Abstract The goal of this thesis is studying the effect of the main macroeconomic fundamentals in boom and bust states on U.S. real house prices, during the period between 1980 and 2021, using a Markov switching dynamic regression (MSDR) model (Hamilton, 1989). The employed model is able to investigate the non-linear relationship between house prices and changes in macroeconomic variables in different regimes and allows us to identify which are the periods of boom and bust in house prices, using the graphical representation of the smoothed probabilities. Different MSDR models have been performed, with different number of states and variances, and the best MSDR model have been chosen considering the lowest Akaike information criterion (AIC), because it should minimize the information loss. The data employed are 165 U.S. quarterly observations from 1980 Q1 (1st January 1980) to 2021 Q1 (1st January 2021), downloaded from the online database of the Federal Reserve Bank of St. Louis (FRED 2021) and Bloomberg (Bloomberg 2021). The growth rate transformation of the time series data have been considered to solve the problem of non-stationarity. The model consider the following regressors: the U.S. real GDP per-capita, that allow us to capture the overall national wealth, which is expected to have a positive effect on house prices (Tsatsaronis and Zhu, 2004); the U.S. inflation rate, which is expected to have a negative impact on U.S. house prices (Tsatsaronis and Zhu, 2004); as short term interest rate the 3-month Treasury Bill Secondary market rate, which has a negative effect on property prices (Adams and Füss, 2010); the term structure of interest rate, which is the difference between the Constant maturity rate on 10-year Treasury, as long term rate, and the 3-month Treasury Bill Secondary market rate, which has a negative impact on house prices (Nneji, Brooks, and Ward, 2013); finally the U.S. working-age population, which is expected to have a positive effect on property prices (Takats, 2012). Considering the model results, in the crash regime, only the per-capita GDP’s regressor is significant below 10% and positive as expected. When one of the GDP’s components increase, the demand for houses increases causing the increase in the house price. In the second regime, so the boom regime, only the regressors of percapita GDP and short term rate are significant below 10% and present the expected sign. In boom regime we have a positive effect of the per-capita GDP on house prices, so when one of the GDP’s components increase, the demand for houses and its price increase too, and a negative effect of the short term rate because when the 3-month Treasury Bill Secondary market rate increases, the mortgage rates and the cost of financing for construction firms increase, causing the rise in house prices. The above results have important policy implications. Policy makers can control house prices in both crash and boom periods by intervening on consumption, investment, government spending and net exports, so one of the GDP’s components. However, controlling the 3-month Treasury Bill Secondary market rate has an effect on house prices only in boom periods. To conclude, understanding what moves house prices can help policy makers to prevent periodical peaks and drops, that can cause the impossibility for citizens to buy an house in boom periods or the destruction of wealth for householders and investors in crash periods.
Riassunto L'obiettivo di questa tesi è quello di studiare l'effetto dei principali fattori macroeconomici sui prezzi reali delle case statunitensi nei periodi di boom e bust tra gli anni 1980 e 2021, utilizzando un modello Markov Switching Dynamic Regression (MSDR) (Hamilton, 1989). Il modello utilizzato è in grado di indagare la relazione non lineare tra i prezzi delle abitazioni e le variazioni delle variabili macroeconomiche nei diversi regimi e permette di identificare quali sono i periodi di boom e bust dei prezzi delle abitazioni, utilizzando la rappresentazione grafica delle probabilità smoothed. Sono stati eseguiti diversi modelli MSDR, considerando diversi stati e varianze, e il miglior modello è stato selezionato considerando il valore AIC più basso. I dati utilizzati sono 165 osservazioni trimestrali statunitensi, dal primo trimestre del 1980 (1 gennaio 1980) al primo trimestre del 2021 (1 gennaio 2021), scaricate dal database online della Federal Reserve Bank di St. Louis (FRED 2021) e Bloomberg (Bloomberg 2021). Sono state considerate le trasformazioni del tasso di crescita delle serie temporali per risolvere il problema della non stazionarietà. I regressori considerati sono: il PIL reale pro-capite degli Stati Uniti, che ci consente di catturare la ricchezza nazionale complessiva, che dovrebbe avere un effetto positivo sui prezzi delle case (Tsatsaronis e Zhu, 2004); il tasso di inflazione degli Stati Uniti, che dovrebbe avere un impatto negativo sui prezzi delle case negli Stati Uniti (Tsatsaronis e Zhu, 2004); come tasso di interesse a breve termine, il tasso di mercato secondario del Buono del Tesoro a 3 mesi (3-month Treasury Bill Secondary market rate), che ha un effetto negativo sui prezzi degli immobili (Adams e Füss, 2010); la struttura a termine del tasso di interesse, che è la differenza tra il Constant maturity rate on 10-year Treasury, come tasso a lungo termine, e il 3-month Treasury Bill Secondary market rate, che ha un impatto negativo sui prezzi delle case (Nneji, Brooks e Ward, 2013); infine la popolazione in età lavorativa degli Stati Uniti, che dovrebbe avere un effetto positivo sui prezzi degli immobili (Takats, 2012). Considerando i risultati del modello, nel regime di crash, solo il regressore del PIL pro-capite è significativo al 10% e positivo come previsto. Quando una delle componenti del PIL aumenta, la domanda di case aumenta provocando l'aumento del prezzo delle case. Nel secondo regime, quindi il regime del boom, solo i regressori del PIL pro-capite e del tasso a breve sono significativi al 10% e presentano il segno atteso. In regime di boom si ha un effetto positivo del PIL pro-capite sui prezzi delle case e un effetto negativo del tasso a breve scadenza perché quando il 3-month Treasury Bill Secondary market rate aumenta, aumentano i tassi dei mutui e il costo del finanziamento per le imprese di costruzione, provocando l'aumento del prezzo delle case. I risultati di cui sopra hanno importanti implicazioni politiche. I principali policy maker possono controllare i prezzi delle case sia nei periodi di crisi che in quelli di boom intervenendo su consumi, investimenti, spesa pubblica ed esportazioni nette, quindi una delle componenti del PIL. Tuttavia, il controllo del 3-month Treasury Bill Secondary market rate ha un effetto sui prezzi delle case solo nei periodi di boom. Per concludere, capire cosa muove i prezzi delle case può aiutare i policy maker a prevenire picchi e cali periodici, che possono causare l'impossibilità per i cittadini di acquistare una casa nei periodi di boom o la distruzione della ricchezza per le famiglie e gli investitori nei periodi di crisi.
Determinanti macroeconomiche dei prezzi delle case statunitensi nei periodi di boom e bust
CHIESA, SARA
2021/2022
Abstract
Abstract The goal of this thesis is studying the effect of the main macroeconomic fundamentals in boom and bust states on U.S. real house prices, during the period between 1980 and 2021, using a Markov switching dynamic regression (MSDR) model (Hamilton, 1989). The employed model is able to investigate the non-linear relationship between house prices and changes in macroeconomic variables in different regimes and allows us to identify which are the periods of boom and bust in house prices, using the graphical representation of the smoothed probabilities. Different MSDR models have been performed, with different number of states and variances, and the best MSDR model have been chosen considering the lowest Akaike information criterion (AIC), because it should minimize the information loss. The data employed are 165 U.S. quarterly observations from 1980 Q1 (1st January 1980) to 2021 Q1 (1st January 2021), downloaded from the online database of the Federal Reserve Bank of St. Louis (FRED 2021) and Bloomberg (Bloomberg 2021). The growth rate transformation of the time series data have been considered to solve the problem of non-stationarity. The model consider the following regressors: the U.S. real GDP per-capita, that allow us to capture the overall national wealth, which is expected to have a positive effect on house prices (Tsatsaronis and Zhu, 2004); the U.S. inflation rate, which is expected to have a negative impact on U.S. house prices (Tsatsaronis and Zhu, 2004); as short term interest rate the 3-month Treasury Bill Secondary market rate, which has a negative effect on property prices (Adams and Füss, 2010); the term structure of interest rate, which is the difference between the Constant maturity rate on 10-year Treasury, as long term rate, and the 3-month Treasury Bill Secondary market rate, which has a negative impact on house prices (Nneji, Brooks, and Ward, 2013); finally the U.S. working-age population, which is expected to have a positive effect on property prices (Takats, 2012). Considering the model results, in the crash regime, only the per-capita GDP’s regressor is significant below 10% and positive as expected. When one of the GDP’s components increase, the demand for houses increases causing the increase in the house price. In the second regime, so the boom regime, only the regressors of percapita GDP and short term rate are significant below 10% and present the expected sign. In boom regime we have a positive effect of the per-capita GDP on house prices, so when one of the GDP’s components increase, the demand for houses and its price increase too, and a negative effect of the short term rate because when the 3-month Treasury Bill Secondary market rate increases, the mortgage rates and the cost of financing for construction firms increase, causing the rise in house prices. The above results have important policy implications. Policy makers can control house prices in both crash and boom periods by intervening on consumption, investment, government spending and net exports, so one of the GDP’s components. However, controlling the 3-month Treasury Bill Secondary market rate has an effect on house prices only in boom periods. To conclude, understanding what moves house prices can help policy makers to prevent periodical peaks and drops, that can cause the impossibility for citizens to buy an house in boom periods or the destruction of wealth for householders and investors in crash periods.È consentito all'utente scaricare e condividere i documenti disponibili a testo pieno in UNITESI UNIPV nel rispetto della licenza Creative Commons del tipo CC BY NC ND.
Per maggiori informazioni e per verifiche sull'eventuale disponibilità del file scrivere a: unitesi@unipv.it.
https://hdl.handle.net/20.500.14239/1808