自1997年以降，受到亞洲金融風暴的衝擊，亞洲各國匯率巨幅波動，於是如何增進匯率預測的準確度已成為重要的研究課題。而自1973年布列敦森林體制崩潰，各工業國家改採浮動匯率以來，匯率巨幅波動致使國際收支理論不再能解釋匯率如何決定，於是1970年代，學者們紛紛提出各種匯率決定理論，其中以貨幣學派模型與資產組合平衡模型最受到重視。然而，自1978年始，這些結構模型的解釋能力逐漸受到質疑，在1983年Meese and Rogoff甚至提出結構模型的樣本外預測能力不如隨機漫步模型的樣本外預測表現，引起學者們的討論到底何者的樣本外預測表現較佳。而隨著計量方法的演進實證研究已由恆定的計量方法演進至非恆定的計量方法，在非恆定的計量方法方面，MacDonald and Taylor(1993、1994)、吳宜璋(1996)等人的研究皆採誤差修正模型來做預測。 本研究亦採誤差修正模型來做預測，但對其他學者的研究稍作改良：1.加入結構變動虛擬變數2.以向量誤差修正模型而非一條誤差修正的式子來做預測，在此以整個體系的觀點來做預測3.以背氏方法加入相驗情報來改善預測。 結論為在金融風暴發生期間，匯率受非基本面因素影響較大時，貝氏向量自迴歸模型預測表現較佳。而在金融風暴發生之前，匯率受基本面影響較小時，以貝氏向量誤差修正模型為良好的預測模型。 / This study improves other scholars' empirical studies by testing structure changes and by using Vector Error Correction Model to forecast N.T. Dollars. Futhermore,use Bayesian Method to improve predition .The conclusion is Bayesian VAR Model perform better when forecasting period include Asian finanl crisis . And Bayesian VECM Model is better model when forecasting period don't include Asian financial crisis.And the out of sample prediction performance of structure model is better than Random Walk Model.

No description available.

學界間廣泛地認為一般金融資產報酬具有的特性是:線型不可預測性,條件 異質變異數,非條件尖峰態 ... 等特性o 固然金融資產報酬具有線型不可 預測之特性,可是並不能否決其間可能有非線型依存關係的存在o目前大部 份經濟計量分析方法中的模式建構問題均是在假設模式的結構訊息已知的 條件下求解,然若真實體系的結構訊息未知或不明朗時,貿然地假設為某種 特定的模式結構,則可能又難於避免模式設定錯誤的困擾,因而對於真實體 系行為的描述亦將可能是誤導且不合理的,這意味著:除非該特定的模式結 構正是真實體系的表徵, 否則無論該特定模式的結構特性多完美,均難以 建構一令人信服的數理化模式來表徵真實體系之行為o 不幸地,此一問題 在高度非線型的動態隨機體系中尤其嚴重, 甚至是否存在一 ``真實'' 模式來據以表徵體系之行為,亦是相當值得懷疑, 故考慮一種無需特定結 構訊息假設的無母數方法或函數逼近法實屬必要o 類神經網路中的倒傳遞 網路模式即是符合此種特性的方法之一o然而學界間仍無法確定的是金融 資產報酬序列資料所產生的 ARCH 效果本身是否為真實序列資料產生機制 特性之顯現, 還是應歸咎於被忽略掉條件均數方面之非線性所衍生模式設 定錯誤情況下的代用模式, 並不得而知;另一方面, ARCH 模式的顯著成就 及其價值亦不能予以輕易地漠視, 因此, 試圖將 ARCH 模式所能提供的攸 關訊息納入倒傳遞網路模式的考量之中而形成倒傳遞網路-自迴歸條件異 質變異數 (BPN-ARCH) 模式以增進樣本外預測能力的精度便是本論文最 主要的嘗試重點與目的o

地方稅捐稽徵機關績效之良窳攸關地方政府庫收，更直接影響地方經濟成長、資源配置效率與所得分配之公平；本文試圖採用能排除外在因素與隨機干擾之三階段產出導向資料包絡分析法（以下簡稱DEA），針對23個地方稅捐稽徵機關2004年迄2008年資料進行管理效率評估，再以隨機邊界分析法（以下簡稱SFA）分離外生因素及隨機干擾以調整各機關產出至相同基準後，再評估排除外在與隨機干擾因素的管理效率。 未考慮外生因素與隨機干擾的DEA效率評估結果顯示，有高達88.7%及82.6%的地方稅捐稽徵機關分別於技術效率及純技術效率上尚有改善空間。第二階段SFA估計結果顯示，土地移轉現值、機關內大專畢業以上員額之比例及員額平均年齡對管理效率皆有正向影響；而總統大選期間及有高鐵停靠站之縣市除對部分產出無顯著影響外，對管理效率之提升亦具優勢；然而，服務轄區土地面積與實徵淨額之效率呈負相關，卻與違章漏稅裁罰效率呈正相關；地方首長選舉期間除為避免清理欠稅招致民怨而降低技術效率外，對其他管理效率則皆呈正相關；又都市計畫面積占稽徵區域比例與地方稅實徵淨額之管理效率呈正相關，卻與欠稅清理效率呈負相關。調整後之DEA結果顯示各項效率值與調整前比較皆存在顯著差異，顯示排除外生因素與隨機干擾影響以避免效率值被錯估確有其必要性；但仍有高達93.05%及68.7%之地方稅捐稽徵機關分別於技術效率及純技術效率上存有改善空間；又多數稽徵機關處於規模報酬遞增階段，即產能過剩而造成資源浪費。另與財政部稽徵業務考核成績比較分析，在規模效率平均值排名方面，除臺南市外，甲組機關排名普遍優於乙組機關；但純技術平均效率值之排名卻有一半以上之甲組機關表現反而不如部分乙組受評單位；顯示甲、乙組之分類歷經多年仍沿襲舊有分組將使各機關未能於適合之群組中受考而錯估其績效。 基於上述實證研究結果，本文提出下列政策性建議： 一、建議逐期分階段調整人力及預算至最適規模，以善加運用資源降低產能過剩情形。 二、建議各機關應引用環保之共乘概念，加強政府機關間橫向溝通、聯繫與合作。 三、若情況允許，建議可不區分甲乙組針對全體地方稅捐稽徵機關進行考核。若人力、時間或其他情況不允許，建議研擬具體方案隨各機關規模改變而有重新分級之機制。 四、建議甲組機關亦可選擇純技術效率較佳之乙組機關作為觀摩學習之對象。 五、建議財政當局可考慮將外在因素之影響納入評核，以提升考核之信度與效度。 最後，臺灣自2010年底起將有部分縣市改制為直轄市，考核編組方式勢將有所變革，有待後續研究者追蹤探討；而改制後所引起之資源重分配亦可作為未來之研究議題。 / The Performance of Local Tax Bureaus is relevant to the revenue of Public Treasury, and even has direct impacts on local economic growth, efficiency of resource allocation and equity of income distribution. This paper attempts to use three-stage out-oriented DEA which can rule out the external factors and the statistical noises to evaluate the efficiency of 23 Local Tax Bureaus in Taiwan during the period of 2004 to 2008. After measuring slack variables of each Bureau in the first stage, the Stochastic Frontier Analysis (SFA) model is used to separate the external factors from the statistical noises, and then adjust the output of each Bureau to the same benchmark. Finally, DEA is again used to evaluate the efficiency of 23 Local Tax Bureaus. The DEA efficiency evaluation results in the first stage show that up to 88.7% and 82.6% of the Local Tax Bureaus still have an ample room to improve their technical and pure technical efficiency. In the second stage, the SFA model estimates show that “the present value of land transfer”, “the proportion of post-graduates in the Bureaus” and “the average age of the staff” have positive effects on the efficiency. “The period of Presidential Election” and “the cities or counties that Taiwan High Speed Rail have set station up” have insignificant impact on part of the outputs, but still have the advantages to enhance the efficiency. However, “the expanse of land in service area” has a negative correlation with the efficiency of net taxation, but is positively related to the efficiency with the fine of illegal tax evasion. “The election period of Local County Executive” has a positive correlation with the efficiency except that tax arrears liquidation might reduce technical efficiency. “The ratio of urban-planed area to the tax levy regional” and the efficiency of net taxation are positively correlated, but negatively related to the efficiency of tax arrears liquidation. After excluding external factors and the statistical noises, the third-stage DEA evaluation results are significantly different from those in the first stage, indicating that it is necessary to exclude impacts of external factors and statistical noises in order to avoid the misjudged value of efficiency. There are still as high as 93.05 % and 68.7% of the Local Tax Bureaus have an ample space for improvement respectively on technical efficiency and pure technical efficiency. Most of the Tax Bureaus are at the increasing return of scale stage, implying that the surplus of capacity cause the waste of resources. Finally, compared with the performance evaluation held by Ministry of Finance, the empirical results in this study show that although the classification of Local Tax Bureaus has been adopt for many years, Bureaus’ performance evaluation results might be misjudged in the unsuitable group. Based on the above empirical results, this research attempts to propose the following policy suggestions: 1. It is recommended to adjust the phase of manpower and budget to the optimum scale for the best using of resources and to reduce inefficiency of excess capacity. 2. Enhancing horizontal communication among government agencies may improve the efficiency of Local Tax Bureaus. 3. If possible, it is recommended to evaluate Local Tax Bureaus without classification. If not, a mechanism of re-rating according to the scale change of Local Tax Bureaus may be needed. 4. It is proposed that Local Tax Bureaus of Group A can take lesson from those of Group B with better pure technical efficiency to learn from. 5. It is suggested that government authorities have to exclude the effect of external factors to improve the reliability and validity of performance evaluation. At the end of year 2010, several counties will be restructured in municipalities. The classification of Local Tax Bureaus for performance assessment must be changed. The reallocation of resources caused by restructuring may be used in future studies.

傳統迴歸分析是假設觀測值的不確定性來自於隨機現象，本文則應用模糊隨機變數概念於迴歸模式的架構，考慮將隨機現象和模糊認知並列研究。針對樣本模糊數(x_i, Y_i)，我們進行模糊迴歸參數估計，並稱此為模糊迴歸模式分析。模糊迴歸參數估計大都採用線性規劃，求出適當區間，將觀測模糊數Y_i的分佈範圍全部覆蓋。但是此結果並不能充分反映觀測樣本Y_i的特性。本研究提出一套模糊迴歸參數的估計方法，其結果對觀測樣本的解釋將更為合理，且具有模糊不偏的特性。在分析過程中，我們亦提出一些模糊統計量如模糊期望值、模糊變異數、模糊中位數的定義，以增加對這些參數的模糊理解。最後在本文中也針對台灣景氣指標與經濟成長率作實務分析，說明模糊迴歸模式的適用性。 / Conventional study on the regression analysis is based on the conception that the uncertainty of observed data comes from the random property. However, in this paper we consider both of the random property and the fuzzy perception to construct the regression model by using of fuzzy random variables. For the fuzzy sample (x_i,Y_i), we will process the parameters estimation of the fuzzy regression, and we call this process as fuzzy regression analysis. The parameters estimation for a fuzzy regression model is generally derived by the linear programming scheme. But it's result usually doesn't sufficiently reflect the characteristics of the observed samples. Hence in this paper we propose an alternative technique for parameters estimation in constructing the fuzzy regression model. The result will describe the observed data better than the conventional method did, moreover it will have the fuzzy unbiased properties. For the purpose of fuzzy perception on the fuzzy random variables, we also give definitions for certain important fuzzy statistics such as fuzzy expected value, fuzzy variance and fuzzy median. Finally, we give an example about the Taiwan Business Cycle and the Taiwan Economic Growth Rate for illustration.

本論文針對隨機數位網路提出一套可行的計算機制，以提供網路管理者進行資源分配與壅塞管理的分析工具。我們研究兩種利潤最佳化模型，探討在預算控制下的頻寬分配方式。因為資源有限，網路管理者無法隨時提供足夠頻寬以滿足隨機的網路需求，而量測網路連結成功與否的阻塞機率(Blocking Probability)為評估此風險之一種指標。我們利用頻寬分配、網路需求量和虛擬端對端路徑的數量等變數，推導阻塞機率函數，並證明阻塞機率的單調性(Monotonicity)和凸性(Convexity)等數學性質。在不失一般性之假設下，我們驗證阻塞機率是(1)隨頻寬增加而變小；(2)在特定的頻寬分配區間內呈凸性；(3)隨網路需求量增加而變大；(4)隨虛擬路徑的數量增加而變小。 本研究探討頻寬分配與阻塞機率之關係，藉由推導單調性和凸性等性質，提供此兩種利潤模型解的最適條件與求解演算法。同時，我們引用經濟學的彈性概念，提出三種模型參數對阻塞機率變化量的彈性定義，並分別進行頻寬分配、網路需求量和虛擬路徑數量對邊際利潤函數的敏感度分析。當網路上的虛擬路徑數量非常大時，阻塞機率的計算將變得複雜難解，因此我們利用高負荷極限理論(Heavy-Traffic Limit Theorem)提供阻塞機率的估計式，並分析其漸近行為(Asymptotic Behavior)。本論文的主要貢獻是分析頻寬分配與阻塞機率之間的關係及其數學性質。網路管理者可應用本研究提出的分析工具，在總預算限制下規劃寬頻網路的資源分配，並根據阻塞機率進行網路參數的調控。 / This thesis considers the problem of bandwidth allocation on communication networks with multiple traffic classes, where bandwidth is determined under the budget constraint. Due to the limited budget, there exists a risk that the network service providers can not assert a 100% guaranteed availability for the stochastic traffic demand at all times. We derive the blocking probabilities of connections as a function of bandwidth, traffic demand and the available number of virtual end-to-end paths for all service classes. Under general assumptions, we prove that the blocking probability is directionally (i) decreasing in bandwidth, (ii) convex in bandwidth for specific regions, (iii) increasing in traffic demand, and (iv) decreasing in the number of virtual paths. We also demonstrate the monotone and convex relations among those model parameters and the expected path occupancy. As the number of virtual paths is huge, we derive a heavy-traffic queueing model, and provide a diffusion approximation and its asymptotic analysis for the blocking probability, where the traffic intensity increases to one from below. Taking the blocking probability into account, two revenue management schemes are introduced to allocate bandwidth under budget control. The revenue/profit functions are studied in this thesis through the monotonicity and convexity of the blocking probability and expected path occupancy. Optimality conditions are derived to obtain an optimal bandwidth allocation for two revenue management schemes, and a solution algorithm is developed to allocate limited budget among competing traffic classes. In addition, we present three elasticities of the blocking probability to study the effect of changing model parameters on the average revenue in analysis of economic models. The sensitivity analysis and economic elasticity notions are proposed to investigate the marginal revenue for a given traffic class by changing bandwidth, traffic demand and the number of virtual paths, respectively. The main contribution of the present work is to prove the relationship between the blocking probability and allocated bandwidth under the budget constraint. Those results are also verified with numerical examples interpreting the blocking probability, utilization level, average revenue, etc. The relationship between blocking probability and bandwidth allocation can be applied in the design and provision of broadband communication networks by optimally choosing model parameters under budget control for sharing bandwidth in terms of blocking/congestion costs.

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.

本研究建構一社會網路互動下的新凱因斯動態隨機一般均衡模型，探討效用基礎下波茲曼分配背後的網路結構，以及，社會網路對新凱因斯動態隨機一般均衡模型參數的影響。根據本論文模擬結果，效用基礎下波茲曼分配背後所隱含的社會網路結構呈現局部區域性連結拓璞，此結論與熱力學對波茲曼分配中粒子互動方式的假設相同，然而，區域性連結之網路結構（如環狀網）並非目前實證研究所觀察到的網路型態（如冪分布網路或高群集係數之小世界網路），故吾人是否得以直接利用效用基礎下波茲曼分配來描述社會上人與人之間的互動現象必需更忱慎考量之。另外，社會網路互動也將使新凱因斯動態隨機一般均衡模型之參數估計產生偏誤，依本研究估計結果觀之，只要加入社會互動，總合需求曲線中實質利率之參數估計將為正號，即實質利率對產出缺口的影響為負向影響，也就是文獻上的投資儲蓄迷思(IS puzzle)，若進一步觀察社會網路結構對該實證迷思的影響則可發現當社會網路群聚程度越高時，該估計偏誤將越嚴重。 / We construct a social network-based New Keynesian DSGE (Dynamic Stochastic General Equilibrium) Model to investigate the underlying social network structure derived from the performance-based Boltzmann-Gibbs model, and thus interpret the process that social network structures affect the estimation bias in the New Keynesian DSGE framework. According to our simulation results, the underlying social network structure derived from the performance-based Boltzmann-Gibbs model should be local. This finding is consistent with the study of thermodynamics, which the Boltzmann-Gibbs distribution is based upon, i.e. the local interaction. However, it contradicts not only the purpose of combining the performance-based Boltzmann-Gibbs machine and New Keynesian DSGE model, but also empirical studies of social network structures in the real world. Accordingly, maybe we have to consider further whether the performance-based Boltzmann-Gibbs machine is a suitable tool for calibrating social interaction under the stylized New Keynesian DSGE framework. Furthermore, if we embedded interaction behavior in the stylized New Keynesian model, the so-called “IS Puzzle” can be consequently observed. We also realized that “IS Puzzle” is connected with network structures. The more clustering the network structure is, the more significant “IS Puzzle” would be.

針對金融時序資料變異數不齊一的性質，隨機波動模型除了提供於ARCH族外的另一選擇；且由於其設定隱含波動本身亦為一個隨機波動函數，藉由設定隨時間改變且自我相關的條件變異數，使得隨機波動模型較ARCH族來得有彈性且符合實際。傳統上處理隨機波動模型的參數估計往往需要面對到複雜的多維積分，此問題可藉由貝氏分析裡的馬可夫鏈蒙地卡羅法解決。本文主要的探討標的，即在於利用馬可夫鏈蒙地卡羅法估計美元/新台幣匯率隨機波動模型參數。除原始模型之外，模型的擴充分為三部分：其一為隱含波動的二階自我回歸模型；其二則為藉由基本模型的修改，檢測匯率市場上的槓桿效果；最後，我們嘗試藉由加入scale mixture的方式以驗證金融時序資料中常見的厚尾分配。

二十一世紀初的經濟不景氣橫掃了全球的DRAM製造產業，讓全世界的DRAM製造產商大賠了120億美金，國內廠商受傷尤其嚴重，甚至發生公司債之債務不履行事件，國內廠商岌岌可危。本論文主要係以案例探討之方式，研究國內DRAM製造廠商之經營模式、產業特性與風險管理問題，尤其在面對國際間產業劇烈之競爭下，國內之DRAM製造商的經營條件比國際大廠更為艱困，例如：金融環境、政府支援程度、生產規模、技術自主問題等與國外廠商皆有一段差距，因此在經營上所面對之風險與其他國家製造商相較，其實更為險峻。 / 本文以案例公司發生公司債之債務不履行事件為切入點，深入地了解一家在本國企業中屬於中大型企業之DRAM製造公司，為何會有債務不履行之情況發生？其近因似為案例公司在財務上過度倚賴公司債為籌資工具，且公司債之到期或轉換公司債之履約期間過於密集，以致產生流動性問題，然而其遠因乃在於DRAM產品價格快速的滑落，廠商缺乏適當的風險管理工具及機制以應付DRAM之價格風險。DRAM產品為成本競爭導向之標準產品，成本競爭來自於生產良率、製程微縮與新建更大尺寸廠房，當每家廠商都競逐於經濟規模以降低成本時產業會變得不穩定而暴起暴落，在產品價格處於高點時，所有廠商將產能利用率(稼動率)推到最高，此時因產能稼動率高，因此平均每單位晶片之生產成本較低，所以廠商獲利頗豐，並可輕易自資本市場取得資金擴充產能；等到市場供過於求，產品價格下跌處於低點時，廠商只好減產以降低損失，在其他條件不變下，此時因產能稼動率低，因此平均每單位晶片之生產成本反而較產品價格好時還要高，產品價格下跌所帶來的巨額損失，對廠商的虧損有乘數效果，此時廠商在資本市場或銀行等間接金融市場都不容易籌措到資金，本文以案例公司所面對之風險管理問題，提供幾個避險之建議，其中包括金融業、政府等應該能夠扮演更積極的角色，創造共贏共榮的局面，並避免類似之事件再發生，此為本文最大之貢獻。