DO DIVIDEND FLOWS AFFECT STOCK RETURNS?

• Date: 01 March 2018
• Published date: 01 March 2018
• Author: Snorre Lindset,Joakim Kvamvold
• DOI: http://doi.org/10.1111/jfir.12142

DO DIVIDEND FLOWS AFFECT STOCK RETURNS?

Joakim Kvamvold

Folketrygdfondet

Snorre Lindset

Norwegian University of Science and Technology

Abstract

We examine price impacts from dividend flows. Event-study estimates show that stocks

experience abnormal returns on the dividend distribution day. Results also show a

spillover effect to non-dividend-paying stocks that are likely to be part of the same

benchmark portfolio as the dividend-paying stocks. Regression results indicate that the

effect is dependent on the ownership share by professional investors. The temporary

nature of the effect on returns is in line with the literature’s demand-driven price pressure

hypothesis.

JEL Classification: G11, G12, G14, G23

I. Introduction

A large body of literature finds that stock prices are positively correlated with flows to

investors. One type of flow is distributions of dividends. In this article, we analyze how

dividend distributions affect stock prices. To this end, we analyze a portfolio consisting

of stocks from two categories. For each trading date, we distinguish between a

subportfolio consisting of dividend payers and a subportfolio consisting of non-dividend

payers. Our empirical results show a clear, positive relation between the dividend

distributions and the returns on the dividend-paying stocks. The results also indicate

that there is a positive relation between the distributions and the returns on the non-

dividend-paying stocks.

Four dates are important in the dividend payment process (see Figure I). At the

declaration date, the dividend-paying company announces the ex-dividend, record, and

We are grateful to Lars-Erik Borge, Honghui Chen, Vladimir Gatchev, Torgeir Kråkenes, Ajai Singh, Magne

Valen-Senstad, and Qinghai Wang for comments and discussions. In particular, we thank the associate editor (Paul

Koch) and an anonymous referee for valuable comments. We also thank participants at the 6th Research School

Conference organized by The National Research School in Business Economics and Administration, the 2015 Paris

Financial Management Conference, and the 23rd Annual Conference of the Multinational Society for valuable

comments. Parts of this article were written while Kvamvold was a visiting scholar at Columbia Business School

and while Lindset was a visiting scholar at the University of Central Florida. It also appeared as chapter 1 in

Kvamvold’s PhD thesis at the Norwegian University of Science and Technology. Kvamvold is currently employed

at Folketrygdfondet (manager of the Government Pension Fund Norway). The views in this article are those of the

authors and do not necessarily reflect the views of Folketrygdfondet. This research did not receive any specific

grant from funding agencies in the public, commercial, or not-for-profit sectors.

The Journal of Financial Research Vol. XLI, No. 1 Pages 149–174 Spring 2018

149

© 2018 The Southern Finance Association and the Southwestern Finance Association

payment dates. The size of the dividend and all other relevant information are also made

public. Thus, no new information regarding the dividend payment is made available to

the market after the declaration date. All holders of the company’s stock before the

ex-dividend date are entitled to the dividend payment. After the ex-dividend date, buyers

of the stock do not have the right to receive the dividend. The record date is usually two

trading days after the ex-dividend date. All holders of the stock on record receive the

dividend. The record date is set so that the company can get on record all investors that

held the stock one day before the ex-dividend date. Finally, the dividend is transferred to

investors on the payment date. The payment date is usually two to four weeks after the

ex-dividend date. Some companies offer investors the opportunity to participate in

dividend reinvestment plans (DRIPs). If an investor participates in such a plan, dividends

are automatically reinvested in the stock of the dividend-paying company.

Most mutual funds and institutional investors measure their portfolio’s return

against a benchmark index and have upper bounds on tracking errors. When an index

constituent goes ex-dividend, the index provider typically “reinvests”the dividend in all

the index constituents according to their market capitalization. Unless investors have

cash in their portfolio, they cannot do similarly until they receive the dividends on

the distribution day. This time gap between the change in the index on the ex-day and the

distribution date poses a problem for portfolio managers with a tight tracking error as

their exposure to the portfolio to which they are benchmarked becomes too low. They

therefore have an incentive to reinvest the dividends as early as possible. They also have

an incentive to invest in such a way that they obtain the right exposure to their

benchmark, that is, to invest broadly in the constituents.

In a frictionless market, with equally well-informed investors, unexpected

changes in asset prices are a result of new information. Edelen and Warner (2001) find

that flows to investors and stock returns are thus positively correlated. That flows contain

new information is a common problem when analyzing price impacts. With new

information, it is difficult to disentangle any demand effect on prices from the effect of

new information. Little attention has been devoted to distributions of dividends, which

add to flows to investors. A particular feature of analyzing dividend distributions is that

they are not associated with any new information. The announcement of the size of the

dividend payment is made weeks before the distribution date. When investors receive the

dividends, the size and timing of the payment merely confirm what they already know. If

Figure I. Dividend Payment Process. This figure illustrates the dates associated with the dividend payment

process. The distance between the dots is not proportional to the expected number of days between the

dates.

150 The Journal of Financial Research