En 1965, Allais proposa une reformulation tout à fait originale de la théorie quantitative de la monnaie. Il s’agit de la théorie Héréditaire et Relativiste (HR) de la demande de monnaie. Apparue une décennie après la reformulation friedmanienne et la publication du modèle de Cagan (1956) relatif aux hyperinflations, cette théorie n’a pas réussi à se frayer une voie dans le champ de l’analyse monétaire. Plusieurs raisons ont concouru au non succès de cette théorie dont notamment son cadre conceptuel tout à fait étrange par rapport aux approches alors dominantes. L’objet de notre thèse est d’interroger l’apport de la théorie HR en tant que reformulation de la théorie quantitative et ce par rapport à la question de la stabilité de la demande de monnaie.Cette thèse est composée de trois parties. La première partie développe certains préludes nécessaires à l'analyse de la théorie HR (partie I). Les deux dernières parties contiennent les deux principaux résultats de notre travail. Le premier est que la théorie HR constitue une reformulation « ontologique » de la théorie quantitative, qui passe par la considération du temps psychologique-le temps tel que ressenti par l’ensemble des agents économiques (Partie II). Le second résultat est que la théorie HR, en tant que théorie macroéconomique, est grosse d’un changement paradigmatique qui fait écho à celui introduit en physique par la théorie de la relativité : pour certains phénomènes monétaires, la théorie HR substitue à l’explication par des relations causales entre agrégats une explication par la seule déformation psychologique du temps (Partie III). / In 1965, Allais proposed an original restatement of the quantity theory of money. It is the Hereditary and Relativistic (HR) theory of the money demand. Published a decade after Friedman’s restatement and Cagan’s model of hyperinflations, the HR theory remained unknown. Many reasons contributed to the lack of success of this theory, one of which is related to its conceptual framework which is incongruous with the standard approach. The HR theory is based upon the notion of time relativity from a psychological point of view, and the idea that the behavior of economic agents is conditioned by a hereditary effect of past events.Our thesis aims to investigate the contribution of the HR theory as a restatement of the quantity theory with regard to the question of the stability of money demand.The thesis is composed of three parts. The first part contains necessary preludes to the analysis of the HR theory (Part I). The second and the third part contain the two main results of the thesis. The first one is that the HR theory represents an ontological restatement of the quantity theory based on the notion of “psychological time”- time as experienced by the collectivity as a whole (Part II). The second result is that the HR theory, as a macroeconomic theory, contains a paradigmatic shift which echoes the one introduced in physics by the theory of relativity: in the HR theory, an explanation of some monetary phenomena using the psychological distortion of time is substituted to the explanation through causal relations between aggregates (Part III).

傳統貨幣需求函數的估計，使用的影響因素包括物價、所得及利率。但是近年股市的蓬勃發展，對貨幣需求造成了一定程度的影響。 Friedman 就股市對貨幣需求的影響提出 4 大效果，分別是交易效果、資產組合調整效果、財富效果及替代效果。其中替代效果為負，其他的效果為正。然而並非只有股市會對貨幣需求造成影響，貨幣需求同樣會影響股市。本文採用 Johansen Procedure 估計法，首先建立一般的貨幣需求模型，使用的雙變數包括貨幣需求、物價、所得及利率，實證結果確定這些變數存在 2 條共積關係，一是貨幣需求共積方程式，一是物價共積方程式。然後我們將股市成交量放入，同樣確定這些變數間具有 2 條共積關係。 Johansen Procedure 有 5 種模型，分別適用於不同的情況，我們要事先由資料來判斷使用哪一個模型並不容易，因此本文採用了多項標準，包括共積係數符號及其大小、向量誤差修正模型誤差項常態性與序列相關檢定、重要統計值(RSS、AIC、SC)等，用來作為選擇最適模型的依據。經由實證結果我們發現，不論是否加入股市成交量，模型三都是最適當的模型，也就是資料有不為零的平均數與線性趨勢，但共積方程式只有截距項。 就貨幣需求共積方程式殘差對各變數的影響來看，M1A 與 M1B 的連續增加，都會使股市成交量擴大，而 M1B 的連續增加還會形成物價上漲的壓力。而就物價共積方程式殘差對各變數的影響來看，解釋上較不容易。這可能是因為台灣地區物價長期處於穩定，加上台灣股市受到心理及消息面的影響性很大，要用總體變數作一個完整的解釋並不十分容易。雖然如此，貨幣市場與股票市場間的互動仍然極具有研究價值。 / Traditionally, when estimating the money demand, we use price index, income, and interest rate as its influcing factors. But the stock market that is booming these years has made certain influence on money demand. Milton Friedman pointed out that there are 4 effects that stock market can influcnce money demand. They are trading effect, portfolio reconstruction effect, wealth effect, and subsitution effect. Among these effects, subsitution effect has negative influence on money demand and other 3 effects have positive influence on mondy demand. However, not only does the stock market has influence on mondy demand, money demand also has influence on stock market. In my thesis, I applied Johansen Procedure estimation method. First, I established a traditional model on money demand. The variables I used including money demand, price index, income, and interest rate. From the empirical outcome we are sure that there are 2 cointegration equations among these variables.One is the money demand cointegration equation and the other is the price cointegration equation. Next we add the stock trading volume to the model. We also make sure that there are 2 cointegration relationships among them. There are 5 models in Johansen Procedure estimation method, and they are applied in different situations. It is not easy to decide which model to apply in advance. So in the thesis, we used many criteria, including the value and the sign of the coefficients, the the serial correlation and the normality test of the residuals from the vector error correction model, and important statistics(RSS, AIC, SC) to decide which model to apply. According to the empirical outcome, whether stock trading volume is included, model 3, which is there are means and linear trend in data but the cointegration equation only has intercept is the proper model we selected. About the residuals from the money demand cointegration quation's influence on variables, we find that the continuous increase in M1A and M1B will make enlarge the stock trading volume. Besides, the coutinuous increase in M1B will cause the price to raise. And about the residuals from the price cointegration equation's influence on variables, it is a little bit difficult to interpret. Maybe it is because the price is very stable in Taiwan and the stock market in Taiwan is affected by psychology side and information side easily. So it is not easy to use the macro economic variables to interpret fully. Althought it is the case, the interaction between the money market and the stock market still worth researching.

本文利用Hendry's ECM Methodology的內容，建立台灣貨幣需求模形，並計算其彈性以與國內相關貨幣需求文獻比較。

為了瞭解台灣地區貨幣總計數與經濟活動（如物價、所得）間，是會受到 金融自由化、國際化之影響，故在研究體系中，納入傳統貨幣需求函數中 ，所未放入之變數－－匯率，藉以進一步了解變數間的互動。本文採用 Johansen and Juselius (1990,1992) 發展之共整合分析，以最大概似法 估計共整合向量; 並利用概似比檢定法導出共整合向量個數之檢定統計量 ，及其極限分配；再利用 Johansen 法對匯率進行「弱外生性」檢定； 最後仍利用Granger(1969,1980,1988) 所提出之Granger Causality之觀 念，來探究何種貨幣總計數對GNP 平減指數具有較佳之解釋能力。依據本 文之實證結果，在選擇作為貨幣政策中間目標之指標變數時， M1B 或M1 C 較 M2 及 M1A 為佳。 / The relationships between money aggregates, price, income, interest rate and exchange rate in Taiwan has been investigated in this paper. The Johansen procedure is adopted to estimate cointegration vectors between these variables. Further, weak ex- ogeneity of variables is also tested by way of Johansen procedu- re. Finally, the Granger causality of money and price is invest- igated fo the sake of understanding of predictibility. Main res- ults of this paper can be summarized as follows: First, the nar- row money aggregates, say M1A, M1B and M1C, are closer related to price than M2; M2 is closer related to real GNP than narrow money aggregates. Second, the inclusion of exchange rate and in- terest rate facilitates cointegration relationships between var- iables. Third, the statement "money Granger cause price" is more adequate than "price Granger cause money" in Taiwan. According to above conclusions, it is reasonable to say that M1B and M1C could be better intermediary indicators than M2 for monetary po- licies of stablizing domestic price.

Diese Dissertation beinhaltet vier Aufsätze, die zur Literatur der Panelkointegrationsmethodik beitragen. Der erste Aufsatz vergleicht die Eigenschaften der vier Residuen-basierten Panelkointegrationstests von Pedroni (1995, 1999) mit dem Likelihood-basierten Panelkointegrationstest von Larsson et al. (2001) in endlichen Stichproben. Die Simulationsergebnisse zeigen, dass unter den fünf untersuchten Panelkointegrationsteststatistiken die Panel-t Teststatistik von Pedroni (1995, 1999) die besten Eigenschaften in endlichen Stichproben besitzt. Der zweite Aufsatz präsentiert eine Korrektur des Beweises von Larsson et al. (2001) bezüglich der Endlichkeit der Momente der asymptotischen Trace-Statistik für den Fall, dass die Differenz zwischen der Anzahl der Variablen und der Anzahl der existierenden Kointegrationsbeziehungen eins ist. Im dritten Aufsatz wird ein neuer Likelihood-basierter Panelkointegrationstest vorgestellt, der die Existenz eines linearen Trends in dem datengenerierenden Prozess erlaubt. Dieser neue Test ist eine Erweiterung des Likelihood-Quotienten-Tests von Saikkonen und Lütkepohl (2000a) für trendbereinigte Daten auf die Paneldatenanalyse. Unter der Nullhypothese folgt die Panel-SL Teststatistik einer standardisierten Normalverteilung, wenn die Anzahl der Beobachtungen über die Zeit (T) und die Anzahl der Querschnitte (N) sequentiell gegen unendlich gehen. In einer Monte-Carlo-Studie werden die Eigenschaften des Panel-SL Tests in endlichen Stichproben untersucht. Der neue Test hat ein annehmbares empirisches Signifikanzniveau für wachsende T und N sowie eine hohe Güte in kleinen Stichproben. Der letzte Aufsatz der Dissertation analysiert die langfristige Geldnachfragefunktion in OECD Ländern mit Hilfe von Paneleinheitswurzel- und Panelkointegrationstests. Um eine mögliche Existenz einer stationären langfristigen Geldnachfragefunktion zu untersuchen, werden der Panel-SL Kointegrationstest und die Tests von Pedroni (1999) verwendet. Im Anschluss daran wird eine Paneldatenschätzung für die Geldnachfragefunktion mittels der dynamischen Kleinste-Quadrate-Methode von Mark und Sul (2003) durchgeführt. / This thesis is composed of four essays which contribute to the literature in panel cointegration methodology. The first essay compares the finite sample properties of the four residual-based panel cointegration tests of Pedroni (1995, 1999) and the likelihood-based panel cointegration test of Larsson et al. (2001). The simulation results indicate that the panel-t test statistic of Pedroni has the best finite sample properties among the five panel cointegration test statistics evaluated. The second essay presents a corrected version of the proof of Larsson et al. (2001) related to the finiteness of the moments of the asymptotic trace statistic. The proof is corrected for the case, in which the difference between the number of variables and the number of existing cointegrating relations is one. The third essay proposes a new likelihood-based panel cointegration test in the presence of a linear time trend in the data generating process. This new test is an extension of the likelihood ratio test of Saikkonen and Lütkepohl (2000) for trend-adjusted data to the panel data framework, and is called the panel SL test. Under the null hypothesis, the panel SL test statistic is standard normally distributed as the number of time periods (T) and the number of cross-sections (N) tend to infinity sequentially. By means of a Monte Carlo study the finite sample properties of the test are investigated. The new test presents reasonable size with the increase in T and N, and has high power in small samples. The last essay of the thesis analyzes the long-run money demand relation among OECD countries by panel unit root and cointegration testing techniques. The panel SL cointegration test and the tests of Pedroni (1999) are used to detect the existence of a stationary long-run money demand relation. Moreover, the money demand function is estimated with the panel dynamic ordinary least squares method of Mark and Sul (2003).

Submitted by David Turchick (davidturchick@fgvmail.br) on 2011-03-28T21:58:20Z No. of bitstreams: 1 tese_dout_david.pdf: 1028717 bytes, checksum: ebae39f4145a0e8d55de18fe3b119f10 (MD5) / Approved for entry into archive by Andrea Virginio Machado(andrea.machado@fgv.br) on 2011-03-29T11:47:06Z (GMT) No. of bitstreams: 1 tese_dout_david.pdf: 1028717 bytes, checksum: ebae39f4145a0e8d55de18fe3b119f10 (MD5) / Made available in DSpace on 2011-03-31T17:21:14Z (GMT). No. of bitstreams: 1 tese_dout_david.pdf: 1028717 bytes, checksum: ebae39f4145a0e8d55de18fe3b119f10 (MD5) Previous issue date: 2010-09-24 / This thesis collects four papers on monetary economics written under the supervision of Professor Rubens Penha Cysne. The first of these papers assesses the bias occuring in welfare-cost-of-inflation measures due to failing to take into consideration the substitution potential of interest-bearing monies such as bank deposits. The second one tackles the theoretical issue of comparing the generality of the money-in-the-utility-function- and the shopping-time models by studying the properties of the demand curves they generate. The third of these works revisits a classic paper by Stanley Fischer on the correlation between the growth rate of money supply and the rate of capital accumulation on the transition path. Finally, the fourth one concerns the relative standing of each one of six measures of the welfare cost of inflation (one of which is new) with respect to the other five, and an estimate of the maximum relative error one can incur by choosing to employ a particular welfare measure in place of the others. / Esta tese é uma coleção de quatro artigos em economia monetária escritos sob a supervisão do Professor Rubens Penha Cysne. O primeiro desses artigos calcula o viés presente em medidas do custo de bem-estar da inflação devido a não se levar em conta o potencial substitutivo de moedas que rendem juros, como depósitos bancários.[1] O segundo se concentra na questão teórica de se comparar os escopos dos tradicionais modelos money-in-the-utility-function e shopping-time através do estudo das propriedades das curvas de demanda que eles geram.[2] O terceiro desses trabalhos revisita um artigo clássico de Stanley Fischer sobre a correlação entre a taxa de crescimento da oferta monetária e a taxa de acumulação de capital no caminho de transição.[3] Finalmente, o quarto diz respeito à posição relativa de cada uma de seis medidas do custo de bem-estar da inflação (uma das quais é nova) em relação às outras cinco, e uma estimativa do erro relativo máximo em que o pesquisador pode incorrer devido a sua escolha de empregar uma dessas medidas qualquer vis-à-vis as outras.