Return Predictability, Dividend Growth and the Persistence of the Price-Dividend Ratio

Adam Golinski; Joao Antonio Rodrigues Madeira; Dooruj Rambaccussing

Return Predictability, Dividend Growth and the Persistence of the Price-Dividend Ratio

Adam Golinski, Joao Antonio Rodrigues Madeira, Dooruj Rambaccussing

Economics

Research output: Contribution to journal › Article › peer-review

Abstract

Empirical evidence shows that the order of integration of returns and dividend growth is approximately equal to the order of integration of the first differenced
price-dividend ratio, which is about 0.8. Yet, the present-value identity implies the three series should be integrated of the same order. We reconcile this puzzle by showing that the aggregation of antipersistent expected returns and expected dividends gives rise to the price-dividend ratio with properties that mimic long memory in finite samples. In the empirical implementation, we extend and estimate the state-space present-value model by allowing for fractional
integration in expected returns and expected dividend growth. This extension improves the model's forecasting power in-sample and out-of-sample. In addition, expected returns and expected dividend growth modelled as ARFIMA
processes are more closely related to future macroeconomic variables, which makes them suitable as leading business cycle indicators.

Original language	English
Number of pages	47
Journal	International journal of forecasting
Publication status	Accepted/In press - 22 Mar 2024

Embargoed Document

Project_ADJ
Accepted author manuscript, 572 KB
Licence: CC BY
Embargo ends: 31/12/99

Cite this

@article{06fe83797b704b059e6cbe2f994dee0c,

title = "Return Predictability, Dividend Growth and the Persistence of the Price-Dividend Ratio",

abstract = "Empirical evidence shows that the order of integration of returns and dividend growth is approximately equal to the order of integration of the first differencedprice-dividend ratio, which is about 0.8. Yet, the present-value identity implies the three series should be integrated of the same order. We reconcile this puzzle by showing that the aggregation of antipersistent expected returns and expected dividends gives rise to the price-dividend ratio with properties that mimic long memory in finite samples. In the empirical implementation, we extend and estimate the state-space present-value model by allowing for fractional integration in expected returns and expected dividend growth. This extension improves the model's forecasting power in-sample and out-of-sample. In addition, expected returns and expected dividend growth modelled as ARFIMA processes are more closely related to future macroeconomic variables, which makes them suitable as leading business cycle indicators.",

author = "Adam Golinski and {Rodrigues Madeira}, {Joao Antonio} and Dooruj Rambaccussing",

year = "2024",

month = mar,

day = "22",

language = "English",

journal = "International journal of forecasting",

issn = "0169-2070",

publisher = "Elsevier",

}

TY - JOUR

T1 - Return Predictability, Dividend Growth and the Persistence of the Price-Dividend Ratio

AU - Golinski, Adam

AU - Rodrigues Madeira, Joao Antonio

AU - Rambaccussing, Dooruj

PY - 2024/3/22

Y1 - 2024/3/22

N2 - Empirical evidence shows that the order of integration of returns and dividend growth is approximately equal to the order of integration of the first differencedprice-dividend ratio, which is about 0.8. Yet, the present-value identity implies the three series should be integrated of the same order. We reconcile this puzzle by showing that the aggregation of antipersistent expected returns and expected dividends gives rise to the price-dividend ratio with properties that mimic long memory in finite samples. In the empirical implementation, we extend and estimate the state-space present-value model by allowing for fractional integration in expected returns and expected dividend growth. This extension improves the model's forecasting power in-sample and out-of-sample. In addition, expected returns and expected dividend growth modelled as ARFIMA processes are more closely related to future macroeconomic variables, which makes them suitable as leading business cycle indicators.

AB - Empirical evidence shows that the order of integration of returns and dividend growth is approximately equal to the order of integration of the first differencedprice-dividend ratio, which is about 0.8. Yet, the present-value identity implies the three series should be integrated of the same order. We reconcile this puzzle by showing that the aggregation of antipersistent expected returns and expected dividends gives rise to the price-dividend ratio with properties that mimic long memory in finite samples. In the empirical implementation, we extend and estimate the state-space present-value model by allowing for fractional integration in expected returns and expected dividend growth. This extension improves the model's forecasting power in-sample and out-of-sample. In addition, expected returns and expected dividend growth modelled as ARFIMA processes are more closely related to future macroeconomic variables, which makes them suitable as leading business cycle indicators.

M3 - Article

SN - 0169-2070

JO - International journal of forecasting

JF - International journal of forecasting

ER -