誤。於是本文再引進Johansen共積法，擬找出變數間的長期關係，導入錯誤校正模型，以對匯率進行預測的工作。最後，藉由Hamilton所發展的馬可夫轉轍模型，將不可觀察的隨機變數融入模型中，透過機率控制狀態變數的變動，再對匯率進行統計的推估與預測。 　　基於本文採用的資料與樣本期間內，本文作成下列結論： 　　1.貨幣學派結構模型的實證表現不佳，實證的係數符號與理論設定的相差甚多，而其樣本外預測表現也遠不如隨機遊走模型。 　　2.導入共積關係的錯誤校正模型，其樣本外預測表現仍舊不及隨機遊走模型，然而相較於結構模型，其有明顯的改善。 　　3.馬可夫轉轍模型的樣本外預測表現，與隨機遊走模型接近，而對匯率變動方向的預測其表現良好。 　　4.將所有模型一併考慮，則樣本外預測表現以馬可夫轉轍模型最佳，錯誤校正模型次之，而以結構模型為最差。

本篇論文擴充Ang et al. (2007)之基本架構,分別建構台灣各式月資料與季資料的物價指數預測模型,並進行預測以及實證分析。我們用以衡量通貨膨脹率的指標為 CPI 年增率與核心CPI 年增率。我們比較貨幣模型、成本加成模型、6 種不同設定的菲力浦曲線模型、3 種期限結構模型、隨機漫步模型、 AO 模型、ARIMA 模型、VAR 模型、主計處(DGBAS)、中經院(CIER) 及台經院(TIER) 之預測。藉由此研究,我們可以完整評估出文獻上常用之各式月資料及季資料預測模型的優劣。 我們實證結果顯示,在月資料預測模型樣本外預測績效表現方面, ARIMA 模 型對 2 種通貨膨脹率指標的樣本外預測能力表現最好。至於季資料預測模型樣本外預測績效表現, ARIMA 模型對未來核心 CPI 年增率的樣本外預測能力表現最好; 然而,對於 CPI 年增率為預測目標的預測模型則不存在最佳的模型。此外,實證分析中我們也發現本研究所建構的模型預測表現仍遜於主計處的預測,但部份模型的樣本外預測能力表現則比中經院與台經院的預測為佳。 / This paper compares the forecasting performance of inflation in Taiwan. We conduct various inflation forecasting methods (models) for two inflation measures(CPI growth rate and core-CPI growth rate) by using monthly and quarterly data. Besides the models of Ang et al. (2007), we also consider some macroeconomic models for comparison. We compare some Monetary models, Mark-up models, six variants of Phillips curve models, three variants of term structure models, a Random walk model, an AO model, an ARIMA model, and a VAR model. We also compare the forecast ability of these model with three different survey forecasts (the DGBAS, CIER, and TIER surveys). We summarized our findings as follows. The best monthly forecasting model for both inflation measures is ARIMA model. For quarterly core-CPI inflation, ARIMA model is also the best model; however, when comparing the quarterly forecasts for CPI inflation, there does not exist the best one. Besides, we also found that the DGBAS survey outperforms all of our forecasting methods/models, but some of our forecasting models are better than the CIER and TIER surveys in terms of MAE.

In this dissertation, generalized additive neural networks (GANNs) and multilayer perceptrons (MLPs) are studied and compared as prediction techniques. MLPs are the most widely used type of artificial neural network (ANN), but are considered black boxes with regard to interpretability. There is currently no simple a priori method to determine the number of hidden neurons in each of the hidden layers of ANNs. Guidelines exist that are either heuristic or based on simulations that are derived from limited experiments. A modified version of the neural network construction with cross–validation samples (N2C2S) algorithm is therefore implemented and utilized to construct good MLP models. This algorithm enables the comparison with GANN models. GANNs are a relatively new type of ANN, based on the generalized additive model. The architecture of a GANN is less complex compared to MLPs and results can be interpreted with a graphical method, called the partial residual plot. A GANN consists of an input layer where each of the input nodes has its own MLP with one hidden layer. Originally, GANNs were constructed by interpreting partial residual plots. This method is time consuming and subjective, which may lead to the creation of suboptimal models. Consequently, an automated construction algorithm for GANNs was created and implemented in the SAS R statistical language. This system was called AutoGANN and is used to create good GANN models. A number of experiments are conducted on five publicly available data sets to gain insight into the similarities and differences between GANN and MLP models. The data sets include regression and classification tasks. In–sample model selection with the SBC model selection criterion and out–of–sample model selection with the average validation error as model selection criterion are performed. The models created are compared in terms of predictive accuracy, model complexity, comprehensibility, ease of construction and utility. The results show that the choice of model is highly dependent on the problem, as no single model always outperforms the other in terms of predictive accuracy. GANNs may be suggested for problems where interpretability of the results is important. The time taken to construct good MLP models by the modified N2C2S algorithm may be shorter than the time to build good GANN models by the automated construction algorithm / Thesis (M.Sc. (Computer Science))--North-West University, Potchefstroom Campus, 2011.

In this dissertation, generalized additive neural networks (GANNs) and multilayer perceptrons (MLPs) are studied and compared as prediction techniques. MLPs are the most widely used type of artificial neural network (ANN), but are considered black boxes with regard to interpretability. There is currently no simple a priori method to determine the number of hidden neurons in each of the hidden layers of ANNs. Guidelines exist that are either heuristic or based on simulations that are derived from limited experiments. A modified version of the neural network construction with cross–validation samples (N2C2S) algorithm is therefore implemented and utilized to construct good MLP models. This algorithm enables the comparison with GANN models. GANNs are a relatively new type of ANN, based on the generalized additive model. The architecture of a GANN is less complex compared to MLPs and results can be interpreted with a graphical method, called the partial residual plot. A GANN consists of an input layer where each of the input nodes has its own MLP with one hidden layer. Originally, GANNs were constructed by interpreting partial residual plots. This method is time consuming and subjective, which may lead to the creation of suboptimal models. Consequently, an automated construction algorithm for GANNs was created and implemented in the SAS R statistical language. This system was called AutoGANN and is used to create good GANN models. A number of experiments are conducted on five publicly available data sets to gain insight into the similarities and differences between GANN and MLP models. The data sets include regression and classification tasks. In–sample model selection with the SBC model selection criterion and out–of–sample model selection with the average validation error as model selection criterion are performed. The models created are compared in terms of predictive accuracy, model complexity, comprehensibility, ease of construction and utility. The results show that the choice of model is highly dependent on the problem, as no single model always outperforms the other in terms of predictive accuracy. GANNs may be suggested for problems where interpretability of the results is important. The time taken to construct good MLP models by the modified N2C2S algorithm may be shorter than the time to build good GANN models by the automated construction algorithm / Thesis (M.Sc. (Computer Science))--North-West University, Potchefstroom Campus, 2011.