Dividend theory suggests that dividend is gluey and it can be used to signal the hereafter gaining quality of the houses. However, from past empirical groundss a consensus is yet excessively reached as varied decisions are reached. Recent literature on the topic holding given some grounds of the cogency of the ICD hypothesis when houses are analysed based on their firm-specific features, such as growing, adulthood degrees, age and corporate administration manner adopted. Sequel to this, this survey seeks to determine the cogency and cosmopolitan pertinence of this grounds by supplying empirical support of the Information Content Dividend on a cross-section of Corporate Firms listed on the Johannesburg Stock Exchange ( FTSE/JSE All Share Index ) utilizing the cross-sectional Fama and Macbeth arrested development method and panel arrested development methods such as fixed effects and random consequence trials. The consequence obtained show
Key-words: Firm ‘s Net incomes, Dividend Changes, Information Content Dividend hypothesis.
This thesis is the consequence of my ain work. Material from the published or unpublished work of others, which is referred to in the thesis, is credited to the writer in inquiry in the text. The thesis is about 9500 words in length.
Table of Content
Index of Figures
Index of Tables
List of Abbreviations Used
I would wish to admit the Almighty God for His nutriment and Grace bestowed upon me to successfully finish my MSc plan at Durham. My bosom felt gratitude goes out to my parents, full household and friends for all their forfeits, love, supplications and huge support showered on me. I would besides wish to appreciate the supervising rendered by Dr Xing Wang. To all my lectors here in Durham and everyone who in one manner or the other has imparted my life therefore far, I besides say a large thank you. God bless you all
Chapter One: Introduction To The Research.
Research Title: Make Dividend Changes Predict Future Earnings of Firms: An Evidence
1.2 Back-ground to the Study
The topic of dividend policy remains a controversial subject in malice of old ages of extended theoretical and empirical research ( Allen & A ; Rachim ( 1996 ) as cited in Adaramola ( 2012 ) . Evidence from past researches strongly supports the prognostic relationship between dividend alterations, and stock-prices-that is dividend additions ( lessening ) impel an addition ( portion ) such houses. However, despite this widely recognised relationship between portion monetary values and dividend alterations the grounds of the being relationship between dividends and future net incomes is far from conclusive. ( Benartzi et al, 1997 )
The Informational Content of Dividends ( ICD ) states that dividend alterations portray information about the house ‘s public presentation in footings of future profitableness and hard currency flows. Get downing with the polar work of Lintner ( 1956 ) who proposed that houses ‘ increase their dividend pay-out merely when net incomes are expected to increase for good in the hereafter. This proposition was farther supported by the research carried out by Miller and Modigliani ( 1961 ) and this was explicitly coined as the ‘Informational Content of Dividends ‘ hypothesis.
In the recent old ages, varied decisions have been reached with regard to the effectivity of the ICD hypothesis. Amongst which are DeAngelo et Al ( 1996 ) , they obtained consequences proposing that the determination to increase dividends by directors was propelled by the over-optimistic prognosiss about future net incomes after analyzing the content of directors ‘ dividend determinations for 145 NYSE houses whose one-year net incomes declined after nine or more back-to-back old ages of growing. A twelvemonth subsequently Bernartzi et Al ( 1997 ) , closely supported the above findings of DeAngelo et Al, after obtaining consequences proposing a really minimum impact of dividends addition on the house ‘s net incomes _ they conclude that houses that increase dividends experience important net incomes in the twelvemonth earlier and in the same twelvemonth but non in the subsequent old ages, while houses which implemented a cut in dividends, showed a lessening in net incomes in the twelvemonth before the lessening and in the current twelvemonth but thenceforth they went on to demo an addition in subsequent old ages. However, as portion of their findings, they find that houses that increase dividends are less likely to see a bead in future net incomes than in non-changing houses, which proves to be consistent, with Lintner ‘s theoretical account.
Although many documents do non back up ‘the ICD hypothesis, Nissim and Ziv ( 2001 ) and more late _Choi et Al ( 2011 ) amongst others, in their research carried out on US houses and Korean houses severally, give strong support of the theory as they find that dividend alterations possess built-in qualities capable of signalling the degree of profitableness in subsequent old ages, they besides reach the decision that dividend alterations are positively related to alterations in net incomes in each of the two old ages subsequent to the alteration. Although the former was faulted by Grullon et Al ( 2005 ) for an alleged abuse of additive arrested development theoretical accounts the latter Choi et Al ( 2011 ) re-confirms their findings after 10years, giving farther grounds that the ICD hypothesis is better supported when house particular features are incorporated into the analysis.
From the above, it can be deduced that despite extended research undertaken on the consequence of dividend alterations on future net incomes, it still remains one of the unresolved issues in finance. One ground appears to be the changing methodological analysis employed in informations analysis and besides the distinctive features of the different populations in which the trial has been carried out which limits the grounds of the cosmopolitan pertinence of this hypothesis.
This thesis will try to make full the bing spread in the empirical literature by measuring whether and how alterations in dividends maps into future net incomes of Corporate Firms widening this trial to the African Continent utilizing a most recent informations set. Further buttressing, the cogency of old surveies and to determine and give acceptance to its cosmopolitan pertinence, an appraisal and rating of the ICD hypothesis in South Africa will be considered transporting out luxuriant statistical analysis to deduce a concrete decision. It is hoped that the thesis will cast some visible radiation on how to get the better of these restrictions.
1.3. Purposes and Aims.
This thesis seeks to do a part to the turning literature on the ICD hypothesis through the application and extension of Choi et ALSs ‘ ( 2011 ) methodological analysis. By using South African informations associating more recent period ( 1999-2011 ) , the research worker will analyze the hardiness of grounds gathered from past researches on other states. The research worker will besides be traveling farther to prove the cogency of the ICD utilizing an alternate net incomes measurement _Total market value deflated by the capital outgo of houses which is yet to be incorporated in the trial of the ICD hypothesis. Furthermore, as an extension to Choi et Al, the research worker shall transport out the fixed and random consequence panel analysis on the information set.
The survey will be looking to place the periods of dividend alterations and its impact on future net incomes and capital outgo of the Corporate Firms contained in the sample size.
The cardinal purpose of the survey is to reply the inquiry ‘Does Dividend Changes Predict Future Earnings of Firms? ‘ through the probe of the alteration in dividend pay-out and the placeholder for the net incomes for the sample period selected. The dependance of the ICD hypothesis on firm-specific features such as the house growing degree will besides be assessed by spliting the houses into high-growth and low-growth houses. The survey aims are as follows:
Table 1.0: Purposes and Objective of the Research.
To separate the current literature on the Information Content of Dividend Hypothesis. To look into the relationship between dividend alterations and the future net incomes of Corporate Firms.
To look into the relationship between the ICD hypothesis and Firm-specific features of Corporate Firms.
To reply the inquiry ‘Do dividend alterations predict Future Net incomes of Corporate Firms. ? ‘ utilizing informations from South Africa.
To look into if, so Firms that addition ( lessening ) their dividends see an addition ( lessening ) in their subsequent, future net incomes.
1.3. Structure Outline.
Chapter one introduces the reader to the context of the survey whilst offering the justification for the research worker ‘s desire to look into the ICD hypothesis. The purposes and aims of the survey are besides presented. The constructions of the survey will go on as follows:
Past literatures associating to the ICD will be reviewed within Chapter Two. The literary analysis to be conducted will assist solidify the purposes and aims to be achieved in the survey by leting the research worker to develop the research inquiries of the survey.
Chapter Three focal points on the presentation of the research methodological analysis adopted. An account of the building of the dataset to be employed is presented whilst presenting the variables cardinal to the empirical analysis.
Chapter Four will show the Empirical Findingss obtained from the information analysis of Chapter Three before a thorough analysis is conducted in footings of the research inquiries and literary grounds gathered.
The survey will officially reason in Chapter Five with a reappraisal of the work undertaken and suggestions for farther research will be made.
Chapter Two: Literature Reappraisal
2.1 Purposes and Objective
This chapter aims to research and critically analyze the relevant literatures sing the Information Content Dividend hypothesis and its relevancy to South African corporate houses.
The chapter will be dedicated to four chief issues ; definition of the chief footings of the survey, separate the current literature on the Information Content of Dividends hypothesis, place the extensions to be made to old methodological analysiss and research, develop cardinal research inquiries to be answered throughout the survey and conclude with a strong model to prove in the extroverted chapter.
2.2 Definition of Footings
2.2.1 Dividend Policy
Dividend policy is indispensable in the finding of how much of a houses ‘ net income is retained and how much is paid back to investors and stockholders. Harmonizing to ( Lease et al. , 2000:29 ) the term ‘dividend policy ‘ refers to “ the pattern that direction follows in doing dividend pay-out determinations or, in other words, the size and form of hard currency distributions over clip to stockholders ” . The existent payments of which are normally distributed capable to the handiness of hard currency or they are distributed as portion dividends. Thus determinations associating to dividend pay-out remain critical to directors, necessitating a careful consideration non merely of current but future deductions Hagin ( 1979: 63 ) as it goes a long manner in finding the going-concern capablenesss of a concern. An inordinate payment of dividends, may take to the keeping of fewer financess for investing intents, and this may take to the forced reversion to capital markets to derive financess. While a meager pay-out may signal hapless public presentation by investors. Therefore, dividend pay-out determinations must be reached in such a mode so as to equitably apportion the distributed net incomes and maintained net incomes. ( Baker and Powell, 1992 ) .
2.2.2 Dividend Changes
Dividend alterations signifies an upward/ downward motion in the sum of dividends paid -out from one period ( quarterly, semi-annually or yearly ) to the following. Past theoretical accounts such as that of Bhattacharya ( 1979 ) , John and Williams ( 1985 ) , and Miller and Rock ( 1985 ) asserts that alterations in dividend policy convey intelligence about future hard currency flows-in specific footings that, dividend additions convey good intelligence, and dividend lessenings convey bad intelligence ( Grullon et al, 2002 ) .A ground given for this is the being of asymmetric information, whereby investors do non hold free entree to insider information and as such, can merely trust on decoding relevant information from the actions of directors, in this instance they interpret a alteration in dividend as a dependable signal to foretell the future public presentation of houses. Thus dividend alterations can be defined as the awaited signals conveyed to external parties by directors for future net incomes or hard currency flows.
2.2.3 Net incomes
Net incomes refer to the sum of net income made by a Firm for an operational twelvemonth. A houses ‘ net incomes is of great relevancy to both internal and external stakeholders of houses has it gives valid arrows to the wellness, growing, effectivity and efficiency of a houses ‘ operations and its ‘ ability to prolong concern in the foreseeable hereafter. Past research workers such as Edwards and Bell ( 1961 ) , Ohlson ( 1991 ) , Bernard ( 1994 ) have successfully distinguished net incomes into normal and unnatural constituents. Abnormal net incomes ” are defined as the difference between entire net incomes and normal net incomes, where “ normal net incomes ” are defined as the needed return to proprietors based on the cost and degree of invested equity capital. Edwards and Bell ( 1961 ) .This differentiation is of great relevancy in trial of the ICD as it represents the proportion of the net incomes which may impel the signals or surprise by stockholders.
In old surveies the ability of net incomes prognosis to possess information content has been elaborated upon. Miller and Modigliani ( 1958 ) in their investing theory, posits the being of a relationship between a houses ‘ net incomes and its value, therefore they concluded it was necessary to see the capitalized value of the “ watercourse of net incomes ” over clip. In the same vena, the capital plus pricing theoretical account of developed by Sharpe ( 1964 ) and Lintner ( 1965 ) and the alteration of the Modigliani and Miller Model by Hamada ( 1969 ) , likewise proves that, under certain restrictive premises[ 1 ], that the value of a house was dependent merely on the chance distribution of the house ‘s future net incomes ( considered jointly with other houses ) and market factors which set the risk-return final payment construction.
2.2.4 Information Content of Dividend Hypothesis
The Informational Content of Dividends ( ICD ) states that dividend alterations portray information about the house ‘s public presentation in footings of future profitableness and hard currency flows as such they propel stock-returns. The ICD hypothesis which is the chief focal point of this survey, posits that with the happening of asymmetric information _ that is where directors possess information non readily available to investors, investors will construe a dividend addition as a signal that direction anticipates for good higher hard currency flows, and a dividend lessening as a signal that direction expects for good lower hard currency flows.
Ross ( 1977 ) amongst other account elaborates on the prevalence of dissymmetry information on issues associating to dividend policy utilizing a two dimensional attack which establishes a close positive relationship between unexpected dividend alterations, proclamations on stock returns and future net incomes. However, although many past literatures have given much grounds on the latter relationship between dividends and stock monetary values, concrete grounds on the former- relationship between dividend alterations and future net incomes which instigates the information content dividend theory is yet to be reached.
2.3 Brief Insights into the South African Market
The JSE was founded as far back as 1887 and in the 2011/2012 World Economic Forum ( WEF ) study it was ranked first out of 142 states for its ordinance of securities ( Press release, JSE,2011 ) , while in 2006 it was ranked 17th by the World Federation of Stock Exchanges, based on its market capitalization in US dollar footings. Having its major constituent for long-run financess typically as long-run bank loans and sedimentations, the bond market and the equity market. The South Africa ‘s capital markets have enjoyed betterments and growing over the last few old ages. ( Mboweni, 2006 ) .
Over the past 112 old ages, the mean return of the JSE has been 7.5 % per twelvemonth above rising prices. With analyst expecting the normal expected returns from the JSE to lift to13.5 % per twelvemonth if rising prices remains close to six per centum, . The FTSE/JSE All Share Index ( ALSI ) have compounded at 12 % ( sole of dividends and 15 % with dividends inclusion ) per annum at an Index degree. ( FPISA, 2012 ) . This returns makes the market an attractive for many investors sing the brighter hereafter it signals
Comparing the above consequences with their US opposite number, it can be deduced that although, the US markets and several developed markets greatly outperformed their South African opposite number in the 1990s, recent comparing of their public presentation over the same period with the S & A ; P 500 has been level in rand footings and has lost 2 % per annum in dollar footings at an Index degree ( excepting dividends ) , Shaun le Roux ( 2011 ) .This diminution harmonizing to analyst is as a consequence of the sky-rocketing monetary values experienced in the US market -laying much accent on the extreme importance of rating in the finding of future net incomes of Firms.
Figure 1: Graph demoing public presentation of the FTSE/JSE AllShare index vis a vis US S & A ; P500
Figure: Graph demoing the performace of the FTSE/AllShare Index vis a visthe US Standard and Poor 500.
*Source: PSG Asset Management, I-Net Bridge.
2.4 Back-ground of the Information Content of Dividends Hypothesis
Although it can be logically inferred following the dividend irrelevance proposition of Modigliani and Miller ( given market efficiency a house ‘s value should be independent of its dividend pay-out ) that since market-wide dividend pay-out ratios are low, higher reinvestment of net incomes should take to faster future aggregative growing and increased pay-out ratios should take to a diminution in maintained net incomes therefore lower investing in capital which culminates in slower sum growing and lower future net incomes. ( Arnot and Asness, 2001 ) .
However, in the existent universe, the being of imperfectnesss distorts this expected opposite relationship. As Dividends are frequently seen as a tool of fiscal signalling and investors normally conceive an addition in dividend by directors to be a possible /bright hereafter for the house while a diminution is normally conceived as a signal for hapless future public presentation. In add-on, bureau job could originate taking to incorrect investings by directors which negatively impact future growing. For case, dividends might signal directors ‘ private information about future net incomes chances, with low pay-out ratios bespeaking fright that the current net incomes may non be sustainable.
Therefore, the MM theory has been criticised for being unable to turn to the importance of dividend payments in concerns. As it has even been observed that houses sometimes go out of their manner to borrow in order to do dividend payments. Therefore leads to the inquiry frequently posed by critics that ‘if dividends are irrelevant as claimed by the theory, why do houses fight to keep dividend payments? ‘ ( Arnot and Asness, 2001 ) .
However, farther question into the work of Lintner, by Modigliani and Miller propelled a contrary determination that under the premises of a perfect capital markets, rational behavior, and nothing revenue enhancements, that the value of a house is independent of the house ‘s dividend pay-out rate this they coined the dividend irrelevancy theorem. This struggle is frequently termed as dividend mystifiers ( Lee, 2010 ) . As in an ideal universe ( with no market imperfectnesss ) , dividend pay-out should take to a decrease in maintained net incomes and a decrease in maintained net incomes should take to drive a decrease in investing which should hitherto, take to a decrease in future net incomes.
Recognizing the consequence of revenue enhancements, rising prices and dealing costs Miller and Modigliani ( 1963 ) revised their earlier theory ; they modify their anterior findings they therefore argue that a house ‘s market value depended on its expected future net incomes and non on current net incomes. Therefore, if net incomes consist of lasting and ephemeral constituents and if dividends depend on the former, dividends would function as a alternate for expected future earn- Immigration and Naturalization Services, and such a alternate relationship might explicate the consequences of the cross-sectional surveies.
Dividends Related Theories in Finance
2.5.1 The Agency Theory and the Free Cash-flow Hypothesis
The relationship that exists between directors and stockholders is widely conceived to be an Agent to Principal relationship. Harmonizing to the jurisprudence of bureau[ 2 ]the bureau is expected to keep a fiduciary/ trust relationship with the Principal.
However, due to struggle of involvement that sometimes arises this trust relationship is breached therefore taking to bureau jobs which biases directors ‘ determinations on the usage of the available scarce resources of the house, major of which is the free hard currency flow[ 3 ]. ( Jensen, 1986 ; Myers, 2001 ) . They opine that the struggles over pay-out policies are even more intense with the handiness of significant free hard currency flow in an Organization. As while Shareholders require frequent dividend pay-out as a wages for their investing, directors are frequently loath to make this as they view pay-outs to stockholders as a decrease of the resources under their control, thereby cut downing their country of power and control, which besides leads to subsequent demand to raise financess which hitherto creates increased monitoring from the capital markets. ( Rozeff, 1982, Easterbrook, 1984, Jensen 1986 )
Therefore Jensen ( 1986 ) argues that, in a command to avoid this director of houses ‘ with free hard currency flows in most instances prefer to overinvest by accepting fringy investing undertakings with negative net nowadays values. Sequel to this an addition in the dividend will, constantly, cut down the extent of overinvestment and increase the market value of the house, while a lessening in the dividends would bring forth a contrary consequence.
However profound the free hard currency flow hypothesis, it is free from unfavorable judgments, as some other empirical surveies such as Howe, He and Kao ( 1992 ) Lang and Litzenberger ( 1989 )
hold found little or no support for the extra hard currency flow hypothesis.
2.5.2 Bird in Hand Hypothesis
The Bird in manus is another hypothesis which sought to explicate the debated relationship between dividend payment and the value of a house. The Southern Cross of this logical thinking was that high dividends culminates in addition in portion values. It advocates that in an ideal universe with no imperfectnesss[ 4 ]rational investors are risk averse therefore they will prefer hard currency dividend payments to additions on the grasp of their investing ( capital additions ) , where the hard currency dividends can metaphorically be termed as the ‘bird in manus ‘ while the capital additions ‘two in the shrub ‘ . This is because higher current dividend minimizes the hazards attached to really obtaining future hard currency flows ( due to uncertainness ) hence cut downing their cost of capital and additions portion value.
2.5.3 Tax Theories
Taxes represent cost straight impacting dividend payment. Based on this association this hypothesis establishes a relationship between expected net incomes and a house ‘s pick of dividend pay-outs. It ‘s posits are that when dividends attracts a higher corporate revenue enhancement rate than retained net incomes, a house paying out more dividends will hold expected net incomes than houses which do non pay at all or every bit much. Therefore doing it necessary for directors to give careful consideration to the difference between revenue enhancement on dividends and capital additions every bit good as the difference between corporate revenue enhancement and retained net incomes.
2.6 Net incomes Measurement
To measure the ability of dividend alterations to foretell profitableness, it is imperative to gauge expected profitableness Nissim and Ziv ( 2001 ) .
Although enquiries are still been made as to high quality of the unnatural steps over net incomes step due to the latter ‘s ascertained restriction of its measuring being prone to mistakes. Measures utilizing unnatural net incomes have been widely adopted because as posited by Nissim and Ziv ( 2001 ) , for net incomes information conveyed by dividends to impact monetary value, the net incomes in inquiry had to be unnatural future net incomes instead than normal future net incomes.
Estimation carried out by most past surveies conjecture that net incomes follow a random walk with a impetus, therefore unnatural net incomes was deemed to be a map of existent alteration in net incomes minus the estimated impetus, accommodating this measuring of net incomes most of the anterior documents find that dividend alterations are non positively related to net incomes alterations. However a alteration by more recent surveies ( Nissim and Ziv, 2001 ) to turn to ascertained restrictions of issues associating to the appraisal of unexpected net incomes which are measurement mistake and omitted correlative variables led to a decision that that dividend alterations are positively associated with net incomes alterations in each of the two old ages following the dividend alteration. As posited by Nissim and Ziv ( 2001 ) , Earnings normal net incomes that result from future retained net incomes and future net stock issues have less impact on current monetary value.
Besides more recent surveies have proffered a apparently improved attack to capture earnings/profitability to extenuate against the restrictions built-in with the usage accounting ratios, several placeholders such as the entire market value deflated by capital outgo and economic net income which have been proven in research worker documents such as ( Harcourt ( 1965 ) , McGowan ( 1983 ) , Hawawini et Al, 2002:4 ) .For the intent of this survey the TMV/CE step will be employed in add-on to other gaining measuring adopted in past literature.
2.7 Prior Empirical Findingss
Since the overture of the Information Content dividend theory the topic has gained broad prominence in most Finance literatures on corporate pay-out policies and every bit served as one of the first accounts of ascertained relationships between dividends and stock monetary values ( Watts, 1973 ) .Having to its recognition much grounds from past, giving acceptance surveies to the being of a positive relationship between dividends and stock monetary values some of which are Petit ( 1972 ) , Aharony and Swary ( 1980 ) and Dielman and Oppenheimer ( 1984 ) .
Premier accounts of the being a prognostic relationship between alterations in dividends and future net incomes can be traced back to the polar work of Lintner ( 1956 ) .In his work he postulates _firms merely increase dividends when it is believed that net incomes would increase for good in the hereafter, this grounds he obtained from his survey of 28 US companies over a period of 23 old ages ( 1918-1941 ) where he paperss that with the exclusion of two companies dividend payment determinations were dependent on its positive desirableness as he observed a form of gluey dividends and dividend smoothening by directors. He therefore concludes that the elements of inactiveness and belief in direction that stockholders prefer a considerable stable rate and it is this stable rates that attract a premium. Past surveies such as Pettit ( 1972 ) , Aharony and Swary ( 1980 ) , Bhattacharya 1979 ; John and Williams 1985 ;
And Miller and Rock 1985 ) , buttress this grounds that proclamations of dividend alterations do convey information to the market.
Pettit ( 1972 ) , transporting proving for the ICD hypothesis in the US market find that the average risk-adjusted return for houses denoting dividend additions is significantly positive over the two yearss environing the proclamation, and for those denoting dividend decreases the two-day return is significantly negative. In the same vena Aharony and Swary ( 1980 ) study same consequences after commanding for contemporary quarterly net incomes studies.
In the same vena Nissim and Ziv ( 2001 ) mensurating utilizing both divisions of net incomes components- normal net incomes and unnatural earnings-they discovery that dividend additions are positively related to net incomes in each of the four subsequent old ages, but dividend lessenings do non related to future profits.Their consequences besides indicate stronger consequences for unnatural net incomes thereby proposing that accounting for the needed return more than offsets the mistake in mensurating unnatural net incomes. However this grounds has besides encountered small or no support from many rap researches some of which are Watts ( 1973 ) , Gonedes ( 1978 ) , Penman ( 1983 ) , DeAngelo, DeAngelo, and Skinner ( 1996 ) , Benartzi, Michaely, and Thaler ( 1997 ) , and Grullon, Michaely, and Swaminathan ( 2002 ) .
Grullon et Al ( 2002 ) , gives grounds bespeaking a steadfast rejection of the hard currency flow signalling hypothesis as it was found that an addition in dividend did non impel an addition in profitableness instead a lessening was observed. In the same visible radiation, net incomes from of dividend decreasing houses tends more to travel upward instead than diminution. However they find out that the wage -out of dividend increasing houses increase for good which indicates a support for the Linters theoretical account.
2.7. Anterior Evidence from Asia
Lonkani and Rachusanti ( 2007 ) proving for the ICD hypothesis in Thailand, argue that due to changing market conditions, past dividends entirely do non give equal accounts of the dividend surprises as proposed by the signalling theory, therefore they employ analyst forecasted dividends which they believe would in their surveies of houses listed on Thailand Stock Exchange. Their consequences show dividend surprises derived from analyst forecasted dividends are a better forecaster of house ‘s future net incomes that dividend surprises derived from conventional past dividends
Although many documents do non back up ‘the ICD hypothesis ‘ , and Choi et Al ( 2011 ) achieve consequences similar to Nissim and Ziv ( 2001 ) in their research carried out on Korean houses, they find strong support of the theory when houses were classified based on their firm-specific features corporate administration ( Chaebol and Non-Chaebol ) and growing degree ( High growing and low growing ) . They find that dividend alterations provide information about the degree of profitableness in subsequent old ages ; they besides reach the decision that dividend alterations are positively related to alterations in net incomes in each of the two old ages subsequent to the alteration.
2.7.3 Evidence from Australia
Lee ( 2010 ) trials for the ICD hypothesis utilizing the Johanssen Vector Error Correction theoretical account show a strong support for the Information/signalling content hypothesis of dividends in the Australian market in the long-term. As it was observed that a unit daze addition in dividend pay-out leads to a lasting addition in future net incomes over clip, and that the pay-out ratio Granger-causes net incomes.
2.7.4 Evidence from Africa
Seneque and Gourley ( 1983 ) one of the premiere subscribers to subject in South Africa, in their research on 145 JSE companies suggests the being of a strong presence of signalling effect/information content of dividend hypothesis. Their consequences besides gives acceptance to Lintner 1956 theoretical account as they conclude that Managerial determinations on dividends were greatly influenced by the chance of future net incomes.
From Nigeria, Uzoaga and Aloizieuwa 1974 in their trials carried out to look into the forms of dividend policies of 13 Nigerian Corporate Firms for a four twelvemonth period. Unlike their South African opposite numbers they found really small supports for the postulates of Lintner ( 1956 ) .Thus, they conclude that a alteration in the degree of dividends paid by the companies was best explained by sentimental factors such as frights and bitternesss instead than conventional factors put frontward in Linters ‘ ( 1956 ) theoretical account.
More recent survey by Adaramola ( 2012 ) on the Nigerian market, using the GLS ( Generalized Least Square ) analysis, gives acceptance to past grounds that dividends possess information content sufficiency to impact stock monetary values.
Summary Table Show of Evidence from Past Researches.
Consequence on Future Net incomes of Corporate Firms
Stock returns and dividend alterations
Free Cash Flow
Firm Specific Characteristics
Chapter Three: Datas and Summary Statisticss
This chapter consists of two chief parts. Part one matches the methodological analysis and hypotheses in visible radiation of the Information Content Dividend Theory. While the 2nd portion identifies and summarizes the informations applied in this survey.
3.1 Research Design
Following the posits of Benartzi et Al, ( 1997 ) , to efficaciously analyze the impact of alterations of dividends on future net incomes it is necessary to deduce the unexpected net incomes for the period. Unexpected net incomes is defined as the difference between the existent net incomes in the twelvemonth been observed and the net incomes forecasted utilizing relevant variables that may be an influencing factor of net incomes with the exclusion of dividend alterations.
Deductions of the Information content:
Firms which show an addition ( lessening ) in dividends in twelvemonth 0 will hold a positive ( negative ) unexpected net incomes in old ages 1,2, etcetera.
Of the houses that increase dividends the larger the dividend addition, the greater the unexpected net incomes in the undermentioned old ages.
3.2 Hypothesis and Methodology
This survey will use different methodological analysiss to measure whether the informations obtained from South Africa supports the Information Content of Dividend hypothesis.. In add-on the sample will be divided based on the size utilizing the book to market value of houses to measure whether house features has a distinguishable consequence on future net incomes of these houses. All methodological analysiss considered are run under a pooled time-series cross-section arrested development footing.
The survey seeks to:
To look into the relationship between dividend alterations and the future net incomes of Corporate Firms.
To look into the relationship between the ICD hypothesis and Firm-specific features of Corporate Firms.
To reply the inquiry ‘Does dividend alterations predict Future Net incomes of Corporate Firms? ‘ utilizing informations from Nigeria and South Africa.
In accomplishing the undermentioned propositions shall be tested utilizing informations from the selected sample
Hypothesis to be tested.
H1: There is an existing relationship between Dividends alterations ( Rdiv ) and the future net incomes of Corporate Firms ( Et-Et-1/B-1 ) .
H2: An addition ( lessening ) in dividends leads to an addition ( diminution ) in future net incomes of Corporate Firms
H3: There will be a positive relationship between firm-specific features and the ICD hypothesis.
3.3.1 Datas Sourcing
For the intent of proving the laid out hypothesis, this survey utilizes informations over a 14 twelvemonth period ( 1998-2011 ) .The sample period and sample size were selected based on the informations handiness standard of houses listed on the Johannesburg Stock Exchange, specifically the FTSE/JSE All Share index.
The informations analysed for the survey was obtained from the FTSE/JSE African Index Series[ 5 ]( JSEOver ) index on the Thomson Reuters DataStream database. However, the database did non supply for all observations the complete set of needed information. Consequently, those with uncomplete information had to be excluded from the original sample, go forthing a set of 911 firm-year observations.
The undermentioned standards were used for make up one’s minding whether to include a given house in the survey set:
The company was listed on the Johannesburg Stock Exchange ( JSE ) .
Datas had to be available either on the DataStream database or on the Thomson One database.
It had to hold paid dividends in two back-to-back old ages, therefore leting dividend alterations to be observed.
It should non hold exhibited unnatural dividend alterations ( over 500 per cent ) owing to the excess normally little sum of dividend payments in the old twelvemonth
Figure 2: The Capitalization of the Johannesburg stock market attributable to the All Share Index
Figure: The Capitalization of the Johannesburg Stock Market attributable to the JSE/All Share Index.
Beginning: JSE, data www.jse.co.za as at 30 September 2008
Graph demoing the capitalisation of the All portion index vis a vis the Full Market as 2008
3.3.2 Data Description
The information and variables used in this survey comprise of one-year accounting values and terminal of twelvemonth market capitalisation information-which includes- dividends pay -out, return of equity, return of assets, . The variable definition was adapted from Thomson Reuters Datastream and applied to prove the information content dividend theory.
Net incomes per portion
Dividends per portion ( DPS )
represents the net incomes of a company before involvement disbursal, income revenue enhancements and depreciation. It is calculated by taking the pretax income and adding back involvement disbursal on debt and depreciation, depletion and amortisation and deducting involvement capitalized.
stand for the net incomes for the 12 months ended the last financial twelvemonth.
represents the entire dividends per portion declared during the financial twelvemonth. It includes excess dividends declared during the twelvemonth.
stand for the financess used to get fixed assets other than those associated with acquisitions. It includes but is non restricted to: Additions to belongings, works and equipment Investings in machinery and equipment.
represents the amount of entire current assets, long term receivables, investing in unconsolidated subordinates, other investings, net belongings works and equipment and other assets.
Ra?†DIV_ this represents alteration in dividends.Di0 agencies dividend in twelvemonth 0 and D-1 means dividend the dividend in the twelvemonth before
Et denotes net incomes before extraordinary points to cipher the net incomes alterations
( E0 -E-1 ) /B-1 denotes net incomes in twelvemonth T deflated by the book value of terminal of the anterior twelvemonth ( at the beginning opf the twelvemonth T )
ROE EBEO/Book-value of equity maps as a control variable in the equations and its calculated by spliting net incomes before extraordinary points by the book value of equity for the same year..
ROEt-1 – this represents one twelvemonth lagged value of ROE
ROEt-2 this represents two twelvemonth lagged value of ROE
DNC this is a dummy variable takeing the value of 1 when alteration in dividends is negative and 1 otherwise.
DPC this is a dummy variable taking the value of 1 when alteration in dividends is positive and 1 otherwise.
DFE represents the value obtained from ROE-E [ ROE ]
E [ ROE ] is the fitted value from the cross sectional arrested development of ROE0 on the logarithm of entire assets in twelvemonth -1, the logarithm of the market-to-book-ratio pf equity in twelvemonth -1 and ROE-1
CE0 is calculated ( E0 -E-1 ) /B-1
NDFED is a dummy variable taking the value of 1 where DFE is negative and 0 otherwise.
PFED is a dummy variable taking the value of 1 where DFE is negative and 0 otherwise.
PCED is dummy variable taking the value of 1 where CE0 is negative and 0 otherwise.
NCED is a dummy variable taking the value of 1 where CE is negative and 0 otherwise.
The database is capable to survivorship prejudice since merely portions listed on the JSEAll Share Index for the period 1999-2011 are considered in the initial sample. The information set contains merely houses that survived during the clip period.
Using the attack employed by past documents such as Hodnett and Hsieh ( 2012 ) , the variables to be examined are winsorized by puting the upper limit and minimal one-year values of each property to 99.5th and 0.5th percentiles severally to take the utmost outliers in each month. In add-on, robust arrested developments are run for all equations to farther neutralize the effects of all outliers in the arrested development consequences.
All arrested developments consequences ( estimated coefficients, t-statistics, chances ) documented in this thesis were all obtained utilizing the STATA 11, package bundle. The dependant and independent variables used in the arrested developments covers the full sample size 1998-2011. However to efficaciously determine the existent relationship between future net incomes and dividend alteration, the arrested development consequences is summarized in two time-periods and in accomplishing this a basal twelvemonth of 1999 is chosen. Therefore for illustration, utilizing the information set 1998-2011, the dependant variable for the arrested development is derived from the net incomes from 1999-2011 for t=0, from 2000-2011 for t=1 and from 2001-2011 for t=2.
Throughout this survey a common base twelvemonth 0 is and -1 which connotes the twelvemonth preceeding the base twelvemonth is systematically used.
Hypothesis 1: To look into the relationship between dividend alterations and Future Net incomes.
Proposition 1: Does a alteration in dividend predict Future net incomes?
For the probe, the two chief variables employed in the analysis are the steadfast net incomes and its dividend alterations. Net incomes before extraordinary points was used as a suited placeholder to mensurate net incomes as harmonizing to Benartzi et Al ( 2004 ) the usage of this placeholder eliminates the ephemeral constituents of income. Annual data instead than quarterly dividends were used due to try specific features and it has besides been argued that dividends respond more to one-year net incomes instead than quarterly net incomes. ( Watts, ( 1973 ) )
Following Michaely et Al. ( 1995 ) and Pettit ( 1972 ) , dividend proclamations and skips are categorized into five categories on the footing of one-year alterations in the payments. These are: additions, decreases, no alteration, inductions and skips. For the intent of this survey three of these categorizations will be adopted_ dividend additions, dividend lessenings and no alteration in dividends.
Following past researches dividend alteration was scaled by the twelvemonth t_1 one-year dividend in a command to do the cross-sectional comparing more meaningful:
Where Di,0 represents the last one-year dividend in twelvemonth 0 for house ‘i ‘ and Di, -1, is the one-year dividend in twelvemonth -1.
Table 1 shows the one-year and entire figure of houses in the survey sample which meets the choice criteria.The entire figure of observations for dividend additions, dividend lessening and no alterations for the survey period are 457,447 and 7 severally. Comparing these figures with those obtained from past surveies such as DeAngelo and DeAngelo ( 1990 ) Nissim and Ziv,2001 ; Grullon et al. , 2005, Choi et al. , 2011 ) from other states, a similar tendency dividend addition outnumbering the dividend decreases can be observed. However this difference is rather minimum. Besides a particular tendency can be observed in the South African Market, a extremely important smaller figure of ‘no dividend alterations ‘ is seen although this conforms with the observation by Choi et Al ( 2011 ) , the difference of is manner below the figure noted by Choi et Al from the Korean market bespeaking more volatile dividend tendencies in the South African markets which suggestively negates Lintner ( 1956 ) proposition of gluey dividends due to inertia of directors.
Table1: Frequency of firm-year observations with at least one dividend event by financial twelvemonth
Year Dividend addition Dividend lessening No alteration Total
1999 17 25 0 42
2000 28 25 1 54
2001 33 29 0 62
2002 20 31 0 61
2003 32 26 1 59
2004 44 23 0 67
2005 24 50 1 75
2006 52 37 1 90
2007 43 46 3 92
2008 47 44 0 91
2009 42 50 0 92
2010 46 42 0 88
2011 37 44 0 80
474 472 7 953
Table studies the motion in the dividend pay-out by houses in the sample when compared to old old ages.
Panel A. Dividend addition
Panel B. Dividend Decrease
Panel C. Whole Sample
Initial ( Basic ) Analysis
( E0 -E-1 ) /B-1 =I±0+ I±1RDIV0+I±2ROEt-1+ Iµt
This analysis takes into awareness the conjectured random walk features of net incomes. Therefore the dependant variable future net incomes proxied by one-year value of net incomes before extraordinary points is adjusted to give unnatural net incomes ( E0 -E-1 ) .In a command to avoid the happening of measurement mistake in the dependant variable, the step of unexpected net incomes is deflated by the book value of the house ‘s equity[ 6 ]as initiated by past documents such as Penman ( 1996 ) , Nissim and Ziv ( 2001 ) and adopted by Choi et Al in their methodological analysis.
The analysis is so carried out utilizing the basic OLS additive arrested development
Cross-sectional Analysis developed by Fama and Macbeth ( 1973 )
( E0 -E-1 ) /B-1 =I±0+I±1 RDIV0+I±2ROEt-1+I±3 RDIV-1+I±4ROEt-2+ I±5 ( E0 -E-1 ) /B-1+vt
The cross-sectional analysis is a alteration of the initial basic analysis which gives a more believable account of the ICD, due to its dynamic nature. It gives room for the accommodation for autoregression in the clip series and the modeling of remainders as a white noise procedure, therefore following Choi et Al ( 2011 ) lagged variables are included to adequately work out the job ensuing from clip variance.Different from the old analysis the Fama and Macbeth ( 1973 ) analysis is undertaken to work out jobs originating from the cross correlativities of the remainders.
Cross-sectional Analysis for Asymmetric Dividend Changes.
( E0 -E-1 ) /B-1 = I±0+ +I±3ROEt-1+ 1
+I±4 ( E0 -E- 1 ) /B-1+t.
Having seen the difference between it the dividend addition and lessening group, a non-symmetrical relationship between dividend additions and lessenings could be hypothesized. Therefore following the methodological analysis of Choi et Al 2011, a separate analysis for the dividend addition and dividend lessening group is undertaken.In a command to gauge the consequence of addition and dividend closely, dummy variables[ 7 ]are introduced in add-on to the lagged variable of the dependant variable, to pick up the documented nonlinearities in the average reversion and autocorrelation of net incomes. ( See Fama and French,2000 ) .
Non-linear Model Analysis.
( E0-E-1 ) /B-1=
The nonlinear theoretical account is an alternate specification by Grullon et Al ( 2005 ) to set for any prejudice caused due to the non -uniformity of the average reversion procedure. Evidence from a major yesteryear surveies ( Brooks and Buckmaster ( 1976 ) and Elgers and Lo ( 1994 ) show that Contrary to the averment of Nissim and Ziv ( 2001 ) that net incomes follow a unvarying average reversion procedure, net incomes are really non-linear and a faster average reversion is observed for big alterations than for little alterations in the same vena it is observed to be faster for negative alterations than for positive alterations.
Panel Data Regression Analysis
Due to the combination of cross-sectional informations and time-series informations, OLS arrested development technique is unsuitable for the analysis ( Leamer, 1978 ) . The appropriate method of analysis involves panel informations arrested development techniques. There are two often used appraisal techniques for panel informations arrested development. These are the fixed effects theoretical account ( FEM ) and the random effects theoretical account ( REM
Chapter Four: Consequences description and reading
( E0 -E-1 ) /B-1 =I±0+ I±1RDIV0+I±2ROEt-1+ Iµt
T I±0 I±1 I±2 R2 N
Panel A. Whole sample
1 7.552615 ( 3.43 ) ** 0.10364 ( -10.09 ) ** 0.02436 ( -0.62 ) 0.2367 884
2 4.779967 ( 3.47 ) ** 0.10409 ( -10.3 ) ** 0.03242 ( -0.84 ) 0.258 834
3 5.222253 ( 3.71 ) ** 0.10486 ( -9.42 ) ** 0.05202 ( -1.27 ) 0.2591 778
This table studies estimations of arrested developments associating net incomes alterations to dividend alterations. The consequences of the arrested developments are divided into three chief parts, whole sample which shows the representative sample, while high growing and low growing shows the consequences when houses are classified based on their house features. High growing and low growing categorizations are made based on the book-to-market ratio ( B/M ) .Firms with a book to market value lower ( higher ) than 50 % are classified as high growing ( low growing ) .
Et represents net incomes before extraordinary points in twelvemonth T ( twelvemonth 0 is the event twelvemonth ) . B-1 is the book value of equity at the terminal of twelvemonth ) 1. RDIV0 is the one-year per centum alteration in the dividend pay-out in twelvemonth 0. ROEt is equal to the net incomes before extraordinary points in twelvemonth T scaled by the book value of equity at the terminal of twelvemonth t. The arrested development is carried out utilizing the STATA statistical package bundle. Following the methodological analysiss of Choi et Al 2011, Nissim and Ziv ( 2002 ) the Fama and Macbeth ( 1973 ) cross-sectional arrested development is utilized in the appraisal of the arrested development co-efficient. In the first phase annually transverse sectional arrested development coefficient estimations are obtained utilizing all observations in that year.In the 2nd phase, the time-series agencies of the cross-sectional arrested development coefficients are estimated.
Chapter Five: Decision
This survey analyses utilizing informations over a most recent period of 1998 to 2011 to measure whether the information content of dividends hypothesis holds for South African listed Companies. By analyzing informations obtained from Thomson Reuters Datastream and first utilizing a robust additive arrested development and thenceforth utilizing the Fama and Macbeth cross-sectional arrested developments to cut down jobs which may from the cross-correlation of remainders, the relationship between dividend alterations and future net incomes is examined both for the full sample and besides after categorizations based on the firm-specific features were made.The Fama and Macbeth cross sectional arrested developments are adjusted to take into consideration dividend dissymmetry and the non-linearity in net incomes. To guarantee hardiness of the consequences separate arrested development equations were run utilizing alternate net incomes measuring. In add-on to the return on plus ratio ( ROA ) which has frequently been used in past literatures on the topic of Information content of Dividends, the survey introduces the Entire market value deflated by Capital outgo as a step of net incomes to set for built-in restrictions of utilizing accounting ratios steps of net incomes.
This survey can merely be declarative and non conclusive due to a figure of
restrictions recognised in the design and executing of this survey: The sample size used for the survey although was grafted from an index on the South African Stock exchange be judged a just representation of the population of Corporate houses in South Africa, consequences may non be generalized due to the frequent job of uncomplete informations encountered, which was dealt with by the straight-out omission of observations this method past groundss show has the possibility of presenting prejudice to the consequences obtained, oppugning the overall dependability of Thomson Reuters Datastream accounting informations. Further, the survey uses merely an index on the market and non the whole population of companies listed on the Johannesburg Stock Exchange. Further, the survey does non measure other firm-specific features such as age and corporate administration
Although legion surveies have examined assorted issues of dividend policy, they have
produced assorted and inconclusive consequences. Possibly the celebrated statement of Fisher Black about divined
policy “ the harder we look at the dividends image, the more it seems like a mystifier, with pieces that
merely make non suit together ” ( Black, 1976, p. 5 ) is still valid.Additions
Recent informations set
Supply a sum-up of what you did in the paper
aˆ?Show what you have added to the literature
aˆ?If possible and relevant, supply a treatment of policy deductions