由於1980年代開始，期間利差(term spread)被發現對於預測未來經濟狀況，存在良好的預測能力，也奠定了期間利差在對於預測未來經濟研究中的地位。因此，本文主要著重於檢視利用台灣利率資料所建構出的期間利差對於預測台灣實質經濟產出，是否也扮演著如此重要的角色。 我們利用台灣過往的利率資料，從2002年第一季開始到2013年第四季，台灣十年期中央政府公債殖利率與31-90天期國庫券次級市場利率所建構之期間利差，除了使用樣本內(in-sample)結果的分析與樣本外(out-of-sample)的預測結果，搭配Haubrich and Dombrosky (1996)的預測方程式與均方根誤差RMSE(Root Mean Square Error)來檢視期間利差的預測實質經濟狀況的能力是否良好。 實證結果發現，樣本內的結果顯示，期間利差的解釋能力大約延續的三個季度；而樣本外的預測結果雖不理想，但期間利差在預測方程式中仍可扮演良好的預測變數之一。 本文發現，雖然在2007第三季發生結構性的轉變，但期間利差對於實質經濟成長率仍有良好的解釋能力，這並不影響預測的結果。而造成預測能力不佳的因素可能有幾點：第一，由Smets and Tsatsaronis(1997)所稱的總和供給面與總和需求面的衝擊導致期間利率的預測能力下降相同；第二，Wright(2006)所稱期間溢酬(term premium)在相對較低情況下，導致預測能力下降；第三，美國的貨幣政策與台灣的貨幣政策執行上有所不同。

By monetary policy decisions, central banks use output gap to keep macroeconomic variables at their natural levels. A substantial disadvantage of this variable is the fact that it is an unobservable variable which is very problematic to measure, although it is possible to estimate it with various methods of estimation. This thesis aims to find the most suitable method of estimation for Czech economy. Thirteen methods have been chosen for this aim: linear trend, quadratic trend, HP filter, band-pass filters, robust trend, univariate unobserved component model, two types of production function, two SVAR models, multivariate HP filter and multivariate unobserved component model. Own estimations have shown that estimated trajectories of unobservable states were not identical. For own selection of the most suitable method of estimation, quantitative (ability to forecast inflation ,a growth of product and data revisions by selected national and international organisations) and qualitative criterions (qualities of methods of estimation, transparency and easy application) have been selected, where emphasis was put on quantitative criterions. Results of this thesis will show that the most suitable method of estimation output gap for Czech economy is multivariate unobserved component model.

Macroeconomic stability is one of the most important national objectives in any country. However, economies are often subjected to a number of shocks (internal and external), which can be destabilising, produce volatility and make it difficult to achieve and maintain economic stability. Consequently, various policies are used to help deal with the various shocks that may affect the economy. Of all the available policies, monetary policy appears to have been ever more at the centre of macroeconomic policymaking. Meanwhile, for monetary policy to be effective, there is a need for a better understanding of the transmission mechanism, i.e., the process through which monetary policy decisions are transmitted into changes in real output and inflation. Whereas extensive research on the transmission mechanism has been conducted in developed countries, such work in developing countries, especially in Africa is lacking. This could be due to the fact that it was not long time ago, around the 1990s that countries in Africa started adopting the more modem central bank operations in a market-based economic and financial system characterised by indirect monetary policy. Such operations require an understanding of the transmission mechanism. Lack of empirical analysis of the monetary transmission mechanism in Botswana and developing countries of Africa in general, is the main motivating factor behind this thesis. The main objective of this thesis is, therefore, to estimate the transmission mechanism of monetary policy in Botswana. Three different, but complementary techniques (the Narrative Approach, Vector Autoregression (VAR) analysis and the Structural Approach involving the estimation of a small structural model for Botswana economy) are used. Results from these methods tell a consistent story and indicate that monetary policy in Botswana affects real output and inflation through the interest rate channel, while the exchange rate channel is not operational. The credit channel is also active but not strong. The structural approach also indicates that devaluation is contractionary in Botswana, but more research is necessary before firmer conclusions could be made.

The Okun’s Law has been frequently addressed in macroeconomic studies around the world, mostly in national levels, with few studies that estimate the relationship at regional level. This work, using 2002q1-2014q3 quarterly data, estimated the Okun coefficients for Brazil and for the metropolitan areas of Belo Horizonte, Porto Alegre and Sao Paulo. The conclusion is that Okun’s Law is valid in Brazil and in one of the three metropolitan regions studied, and there are regional differences in the responsiveness of labor markets to output fluctuations and deviations from their long-term levels, which, according to the literature, implies the need to implement unemployment policies that not only meet the regional disparities in the labor market, but also the elasticity of the labor market to changes in economic activity in each location. / A Lei de Okun tem sido frequentemente abordada em estudos macroeconômicos ao redor do mundo, na sua grande maioria em níveis nacionais, havendo poucos estudos que compreendam a relação em nível regional. Este trabalho, a partir de dados trimestrais de 2002q1 a 2014q3, estimou os coeficientes de Okun para o Brasil e para as regiões metropolitanas de Belo Horizonte, Porto Alegre e de São Paulo. Verificou-se que a Lei de Okun é válida no Brasil e em uma das três regiões metropolitanas estudadas, e que existem diferenças regionais quanto à capacidade de resposta dos mercados de trabalho às flutuações do produto e aos desvios de seus níveis de longo prazo, o que, segundo a literatura, implica a necessidade de implementar políticas de desemprego que observem não apenas as disparidades regionais no mercado de trabalho, mas também a resposta deste a alterações da atividade econômica em cada localidade.