Yes / This paper re-examines the long-run relationship between radiative forcing (including emissions of carbon dioxide, sulphur oxides, methane, and solar radiation) and temperatures from a structural time series modeling perspective. We assess whether forcing measures are cointegrated with global temperatures using the structural time series approach.

[pt] Esta tese tem dois objetivos principais. O primeiro deles é a investigação dos efeitos da aplicação do filtro Hodrick e Prescott( HP) na deteção de ciclos macroeconômicos na série do PIB brasileiro, no período 1900-1992. Comparamos este resultado com os obtidos por Cribari e com a abordagem dos modelos estruturais de Harvey, concluindo que a aplicação do filtro HP pode gerar ciclos espúrios. O outro objetivo é comparar as estimativas da componente sazonal obtidas pela abordagem estrutural de Harvey com o método X11-ARIMA de dessazonalização. São utilizados na comparação séries artificiais com sazonalidade / [en] The thesis has two main objectives. The first one is to investigate the effects of the application of the Hodrick- Prescott filter (HP) in detecting macro-economic cycles in the brazilian GDP series, from 1900 to 1992, We compare the results from the HP method to those of Cribari and the structural approach of Harvey. We conclude that the HP series may generate spurious cycles. The second objective of this thesis is to compare the estimatives of the seasonal component obtained by fitting the structural model of Harvey with those obtained by X-11 ARIMA method for seasonal adjustment. In comparing the two approaches we use artificial generated series with seasonality.

[pt] Este trabalho apresenta um modelo de volatilidade estocástica com especificação multiplicativa, chamado modelo de escala local. O modelo trabalha com a precisão (recíproca da variância) de uma série temporal. A precisão é tratada como componente não observável, caracterizando o modelo como estrutural, e é suposta evoluir segundo um filtro Gama, com um ruído multiplicativo que segue distribuição Beta. A função de previsão para a variância é uma média móvel com amortecimento exponencial (EWMA) no quadrado das observações passadas, a mesma função de previsão do modelo IGARCH(1,1). O fator de amortecimento é estimado por máxima verossimilhança. A densidade de medida é Gaussiana, condicional à precisão não observável, e a densidade preditiva resulta t de Student, cujos graus de liberdade são monitorados pelo fator de amortecimento estimado. A densidade de medida Gaussiaan, embora induza excesso de curtose nas distribuições incondicional e preditiva, pode ser inadequada para modelar dados com um grande excesso de curtose, como é o caso de séries financeiras. Por isso, é testada uma densidade de medida mais genérica, a densidade de potência exponencial, que possui a normal como caso particular. O modelo é chamado modelo de escala local generalizado. A introdução de variáveis explicativas é efetuada de maneira trivial. Intervalos de confiança para os parâmetros do modelo são obtidos via Bootstrap paramétrico. Os resultados obtidos são semelhantes àqueles fornecidos pelos modelos GARCH (1,1) e AR(1)-SV, sendo que o modelo de escala local, além da maior facilidade de implementação, fornece soluções exatas, o que não ocorre no AR(1)-SV, e é mais parcimonioso do que o GARCH(1,1). / [en] In this thesis, we investigate, and develop further, a stochastic volatility modelo named local scale model. This model deals the precision, which is the inverse of the variance unobserved component, and so fits within the framework of structural time series models, the precision is assumed to be a Gamma variable, which evolves through a multiplicative equation, scaled by a Beta variable. The measurement density is Gaussian, conditional on the unobserved precision, and the resulting forecast is a Student`s t density, with a scale which is approximately an exponencially weighted moving average (EWMA) of the sqares of the past observations. The degrees of freedom of the Student`s t distribution are controlled by the size of the discount parameter of the EWMA scheme. The Gaussiannity of the measurement density is potentially inadequate when the model is applied to heavy tailed finance data. Instead, this assumption can be replaced by an exponential power density, which allows the modeling of the observed excess kurtosis. The extension of the model to account for explanatory variables is straightforward. Confidence intervals for the parameters are obtained by Bootstrap. The model fits like the GARCH(1,1)mand AR(1)- SV, but the local scale model, besides being easier to fit, provides a more parcimonious alternative to the GARCH (1,1) model, and has an exact filter, rather than a best linear one, like in the AR(1)-SV.

[pt] Esta tese propõe um modelo para o tratamento de observações diárias e é aplicado na área do setor elétrico, no problema de previsão de carga horária. O modelo proposto é basicamente um modelo estrutural onde a sazonalidade anual (movimentos periódicos dentro do ano) é modelada utilizando a técnica de Splines. Esta técnica também é utilizada na estimação do efeito não linear de uma variável explicativa. O modelo desenvolvido nesta tese também leva em conta os feriados dada a grande influência dos mesmos no consumo de energia elétrica. A metodologia proposta é aplicada à três concessionárias do Sistema Interligado Brasileiro: LIGHT (Estado do Rio de Janeiro); CEMIG (Estado de Minas Gerais) e COPEL (Estado do Paraná). A estimação é levada a cabo utilizando o software STAMP conjuntamente com módulos desenvolvidos no utilitário MATLAB. / [en] This thesis presents a model that deals with daily obsevations applied to the problem of forecasting daily elecricity demand. This approach is basaed on a structural time series model with the annual seasonal pattern being modelled by a Periodic Sppline. The methods of Splines was first used in Harvey and Koopman (1993) to analyse hourly load observations, including temperature used an explanatory variable which is also modelled by a Spline. The main contribuition of this thesis is the treatment of holidays and the temperature response modelled by a spline which considerss the possible vsariations that the effect of temperature has on electricity demand within the year. The methodology is applied to three companies of the Brazilian electrical system: LIGHT (State of Rio de Janeiro), CEMIG (State of Minas Gerais) and COPEL (state of Paraná).

[pt] O presente trabalho trata da melhoria da estatí­stica-teste Multiplicador de Lagrange com distribuição qui-quadrado até ordem n (-1) , baseando-se na expansão de Harris (1985) e na melhoria obtida para os testes Escore, fornecida por Cordeiro e Ferrari (1991 e 1994), Apresentamos uma abordagem totalmente ambientada aos modelos estruturais de séries temporais, utilizando-se tais testes na detecção de ciclos. O trabalho apresenta também uma série de simulações comparando as performances destes testes aperfeiçoados com os tradicionalmente utilizados. / [en] The presente work discusses the improvement of the statistics-test Lagrange Multipliers with chi-squared distribution to order n (-1) , basing itself in Harris´ (1995) expansion and in the improvement for the score tests, furnished by Cordeiro and Ferrari (1991 and 1994). We present a totally adapted aproach to time series structural models, utilizing these tests in the cycles detection. The work aldo presents a serie simulations comparing the perfomances of these improved tests with the ones traditionally utilized.

[pt] Na presente dissertação, além de se expor a fundamentação teórica das Metodologias Estruturais clássica e bayesiana para previsão de séries temporais, analisou-se o comportamento de séries temporais, analizou-se o comportamento da série produção de sacos de polipropileno produzidos pela Indústria Têxtil Oeste Ltda. com observações cobrindo o período de janeiro de 1987 a junho de 1992. Na análise, através dos pacotes computacionais correspondentes: STAMP (clássico) e BATS (bayesiano), utilizou-se variáveis de intervenção e a variável exógena correspondente à produção de metros quadrados de polipropileno, cobrindo período idêntico, ou seja janeiro de 1987 a junho de 1992. Adotando como critério de decisão o erro médio quadráticas previsões no período de ajuste e da análise ex-ante feita com as seis últimas observações (janeiro a junho de 1992), para testar a capacidade extrapolativa dos modelos, escolheu-se um modelo representativo dentro de cada abordagem e em seguida foi feito um estudo comparativo de ambas. / [en] In the present dissertation, besides exposing the theoretical foundations of Structural Models (Classic and Bayesian approaches); we also analysed the series of production of propileno´s bag produced by Indústria Têxtil Oeste Ltda. with observation covering the period from January 1987 to June 1992. We used in the analysis the packages: STAMP (classical) and BATS (Bayesian), with intervention variables and the series of production of square meters of propileno as explanatory variable. As decision criterion we used the mean square error during the period of adjustment and the ex-ante analysis with the last six observation (January up to June of 1992), to test the predictive ability of the models.

[pt] O Modelo Estrutural, recentemente desenvolvido por Harvey, considera a tradicional idéia de modelar uma série a partir de suas componentes básicas não observadas. Em particular, a componente cíclica, que descreve um movimento senoidal amortecido, pode ser utilizada para explicar um comportamento repetitivo ao longo da série. O uso desta componente é motivado pelo fato de que o seu espectro teórico apresenta um pico de valor finito. O modelo original de Harvey define o ciclo da série a partir de uma única senóide amortecida. Porém, a estimação do espectro de algumas séries reais revelou ser razoável supor que a componente cíclica de tais séries representa uma soma de várias senóides amortecidas em diferentes seqüências. Neste trabalho é proposta uma extensão da componente cíclica para o modelo de Harvey e discute-se, para o modelo estendido, a estrutura ARIMA equivalente. Constrói-se um teste de multiplicadores de Lagrange no domínio da freqüência, com o objetivo de verificar a existência de uma freqüência adicional na estrutura do ciclo. Finalmente, a teoria apresentada na dissertação é aplicada de forma a testar a presença de uma segunda freqüência no ciclo da série de índices pluviométricos de Fortaleza. / [en] The Structural Models, recently suggested by Harvey, uses the classic idea of modelling a time series y its non observed components. The cyclical component of the series, which is defined by a smoothed sine, is a special one that may be used for explaining a repetitive behaviour along the series. The motivation for using this component lies in the finite valued peak presnted by its theoretic spectrum. Harvey´s model stays that the cycle of the series is defined by a single smoothed sine. However, the estimated spectrum of certain series showed that it is reasonable to suppose the cyclical component of those series as being a sum of distinct smoothed sines of different frequencies. This thesis proposses an extension of the cyclical component in Harvey´s model and discusses the equivalent ARIMA structure for this extended component. We develop a Lagrange Multipliers test in the frequence domain for verifying the existence of an extra frequence in the cyclical component. The theory presented in the dissertation is applied with the purpose of testing the presence of a second frequence in the cycle of the series of sunspot numbers and in the series of rainfall in Fotaleza.

One fundamental point of departure for this thesis is the importance of addressing all three basic economic research questions: what is produced, with what and for whom and including them in the discussion regarding long-term macroeconomic performance. This could also be stated as that a consistent historical national accounts approach where both aspects of production and distribution are included can significantly enhance the research on macroeconomic historical issues. Built upon this foundation, the objective of this thesis is twofold. To begin with, the objective includes the broadening of the empirical database of the Swedish historical national accounts (SHNA) with accounts for the process involving the horizontal distribution of income. The second objective of this thesis consists of conducting analyses of the Swedish macroeconomic devel­opment using the extended database of the SHNA. An important aspect of the analytical objective involves the exploration of methods that had not widely been applied in Swedish economic historical research. Thus, great emphasis is placed on the methodology used in the analyses of macroeconomic development. These two main objectives forni the disposition of the thesis. The first empirical part consists of work with income accounts in the SHNA. This work has resulted in the establishment of a set of income accounts concur­ring with the procedure recommended in the contemporary national accounting system. In the second part of the thesis, selected macroeconomic issues are examined using the extended SHNA database. The first analysis consists of a closer examination of the presence of periodization patterns in Swedish growth and structural change. In this chapter an analysis based on structural time series models is applied to the SHNA series. The main results of this chapter is that the time series on growth and structural change reveal a pattern that not unconditionally is consistent with the prevailing periodisation pattern recognised in Swedish economic-historical research. Instead, the development pattern reveals features found in international research. The next analysis is concerned with the role of specific institutions for contributing to the slow-down in growth that occurred from the late 1960s and throughout the 1970s and 1980s. In this chapter the importance of the corporate tax system, investment funds and the public pension funds for the efficiency of the resource alloca­tion process is examined. The hypothesis that is examined is that these institutional arrangements altered the distribution of income in such a way that the investment allocation was disturbed and thereby leading to ineffi­ciencies that affected long-term growth negatively. This hypothesis is supported by empirical evidence on changes in the income distribution and changes in long-term rates of growth and structural change. Thus, the investigated institutional arrangements to a certain extent had a negative effect on the Swedish economic per­formance during the 1960s to the 1980s. In the final analytical chapter, the objective is mainly methodological. Here, the focus is on the potential application of CGE-models as a tool for examining Swedish macroeconomic history. A fairly straightforward CGE-model is formulated for the period 1910 to 1930 and estimated using the broadened SHNA. The predic­tions of the model are evaluated against the actual historical development in order to assess the performance of the model. As the model formulated in this chapter generates accurate prediction of the main macroeconomic indicators, it is subsequently used in a counterfactual analysis of the impact of total factor productivity growth on the overall growth performance. In summary, the thesis demonstrates that much can be achieved in the research on the Swedish macroeco­nomic development by utilizing new theoretical approaches and applying state of the art analysis methods as a complement to the structural analytical research that has been conducted previously. However, much research is still required, especially on the improvement of the macroeconomic database where one priority is to create detailed and consistent input-output tables and social accounting matrices. / digitalisering@umu

Deciding the optimal media advertisement spending is a complex issue that many companies today are facing. With the rise of new ways to market products, the choices can appear infinite. One methodical way to do this is to use Marketing Mix Modelling (MMM), in which statistical modelling is used to attribute sales to media spendings. However, many problems arise during the modelling. Modelling and mitigation of uncertainty, time-dependencies of sales, incorporation of expert information and interpretation of models are all issues that need to be addressed. This thesis aims to investigate the effectiveness of eight different statistical and machine learning methods in terms of prediction accuracy and certainty, each one addressing one of the previously mentioned issues. It is concluded that while Shapley Value Regression has the highest certainty in terms of coefficient estimation, it sacrifices some prediction accuracy. The overall highest performing model is the Bayesian hierarchical model, achieving both high prediction accuracy and high certainty.

本篇論文在結構性時間序列模型基礎下，將中國大陸29省市自治區1978-2005年實質人均GDP，拆解出其長期趨勢變動軌跡中的水準值與斜率值，對照傳統上直接利用實質所得數據，以動態縱橫資料方法進行經濟成長條件收斂假說的檢定。本文特色在於加入潛在GDP長期趨勢項的水準值和斜率值，並利用內生解釋變數落後項動態分析。除可驗證隨著時間經過，中國相對貧窮省區是否終將逐漸趕上相對富有省份所得水準外，其次，根據GDP趨勢項一階與二階條件的收斂與否，可進而確認實質GDP收斂的本質。 我們發現，實質人均GDP收斂的本質關鍵在於潛在趨勢水準收斂，潛在GDP趨勢斜率的成長率將左右區域間實質所得收斂速度。大部分樣本中，擴大的Solow模型或考慮不同經濟開放程度因素下的內生成長模型，支持條件收斂假說，而後者設算出的收斂係數明顯較為低。此外，考慮採用Arellano and Bond(1991)的the first difference GMM估計式可能存在弱工具性問題(a weak instruments problem)，以Blundell and Bond(1998)發展出的the system GMM估計式，作為探討初始所得與經濟成長收斂的關係應是較為適合的方法。 / This research examines the economic growth conditional convergence hypothesis. Using the data of 29 provinces in Mainland China between 1978 and 2005, this study applied the structural time series model to deconstruct the provinces’ real GDP per capita into two parts - the level and the slope of trend movement. The characteristics of this paper are to include the level and the slope of trend of potential GDP and to consider the lagged dependent variables into the panel data. This study intends to validate whether the income level of relatively poor provinces will gradually catch up that of the relatively affluent provinces in Mainland China eventually. In addition, this study, based on the convergence or divergence in the first-order and second-order conditions of GDP tendency, will confirm the essence of the convergence in real GDP. The findings are that the essential key of the convergence in real GDP per capita is the convergence of the potential level of GDP. The growth of potential GDP tendency slope would affect the converging speed of real income in regions. The testing results of either the augmented Solow model or the endogenous growth model which considered different economic opening degrees both support the conditional convergence hypothesis in most sample sets, while the estimated convergence coefficients of the later are significantly lower than those of the former. In addition, considering the possible weak instruments problem in the first difference GMM estimator developed by Arellano and Bond (1991), the system GMM developed by Blundell and Bond (1998) should be a more suitable way to observe the relation between initial income level and economic growth convergence.