My dissertation consists of three chapters. In the first chapter “Dancing with the R*: Information Shocks in the "New Normal"”, we quantify the impact of information shocks — signals about the state of the economy conveyed through central bank policy communications — on the real natural rate of interest (R*). Using an internal instruments VAR approach, we find that information shocks associated with policy rate decisions have a statistically significant, economically meaningful, and highly persistent effect on R*. We show that the response of R* to information shocks may operate through a productivity channel and that the effect is sign asymmetric — stronger following unexpected rate cuts than after rate hikes. Overall, our findings underscore the critical importance of central bank communication: if not carefully managed, policy signals can exert powerful and unintended effects on macroeconomic dynamics. In the second chapter “Looser, Tighter, Clearer: A New Financial Conditions Index for the Euro Area”, we proposes the Macro-Finance Financial condition Index ("MF-FCI") for the euro area. The new index is estimated within a novel macro-finance vector autoregressive model where the financial variables are collapsed into one linear combination — the MF-FCI — that best explains the joint dynamics of key macroeconomic variables. The new methodology jointly estimates the FCI weights and VAR coefficients, allowing for mutual feedback between macroeconomic and financial variable blocks. This framework, among other features, enables the estimation of model-implied impulse response functions, supporting dynamic analysis of the transmission of financial conditions to the broader economy. The resulting MF-FCI (i) offers a transparent decomposition into its underlying financial components over time and (ii) outperforms existing euro area FCIs in both in-sample fit and out-of-sample forecasting accuracy. In the third chapter Deconstructing Financial Conditions Using a VAR-based Macro Random Forest, I propose a time-varying extension of the MF-FCI model, presented in the second Chapter. The new methodology for decomposing financial conditions with time-varying weights is based on a macro random forest framework. By capturing dynamic changes in the relationship between financial variables and the broader economy, this methodology improves the flexibility and responsiveness of the FCI. The time-varying structure enables a straightforward interpretation of individual financial variables’ contributions to financial conditions incorporating the standard VAR literature in a more advanced macro-random forest setting.
Il mio lavoro di tesi è composto da tre capitoli. Nel primo capitolo, “Dancing with the R*: Information Shocks in the "New Normal"”, quantifichiamo l’impatto degli information shocks — segnali sullo stato dell’economia trasmessi attraverso le comunicazioni di politica monetaria delle banche centrali — sul tasso di interesse naturale reale (R). Utilizzando un approccio VAR con strumenti interni (internal instruments VAR), troviamo che gli information shocks associati alle decisioni sui tassi di policy hanno un effetto statisticamente significativo, economicamente rilevante e altamente persistente su R*. Mostriamo che la risposta di R* agli information shocks può operare attraverso un canale di produttività e che l’effetto è asimmetrico nel segno — più forte in seguito a riduzioni inattese dei tassi rispetto agli aumenti dei tassi. Nel complesso, i nostri risultati sottolineano l’importanza cruciale della comunicazione delle banche centrali: se non gestiti con attenzione, i segnali di policy possono esercitare effetti potenti e non intenzionali sulle dinamiche macroeconomiche. Nel secondo capitolo, “Looser, Tighter, Clearer: A New Financial Conditions Index for the Euro Area”, proponiamo il Macro-Finance Financial Conditions Index (“MF-FCI”) per l’area dell’euro. Il nuovo indice è stimato all’interno di un innovativo modello VAR macro-finanziario, nel quale le variabili finanziarie sono sintetizzate in una combinazione lineare — l’MF-FCI — che spiega al meglio la dinamica congiunta delle principali variabili macroeconomiche. La nuova metodologia stima congiuntamente i pesi dell’FCI e i coefficienti del VAR, consentendo un feedback tra il blocco delle variabili macroeconomiche e quello delle variabili finanziarie. Questo quadro metodologico, tra le altre caratteristiche, permette la stima delle funzioni di risposta all’impulso (IRF) implicite nel modello, supportando l’analisi dinamica della trasmissione delle condizioni finanziarie all’economia nel suo complesso. L’MF-FCI risultante (i) offre una scomposizione trasparente nelle sue componenti finanziarie sottostanti nel tempo e (ii) supera gli indici FCI esistenti per l’area euro sia in termini di previsione in-sample sia di accuratezza out-of-sample. Nel terzo capitolo, “Deconstructing Financial Conditions Using a VAR-based Macro Random Forest”, propongo un’estensione del modello MF-FCI presentato nel capitolo precedente che accomoda la stima di pesi variabili nel tempo. La nuova metodologia per la scomposizione delle condizioni finanziarie con pesi variabili nel tempo si basa su un quadro di macro random forest. Catturando i cambiamenti dinamici nella relazione tra le variabili finanziarie e l’economia nel suo complesso, questa metodologia migliora la flessibilità e la capacità di risposta dell’FCI. La struttura a parametri variabili consente inoltre un’interpretazione immediata del contributo delle singole variabili finanziarie alle condizioni finanziarie, incorporando la letteratura VAR standard in un contesto più avanzato di machine learning.
Martorana, Giulia, ESSAY ON THE TRANSMISSION OF MONETARY POLICY IN A CHANGING WORLD: AN EMPIRICAL ANALYSIS, Castelnuovo, Efrem, Bletzinger, Tilman, Università Cattolica del Sacro Cuore MILANO:Ciclo XXXVII. [doi:10.83049/unicatt/publicatt/10807_332121] [https://hdl.handle.net/10807/332121] [http://dx.doi.org/10.83049/unicatt/publicatt/10807_332121]
ESSAY ON THE TRANSMISSION OF MONETARY POLICY IN A CHANGING WORLD: AN EMPIRICAL ANALYSIS
Martorana, Giulia
2026
Abstract
My dissertation consists of three chapters. In the first chapter “Dancing with the R*: Information Shocks in the "New Normal"”, we quantify the impact of information shocks — signals about the state of the economy conveyed through central bank policy communications — on the real natural rate of interest (R*). Using an internal instruments VAR approach, we find that information shocks associated with policy rate decisions have a statistically significant, economically meaningful, and highly persistent effect on R*. We show that the response of R* to information shocks may operate through a productivity channel and that the effect is sign asymmetric — stronger following unexpected rate cuts than after rate hikes. Overall, our findings underscore the critical importance of central bank communication: if not carefully managed, policy signals can exert powerful and unintended effects on macroeconomic dynamics. In the second chapter “Looser, Tighter, Clearer: A New Financial Conditions Index for the Euro Area”, we proposes the Macro-Finance Financial condition Index ("MF-FCI") for the euro area. The new index is estimated within a novel macro-finance vector autoregressive model where the financial variables are collapsed into one linear combination — the MF-FCI — that best explains the joint dynamics of key macroeconomic variables. The new methodology jointly estimates the FCI weights and VAR coefficients, allowing for mutual feedback between macroeconomic and financial variable blocks. This framework, among other features, enables the estimation of model-implied impulse response functions, supporting dynamic analysis of the transmission of financial conditions to the broader economy. The resulting MF-FCI (i) offers a transparent decomposition into its underlying financial components over time and (ii) outperforms existing euro area FCIs in both in-sample fit and out-of-sample forecasting accuracy. In the third chapter Deconstructing Financial Conditions Using a VAR-based Macro Random Forest, I propose a time-varying extension of the MF-FCI model, presented in the second Chapter. The new methodology for decomposing financial conditions with time-varying weights is based on a macro random forest framework. By capturing dynamic changes in the relationship between financial variables and the broader economy, this methodology improves the flexibility and responsiveness of the FCI. The time-varying structure enables a straightforward interpretation of individual financial variables’ contributions to financial conditions incorporating the standard VAR literature in a more advanced macro-random forest setting.| File | Dimensione | Formato | |
|---|---|---|---|
|
tesiphd_completa_Martorana.pdf
embargo fino al 01/03/2028
Licenza:
Creative commons
Dimensione
6.45 MB
Formato
Adobe PDF
|
6.45 MB | Adobe PDF | Visualizza/Apri |
I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.



