This diploma thesis evaluates out-of-sample predictive ability of exchange rate models. The first part of the thesis summarizes existing empirical findings about exchange rate predictability and describes exchange rate models chosen to be evaluated. The second part of the thesis evaluates predictive ability of purchasing power parity, uncovered interest parity, monetary model and Taylor rule model. The exchange rate models are evaluated on CZK/EUR and CZK/USD currency pairs. The analysis is made using quarterly data from 1999 to 2013, while 2009 to 2013 period is reserved for forecast evaluation. The predictive ability of exchange rate models is evaluated in one quarter, one year and three years horizons. The exchange rate models are specified in first differences and estimated by ordinary least squares method. The forecasts are made using rolling regression. The exchange rate models are evaluated using RMSE, Theil's U, CW test and direction of change criterion. The diploma thesis concludes with description of own empirical findings.

明天股票的價格是會漲還是會跌呢? 明天到底會不會下雨? 下期樂透開獎會是哪些號碼呢? 未來不知道會發生哪些事情? 大家總是希望能夠未卜先知、洞悉未來! 可是我們要如何進行預測呢? 本文比較了不同時間序列模型的預測績效, 而且測試預測的結合是否能夠改進預測的準確度? 時間序列模型的研究在近年來非常蓬勃地發展, 所以本文簡單介紹了時間序列模型(Time series models)當中的線性AR模型、非線性TAR模型、非線性STAR模型, 以及這些模型該如何來進行在樣本外的預測。 同時本文說明了預測的結合(Combined forecast)該如何進行? 預測結合的目的是希望能夠達到截長補短的效果! 除了傳統迴歸(Regression-based)方法和變動係數(Time-varying coefficients)方法外, 本文提出了兩種非迴歸類型的預測結合方法, 績效權數(Fitness weight)和支向機(Support Vector Machine)。 其中主要的焦點放在支向機, 因為迴歸方法可能會有共線性的問題, 支向機則是沒有這個問題。 本文實證的結果顯示, 在時間序列模型方面, 非線性模型的預測能力, 在大多數的情形底下, 都不如簡單的線性AR模型; 在預測結合的方面, 支向機的績效是和迴歸方法的績效是差不多的, 這兩者都比變動係數方法的績效來得穩固, 可是如果基底模型的預測值存在共線性的問題或樣本數目過少的問題, 那麼支向機的績效是優於迴歸方法的績效。 最後, 時間序列模型的預測績效會受到資料性質的影響, 而有極大的改變, 或許我們可以考慮使用比較保險的預測策略-預測結合, 因為預測結合的預測誤差範圍是小於時間序列模型的預測誤差範圍!

Submitted by EDUARDO MORATO MELLO (eduardo.mello@br.natixis.com) on 2015-02-04T19:07:16Z No. of bitstreams: 1 MPFE_EduardoMello.pdf: 1511350 bytes, checksum: 0c43eb471871651f1d5b9ab8996e0e63 (MD5) / Rejected by JOANA MARTORINI (joana.martorini@fgv.br), reason: Eduardo, Alterar o ano para 2015. on 2015-02-05T15:09:42Z (GMT) / Submitted by EDUARDO MORATO MELLO (eduardo.mello@br.natixis.com) on 2015-02-05T15:14:07Z No. of bitstreams: 1 MPFE_EduardoMello.pdf: 1511130 bytes, checksum: ee2bf1cdb611b05a4c962200c29ff28f (MD5) / Approved for entry into archive by JOANA MARTORINI (joana.martorini@fgv.br) on 2015-02-05T15:15:33Z (GMT) No. of bitstreams: 1 MPFE_EduardoMello.pdf: 1511130 bytes, checksum: ee2bf1cdb611b05a4c962200c29ff28f (MD5) / Made available in DSpace on 2015-02-05T15:21:22Z (GMT). No. of bitstreams: 1 MPFE_EduardoMello.pdf: 1511130 bytes, checksum: ee2bf1cdb611b05a4c962200c29ff28f (MD5) Previous issue date: 2015-01-30 / Este estudo investiga o poder preditivo fora da amostra, um mês à frente, de um modelo baseado na regra de Taylor para previsão de taxas de câmbio. Revisamos trabalhos relevantes que concluem que modelos macroeconômicos podem explicar a taxa de câmbio de curto prazo. Também apresentamos estudos que são céticos em relação à capacidade de variáveis macroeconômicas preverem as variações cambiais. Para contribuir com o tema, este trabalho apresenta sua própria evidência através da implementação do modelo que demonstrou o melhor resultado preditivo descrito por Molodtsova e Papell (2009), o 'symmetric Taylor rule model with heterogeneous coefficients, smoothing, and a constant'. Para isso, utilizamos uma amostra de 14 moedas em relação ao dólar norte-americano que permitiu a geração de previsões mensais fora da amostra de janeiro de 2000 até março de 2014. Assim como o critério adotado por Galimberti e Moura (2012), focamos em países que adotaram o regime de câmbio flutuante e metas de inflação, porém escolhemos moedas de países desenvolvidos e em desenvolvimento. Os resultados da nossa pesquisa corroboram o estudo de Rogoff e Stavrakeva (2008), ao constatar que a conclusão da previsibilidade da taxa de câmbio depende do teste estatístico adotado, sendo necessária a adoção de testes robustos e rigorosos para adequada avaliação do modelo. Após constatar não ser possível afirmar que o modelo implementado provém previsões mais precisas do que as de um passeio aleatório, avaliamos se, pelo menos, o modelo é capaz de gerar previsões 'racionais', ou 'consistentes'. Para isso, usamos o arcabouço teórico e instrumental definido e implementado por Cheung e Chinn (1998) e concluímos que as previsões oriundas do modelo de regra de Taylor são 'inconsistentes'. Finalmente, realizamos testes de causalidade de Granger com o intuito de verificar se os valores defasados dos retornos previstos pelo modelo estrutural explicam os valores contemporâneos observados. Apuramos que o modelo fundamental é incapaz de antecipar os retornos realizados. / This study investigates whether a Taylor rule-based model provides short-term, one-month-ahead, out-of-sample exchange-rate predictability. We review important research that concludes that macroeconomic models are able to forecast exchange rates over short horizons. We also present studies that are skeptical about the forecast predictability of exchange rates with fundamental models. In order to provide our own evidence and contribution to the discussion, we implement the model that presents the strongest results in Molodtsova and Papell’s (2009) influential paper, the 'symmetric Taylor rule model with heterogeneous coefficients, smoothing, and a constant.' We use a sample of 14 currencies vis-à-vis the US dollar to make out-of-sample monthly forecasts from January 2000 to March 2014. As with the work of Galimberti and Moura (2012), we focus on free-floating exchange rate and inflation-targeting economies, but we use a sample of both developed and developing countries. In line with Rogoff and Stavrakeva (2008), we find that the conclusion about a model’s out-of-sample exchange-rate forecast capability largely depends on the test statistics used: it is necessary to use stringent and robust test statistics to properly evaluate the model. After concluding that it is not possible to claim that the forecasts of the implemented model are more accurate than those of a random walk, we inquire as to whether the fundamental model is at least capable of providing 'rational,' or 'consistent,' predictions. To test this, we adopt the theoretical and procedural framework laid out by Cheung and Chinn (1998). We find that the implemented Taylor rule model’s forecasts do not meet the 'consistent' criteria. Finally, we implement Granger causality tests to verify whether lagged predicted returns are able to partially explain, or anticipate, the actual returns. Once again, the performance of the structural model disappoints, and we are unable to confirm that the lagged forecasted returns antedate the actual returns.

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.

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.