The forecast is a natural attitude of men and women: forecasts are made because to make certain decisions today, we need to know how the world will be tomorrow, and therefore how the future state of the world will influence the result of the choices. In finance, the term "forecast" refers to the expectations of individuals on the future trend of the variables studied, based on information or intuitions, starting from the assumption that individuals have a good knowledge of the system in which they operate and political economy. This thesis, developed in three chapters, has as its main theme the irrationality of investors generated by the inefficiency of information. The analysis was carried out on the euro-dollar forecasts presented by various institutions such as banks, research divisions of banks and research centers. The available data comes from the Bloomberg platform. In the first chapter a model was developed that estimates the forecast errors as parameters for the evaluation of the predictive capacity of each predictor, showing whether the difference in absolute value between the spot rate and the forecast or revision issued increases, decreases or remains unchanged in the time to assess the ability of financial institutions to incorporate new information in a complete and timely manner in the time path leading to the given terminal for which the forecast was issued. The results in accordance with the existing literature showed that the revisions worsen the previously issued forecasts rather than improving them, in all the horizons considered. This means that predictors fail to learn from their mistakes and therefore fail to incorporate new information efficiently. Furthermore, it was shown that the predictors would have obtained a prediction error in a lower absolute value, using the random walk model or by issuing a forecast equal to the known spot rate at the time of the forecast. Furthermore, in order to reinforce the results achieved in the first chapter, the predictions of the available predictors were analyzed through the Hurst test. The application of a long-term memory coefficient makes it possible to understand to what extent past forecasts influence future forecasts. The greater the value of the coefficient in a range between 0 and 1, the longer the historical memory series will be. The results showed values almost always greater than 0.5, at which point the efficiency of the analyzed variables can be affirmed. Finally, in the third chapter of this document, the statistical test of Toda and Yamamoto (1995) was applied with the aim of showing that predictors, although clearly inefficient, as shown by the analysis of random walking (first chapter) and by Hurst tests (second chapter) are connected by a cause and effect mechanism, which shows the existence of a mass inefficiency. The idea was to analyze whether the inefficiency of long-term predictors is generated or not by a cause and effect mechanism, which drives the most important banks to act as market leaders, influencing everyone's choices and forecasts other financial institutions.
La previsione è un atteggiamento naturale di uomini e donne: si fanno previsioni perché per prendere determinate decisioni oggi, bisogna sapere come sarà il mondo domani, e quindi come il futuro stato del mondo influenzerà il risultato delle scelte. In finanza, il termine "previsione" si riferisce alle aspettative degli individui sulla tendenza futura delle variabili studiate, sulla base di informazioni o intuizioni, a partire dal presupposto che gli individui abbiano una buona conoscenza del sistema in cui operano e dell'economia politica. Questa tesi, sviluppata in tre capitoli, ha come tema principale l'irrazionalità degli investitori generata dall'inefficienza delle informazioni. L'analisi è stata effettuata sulle previsioni in valuta euro-dollaro presentate da varie istituzioni come banche, divisioni di ricerca di banche e centri di ricerca. I dati disponibili provengono dalla piattaforma Bloomberg. Nel primo capitolo è stato sviluppato un modello che stima gli errori di previsione come parametri per la valutazione della capacità predittiva di ciascun predittore, mostrando se la differenza di valore assoluto tra il tasso spot e la previsione o revisione emessa aumenta, diminuisce o rimane invariata nel tempo per valutare la capacità degli istituti finanziari di incorporare nuove informazioni in modo completo e tempestivo nel percorso temporale che porta alla data terminale per la quale è stata emessa la previsione. I risultati in accordo con la letteratura esistente hanno mostrato che le revisioni peggiorano le previsioni precedentemente emesse anziché migliorarle, in tutti gli orizzonti considerati. Ciò significa che i predittori non riescono a imparare dai propri errori e quindi non riescono a incorporare nuove informazioni in modo efficiente. Inoltre, è stato dimostrato che i predittori avrebbero ottenuto un errore di previsione in un valore assoluto inferiore, utilizzando il modello di camminata casuale o emettendo una previsione pari al tasso spot noto al momento della previsione. Inoltre, al fine di rafforzare i risultati raggiunti nel primo capitolo, le previsioni dei predittori disponibili sono state analizzate attraverso il test di Hurst. L'applicazione di un coefficiente di memoria a lungo termine permette di capire in che misura le previsioni passate influenzano le previsioni future. Maggiore è il valore del coefficiente in un intervallo compreso tra 0 e 1, maggiore sarà la memoria lunga della serie storica. I risultati hanno mostrato valori quasi sempre maggiori di 0,5, punto in cui si può affermare l'efficienza delle variabili analizzate. Infine, nel terzo capitolo di questo documento, è stato applicato il test statistico di Toda e Yamamoto (1995) con l'obiettivo di mostrare che i predittori, sebbene chiaramente inefficienti, come dimostrato dall'analisi del camminare casuale (primo capitolo) e dai test di Hurst (secondo capitolo) sono collegati da un meccanismo di causa ed effetto, che mostra l'esistenza di un'inefficienza di massa. L'idea è stata quella di analizzare se l'inefficienza dei predittori a lungo termine sia generata o meno da un meccanismo di causa ed effetto, che spinge le banche più importanti ad agire come leader del mercato, influenzando le scelte e le previsioni di tutti gli altri istituti finanziari.
The role of information efficiency in exchange rate forecasts: evidence from survey data
VALERI, GIANLUCA
2020
Abstract
The forecast is a natural attitude of men and women: forecasts are made because to make certain decisions today, we need to know how the world will be tomorrow, and therefore how the future state of the world will influence the result of the choices. In finance, the term "forecast" refers to the expectations of individuals on the future trend of the variables studied, based on information or intuitions, starting from the assumption that individuals have a good knowledge of the system in which they operate and political economy. This thesis, developed in three chapters, has as its main theme the irrationality of investors generated by the inefficiency of information. The analysis was carried out on the euro-dollar forecasts presented by various institutions such as banks, research divisions of banks and research centers. The available data comes from the Bloomberg platform. In the first chapter a model was developed that estimates the forecast errors as parameters for the evaluation of the predictive capacity of each predictor, showing whether the difference in absolute value between the spot rate and the forecast or revision issued increases, decreases or remains unchanged in the time to assess the ability of financial institutions to incorporate new information in a complete and timely manner in the time path leading to the given terminal for which the forecast was issued. The results in accordance with the existing literature showed that the revisions worsen the previously issued forecasts rather than improving them, in all the horizons considered. This means that predictors fail to learn from their mistakes and therefore fail to incorporate new information efficiently. Furthermore, it was shown that the predictors would have obtained a prediction error in a lower absolute value, using the random walk model or by issuing a forecast equal to the known spot rate at the time of the forecast. Furthermore, in order to reinforce the results achieved in the first chapter, the predictions of the available predictors were analyzed through the Hurst test. The application of a long-term memory coefficient makes it possible to understand to what extent past forecasts influence future forecasts. The greater the value of the coefficient in a range between 0 and 1, the longer the historical memory series will be. The results showed values almost always greater than 0.5, at which point the efficiency of the analyzed variables can be affirmed. Finally, in the third chapter of this document, the statistical test of Toda and Yamamoto (1995) was applied with the aim of showing that predictors, although clearly inefficient, as shown by the analysis of random walking (first chapter) and by Hurst tests (second chapter) are connected by a cause and effect mechanism, which shows the existence of a mass inefficiency. The idea was to analyze whether the inefficiency of long-term predictors is generated or not by a cause and effect mechanism, which drives the most important banks to act as market leaders, influencing everyone's choices and forecasts other financial institutions.File | Dimensione | Formato | |
---|---|---|---|
Tesi Gianluca Valeri.pdf
accesso aperto
Dimensione
8.45 MB
Formato
Adobe PDF
|
8.45 MB | Adobe PDF | Visualizza/Apri |
I documenti in UNITESI sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.
https://hdl.handle.net/20.500.14242/194624
URN:NBN:IT:UNIMC-194624