Content uploaded by Erekle Pirveli
Author content
All content in this area was uploaded by Erekle Pirveli on Mar 11, 2021
Content may be subject to copyright.
Content uploaded by Erekle Pirveli
Author content
All content in this area was uploaded by Erekle Pirveli on Jun 08, 2020
Content may be subject to copyright.
Content uploaded by Erekle Pirveli
Author content
All content in this area was uploaded by Erekle Pirveli on May 02, 2020
Content may be subject to copyright.
Earnings persistence and
predictability within the emerging
economy of Georgia
Erekle Pirveli
Caucasus School of Business, Caucasus University, Tbilisi, Georgia
Abstract
Purpose –The purpose of this paper is to provide the first empirical assessment of the persistence and
predictability of earnings within the Georgian private sector entities.
Design/methodology/approach –The sample comprises of all the Georgian private sector entities who,
according to the new Law of Georgia on Accounting, Reporting and Auditing (2016), had to submit their audited
financial statements by 1 October 2018. Financial data has been officially withdrawn from the Ministry of
Finance of Georgia and the descriptive data has been obtained by the use of Link Klipper and ScrapeStorm tools
through the official “Reportal”website. The final sample consists of 450 large Georgian private sector entities.
The study uses a simple, one-year-lagged earnings auto-regression to detect the persistence and predictability
within the next series of earnings. A weighted least square method has been used as a statistical procedure.
Findings –The results reveal that current earnings persist within the next year’s series of earnings at less
than 25%, while the reliance on current year’s earnings enables us to predict the next year’s earnings only
with a chance of 20%. Further analysis has witnessed that cash flows from operations persist at less than
40% and are able of predicting the next year’s cash flows at below 35%. Overall, the properties of earnings
and cash flows within the private sector of Georgia are of relatively poor quality, with the latter
demonstrating higher properties compared to earnings.
Practical implications –The general finding on a relatively low property of earnings raises potential
investors and creditors’awareness on the valuation-usefulness of provided financial information within the
private sector of Georgia. The fact that earnings are significantly less persistent and predictable compared to
cash flows from operations, hints on accruals’problematic functioning. The results presented in this paper
should be of interest to a local regulator (SARAS), charged with the responsibility of successfully running a
currently ongoing accounting reform of Georgia.
Originality/value –This is the first study that examines the persistenceand predictability of earnings and
cash flows from operations among the private sector entities ofGeorgia.
Keywords Earnings persistence, Earnings quality, Emerging economy, Accounting quality,
Georgia, Earnings predictability
Paper type Research paper
1. Introduction
Starting from the theoretical proposition of Graham and Dodd (1951) and followed by the
empirical evidence of Ball and Brown (1968), accounting literature’s one of the major
This work was supported by Shota Rustaveli National Science Foundation of Georgia [Grant
Number: FR17_489, Project Title: are Georgian Private Sector Entities Engaged in Financial
Information Manipulation?].
The author is thankful to the editor and two anonymous referees of the JFRA, to Mary E. Barth,
Tak-Jun Wong, Peter F. Pope and workshop participants at the American Accounting Association’s
annual meeting in San Francisco (2019) and First International Scientific-Practical Conference on Increased
Transparency and Financial Information Availability in Tbilisi (2019) for their valuable feedback.
Earnings
persistence
and
predictability
563
Received 19 March2019
Revised 1 January2020
1 April 2020
Accepted 2 April2020
Journal of Financial Reporting and
Accounting
Vol. 18 No. 3, 2020
pp. 563-589
© Emerald Publishing Limited
1985-2517
DOI 10.1108/JFRA-03-2019-0043
The current issue and full text archive of this journal is available on Emerald Insight at:
https://www.emerald.com/insight/1985-2517.htm
findings is to grant earnings information with the credibility of being involved in stock price
formation. As then capital market research in accounting has become a bulk of accounting
research. The scholars questioned not only the ability of earnings to predict stock returns
but also the ability of earningsto predict next year’s earnings and cash flows.
The neoclassical economic theory of consumption suggested that investors in their
decision-making process are based not on a single amount of earnings such as earnings in
year t, but rather on a series of earnings that may hint on potentially collectable lifetime
resources (Pimentel and Lima, 2010). That is, investor decisions, instead of a single (current)
earnings, are relied upon a longer-term income expectation, called as “permanent earnings”.
This stream of research became vital and has been implemented for many of the developed
economies. The evidence is, however, scarce from the emerging markets’perspective. This
study aims to provide the first assessment of the time-series properties of earnings for the
case of the Georgian private sector.
Several reasons make the case of Georgian private sector interesting for the general
public, regulators and standard setters. First, GA’s accounting and audit field currently
undergoes some unprecedented reforms. In the framework of the European Union (EU)-
Georgia association agreement, the Law of Georgia on Accounting, Reporting and Auditing
has been enacted as of 2016 [Law of Georgia on Accounting, Reporting and Auditing (2016)].
As such, for the first time in the history of Georgia, about 700 large and about 3.000 medium-
sized private sector entities had to recently become financially transparent, while about
80.000 small private sector entities have to go public by 1 October 2021. While the
unprecedentedly increased transparency is targeted to promote the local capital market
development, the latter is only possible if a quality –persistent and predictable –financial
information enters the playing field.
Second, GA’s accounting system’s convergence towards the unified international
reporting standards has been delayed in time as its fundamentals are said to be still rooted
to the Soviet-era accounting practices, where accounting served for monitoring/planning
and taxation purposes. To this end, the local accounting and audit profession still brings
relatively less experience of working with accruals, going concerned or fair value
measurement. This makes the properties of earnings in Georgia an interesting topic to be
investigated as over the years, earnings represented a source of monitoring (whether the
planned numbers have been achieved) as opposed of being valuation-oriented (McGee,2008,
2014).
The final feature that makes the topic novel is the local market’s limited demand for
financial information. Having one of the smallest and most illiquid capital markets
internationally vanishes the likelihood that capital markets in Georgia may put tension on
insiders to prepare value-relevant financial statements. As for the creditors, banks, instead
of the reliance on financial reports, normally base their debt covenant decisions on the
amount of collateral, future expectations and website visits (World Bank,C., 2008, 2013;
Alagardova and Manuilova, 2015;Pirveli, 2015). The demand for financial statements is
weak both from investors and creditors, whereas the focus shifts towards tax authorities,
who had access to entity transactions even without the recently declared increased
transparency. As there is a lower incentive for high-quality reporting among accountants,
the auditors may potentially save the game, however, prior literature has also evidenced that
many of the even larger entities avoid a more costly service of larger audit firms (Pirveli,
2019b). As such, the aim of the reform may shift towards an internal rise of funds, for which
the Georgian private sector may necessitate a well-diversified ownership structure.
According to Pirveli and Bendeliani (2020), local ownership structures are significantly
concentrated. So, to whom is the increased transparency, and thus, the published financial
JFRA
18,3
564
statements are directed to? If there are no actual consumers of this information, financial
numbers are likely to merely serve for taxation purposes. It makes earnings less likely to be
an outsider –and, thus, valuation-oriented.
In studying the properties of earnings within the Georgian private sector, this work
examines a non-yet investigated data of 450 large Georgian private sector entities. Financial
data has been officially withdrawn from the Ministry of Finance of Georgia and the
descriptive data has been obtained by the use of Link Klipper and ScrapeStorm tools
through the official “Reportal”website. The properties of accounting components are
detected in the light of their time-series properties of persistence and predictability. To
assess the quality of accounting components, this work uses a simple weighted least square
(WLS) regression of accounting components lagged on one-year lagged accounting
components. The sample covers financial and descriptive information of 450 entities of large
and medium-sized entities.
The results show that reported financial information is of less use for valuation purposes.
The results witness that earnings are lowly (below 25%) persistent and predictable.
Additional analysis shows that cash flows from operations, while they also exhibit generally
lower (compared to developed economies) properties, are significantly more persistent and
predictable compared to earnings. A significant difference between the properties of
earnings and cash flows from operations hints that corporate managers and accountants
face difficulties in a proper reporting of accruals-based transactions (deferrals, estimations,
depreciation, etc.). This, in line with the prior literature’s expectations, may hint on insiders’
lower orientation on the external use of financial information, the limited accounting/audit
profession and a turbulent business environment, diminishing investors’trust in providing
earnings-based estimations. The author recognizes that given analysis are based on limited
time-series information –as such each entity is represented by only two-years of financial
information. Despite the data is limited in time-series, the author highlights the importance
of detecting the very first evidence on the properties of accounting components at the outset
of the ongoing accounting/audit reform in Georgia, as it will help us detect the evolution of
the reform’s outcomes in the future research.
The findings of this work have important implications for the investors/creditors,
regulators and academia. External users of firm-level financial information may learn that
provided financial information can hardly help them in their valuation decisions. Among the
financial components, cash flows, however, could be more heavily envisaged in their
decision process compared to accruals or earnings. The finding on low-quality accruals is in
line to previous literature (Pirveli, 2015;Pirveli and Zimmermann, 2015) but new for the
private sector. As for the regulator, to promote local capital markets’development in
Georgia, the regulator should promote high-quality financial reporting. The regulator needs
to promote entities’reporting incentives to be transposed from taxation- towards valuation-
orientation. The reached finding is an indication that intensive accounting/auditing courses
are necessary to heighten the overall quality of the financial information and help insiders to
timely and properly follow the recently mandated accounting/audit rules. Simultaneous and
tireless efforts should be implemented by the regulators (Ministry of Finance –supervisor of
accounting/audit reform and the National Bank of Georgia –supervisor of the local capital
markets and the financial sector) to promote accounting/audit profession development,
entities’orientation on the external use of the reports and ultimate development of the
capital markets. Finally, this paper enriches accounting literature by providing the first
evidence on the properties of earnings from an emerging economy’s perspective. The work
demonstrates the relevance and topicality of the theme involving the complexity of data
retrieval and challenges of adapting regulation prepared in one type of context into a very
Earnings
persistence
and
predictability
565
different setting. By describing the accounting/audit reform development in Georgia, the
work gives a basis of how the transformation towards international accounting standards is
adopted, implemented and contextualized within an emerging economy. The contribution
and relevance of this work relate to the efficiency of international standards
contextualization in a context of an emerging economy and thus examining a transition
from a situation of disorder –after the independence from the Soviet Union towards a
developed economy with an orderly framework of transparency. This adds up to a debate of
a long literature on accounting standards adoption in emerging countries (Zaman Mir and
Shiraz Rahaman, 2005;Chand and White,2007a, 2007b;van Helden and Uddin, 2016).
In Section 2, by examination of the prior literature, setting up a scene behind the
accounting/audit reform development in Georgia and the development of the hypotheses,
provides a theoretical background of the work. Section 3 describes the sample selection
procedure and provides methodological details. Section 4 offers the results and finally,
Section 5 concludes the paper.
2. Theoretical background
2.1 International literature review
As defined by Thiagarajan (1989), persistence refers to the extent to which a certain value
continues to remain for a long time (longer than expected or continuously) in the future. This
general meaning was applied by Stigler (1963) while examining the departures from
perfectly competitive business environments. According to Stigler (1963), persistence
referred to the correlation of rates of return at two distinct points in time such as tand t1,
where a high correlation implied a high persistence. He noted that non-competitive
(monopolistic) environments are especially characterized with monotonic continuance
(persistence) of high rates of return:
Competitive industries will have a volatile pattern of rates of return, for the movement into high-
profit industries and out of low-profit industries will - together with the flow of new disturbances
of equilibrium - lead to a constantly changing hierarchy of rates of return. In the monopolistic
industries, on the other hand, the unusually profitable industries will be able to preserve their
preferential position for considerable periods of time (Stigler, 1963, p. 70).
Several reasons condition an outstanding emphasis of time-series properties such as
persistence and predictability of earnings research in the field of accounting and finance
(Kothari, 2001). First, its high importance stems from a desire to understand the role of
current earnings in the valuation process. This is because the properties of earnings are
either directly or indirectly linked to almost all valuation models (Feltham and Ohlson, 1995;
Ohlson, 1995). Second, accounting data-based capital market research has provided
consistent evidence that security returns are (at least partly) predictable and that the power
of predictability is a function of time-series properties of earnings. In studying the
association between accounting and market data, the relation preciseness is determined by
the degree of accuracy in disentangling an unexpected part of earnings from its expected
part [1]. The mentioned “degree of accuracy”of disentangling these two parts from one
another, is, on its side, a function of the time-series properties of earnings. This dimension of
research has contributed towards a long-standing debate around efficient market
hypothesis (Fama, 1970) and the behavioural finance (Shiller, 2003). Last but not least,
consistent with positive accounting theory (Watts and Zimmerman, 1978), knowledge of the
stochastic process of earnings generation helps managers to identify potential misuse
(smoothing) of accounting data. Detection of the “natural”path of earnings variation is
helpful to reveal the opportunistic part of earnings management (Foster, 1977).
JFRA
18,3
566
Graham and Dodd (1951) suggested that current earnings can be used as a starting point
to predict the next earnings. “In the absence of indications to the contrary, we accept the past
record as at least the starting basis for judging the future (Graham and Dodd, 1951,p.425)”.
This theoretical idea has been then reinforced and empirically tested by Ball and Brown
(1968) and Beaver (1968). Studying the relation between earnings announcements and stock
price movements consists of one of the major bulks of capital market research in accounting
literature. This stream of research has documented a positive and significant association
between the two variables, implying accounting information’s ability to contain value
relevant content [2]. A positive earnings news would, thus, generate higher expected cash
flows and lead to higher stock returns. This empirical finding was based on the fundamental
equality that stock prices stand for discounted cash flows. The reached conclusion has been
found to be robust “across statistical methodologies, time periods and stock exchanges”(Lev
and Ohlson, 1982,p.261).
As such, the approach of Graham and Dodd (1951) is today commonly echoed by
financial-statement analysis books, describing how to use the reported numbers for
valuation purposes (Healy and Palepu, 2012;Penman, 2013). In sum, the idea that financial
statements can be used for valuation if one is wise enough to make the necessary
adjustments have gained wide acceptance. From here on we can observe the next stream of
research “looking for a valuation short cut”to better understand the relation between the
current and permanent earnings (Frankel and Litov, 2009,p.188).
Graham et al. (1962) highlight the importance of information in current earnings and in
its components for estimating the upcoming earnings series of an enterprise. Consistent to
this, Sloan (1996) was another influencing work that has segregated earnings into its two
components such as accruals and cash flows from operations and supplied convincing
evidence that the information content of these components is systematically different for
stock pricing, but that market participants “do not reflect this information fully until it
impacts future earnings”.Sloan (1996) model has decomposed earnings in its two parts and
enables us to address:
whether the decomposition of earnings into its two parts such as accruals and cash
flows improves earnings persistence and predictability;
which component can better explain the next year’s earnings around its mean; and
whether the cash-flow component of earnings can better predict future earnings in
comparison to earnings itself.
His finding was in line with Bernstein (1993, p. 461, as cited by Sloan, 1996), who stated that:
CFO (cash flow from operations), as a measure of performance, is less subject to distortion than is
the net income figure. This is so because the accrual system, which produces the income number,
relies on accruals, deferrals, allocations and valuations, all of which involve higher degrees of
subjectivity than what enters the determination of CFO. That is why analysts prefer to relate CFO
to reported net income as a check on the quality of that income. Some analysts believe that the
higher the ratio of CFO to net income, the higher the quality of that income. Put another way, a
company with a high level of net income and low cash flow may be using income recognition or
expense accrual criteria that are suspect.
Based on the Center for Research in Security Prices file data on The New York Stock
Exchange and The American Stock Exchange firms and by using financial statement data
from 1962 to 1991 for 40,679 firm-year observations, in a pooled model Sloan (1996) detects
persistence of earnings at 84.1% (p. 299), the persistence of accruals at 76.5% (p. 300) and
persistence of cash flows at 85.5% (p. 300). Sloan (1996) argued that investors solely look at
Earnings
persistence
and
predictability
567
earnings and fail to distinguish between the accrual and cash flow components of earnings.
Since 1996, the investigation of the behaviour of earnings components came into force of
market efficiency research. This trend is called as “accruals anomaly”and the rationale
behind is that the financial part of earnings (cash flows) behave differently compared to the
non-financial part (accruals) of earnings in terms of persistence and, thus, the reactions they
cause into stock price volatility is also distinguishable (Galimberti and Cupertino, 2009).
According to Foster (1986,p.134)afirm’s accounting information is conditioned by its
production type, capital leverage, investments, general economic environment and the choice of
accounting techniques. Consistent to this, Lev (1983) showed that various economic factors are
linked with time-series properties of earnings. Based on a sample of 385 S and P firms’data
spanned across 15 consecutive years, Lev (1983) has witnessed various firm- and industry-
specificfactors’role in driving the persistency and predictability levels of earnings. Persistence as
such has been asserted to be a function of both firm- and industry-specific factors. The following
non-exhaustive list of factors mainly roots in industrial organization literature of economics:
Product type –the influence of product type is understood by Friedman’s
“permanent income”theory, stating that consumption of nondurable products and
services leads to more permanent income, while the consumption of durable
products is related to more volatile transitory income component. This implication
has been well evidenced by empirical studies (Darby, 1972). We see that series of
earnings within the companies producing non-durable goods and services behave
on average in a more systematic manner, demonstrated by lower variability and/or
by more significant autocorrelation in earnings series.
Firm size –positive accounting theory has suggested that larger firms are likely to
choose less risky investments to avoid potential government scrutiny that would
accompany “outstandingly”higher returns (Watts and Zimmerman, 1978).
Statistical studies have widely supported the argument that the variability of
growth rates of large firms is lower than for small firms.
Accounting techniques –Lev (1983) argued that the absorption costing method of
cost calculations is likely to yield a smoother earnings stream (with lower levels of
fluctuation) than the more economically rational method of direct costing would do.
Another important finding was revealed by Basu (1997). He showed that the release of
negative information is likely to be immediately realized while positive news will be realized
gradually over time; this let him to conclude that losses will be less persistent compared to
gains.
2.2 Accounting reform in Georgia
To formulate an expectation about the properties of earnings within the Georgian private
sector, we need to overview thelocal accounting field and local literature.
While it is a commonly discussable issue whether the adoption of international
accounting standards has been successful around the world, we are even less aware of how
the process of adoption is implemented and contextualized within the developing countries
(Amidu and Issahaku, 2019;Bananuka et al.,2019;Tawiah, 2019). Georgia is no exception.
As the independence from the Soviet Union, the accounting system in Georgia has been
rather unregulated. Over the years, the regulatory bases have been amended several times –
in 1995, 1999 and 2012. Though, the World Bank report on the observance of standards and
codes reports on accounting and auditing have revealed that these changes have been
unsystematic and significant deviations from the international standards have remained
JFRA
18,3
568
(Kaciashvili, 2003;Group, 2007;Wumburidze, 2013;McGee, 2014;Pirveli, 2014;Alagardova
and Manuilova, 2015). The World Bank, C. (2007) report stated there was a need of
considerable reforms in the field as follows:
an increased transparency and reach to entities’disclosures;
a clearer definition of public interest entities (PIEs)’status;
entities’categorization by size and the consequent allocation of reporting
requirements because of each category;
establishment of audit registry; and
higher attention and resources dedicated to professional training, as well as
materials’translation; and
stricter enforcement of the law.
Following the recommendations of the World Bank, GA enacted the Law of Georgia on
Accounting, Reportingand Auditing in 2016 (Law, 2016). Based on the size, revenues and an
average number of employees, private sector entities have been categorized into four classes,
plus the PIEs (Figure 1). PIEs and groups of Categories I and II had to publish their financial,
managerial (including non-financial reporting) and audit reports of the financial year 2017
immediately, but not later than 1 October 2018. Groups of the Category III had to report their
consolidated financial statements of the financial year 2018 immediately, but not later than 1
October 2019. Groups of the Category IV have to report their consolidated financial
statements of the financial year 2018 immediately, but not later than 1 October 2021 (Law,
2016). Financial statements of PIEs and Category I entities shall be prepared according to
international financial reporting standards (IFRS), while Categories II and III apply IFRS for
small and medium-sized enterprises and Category IV follows a simplified IFRS.
This implies about 80,000 entities have to become transparent by the end of 2021. Before
this massive data set breaks a transparency threshold, we are already now in need to have a
valid estimation of what to expect.
Figure 1.
Legal entity
categorization and
reporting
requirements
Earnings
persistence
and
predictability
569
2.3 Local literature review
In reviewing the existing local literature, we segregate the topics of an investigation into two
pillars, namely, quality of the amended law andenforcement level of theamended law.
As the new law of accounting and auditing has been amended in Georgia, we have seen
some preliminary works assessing the quality of the law in these settings. It has been
highlighted that the new accounting law of Georgia is aligned with most of the aspects [3]of
the international standards and is of higher quality compared to any of its predecessor
versions (Sabauri, 2018;Pirveli, 2019b). The new law has addressed previously existing
deficiencies highlighted by the World Bank reports (World Bank, C., 2007,2008,2013;
Alagardova and Manuilova, 2015). These changes covered:
an increased transparency and reach to entities’disclosures;
a clearer definition of Public Interest Entities’status;
entities’categorization by size and the consequent allocation of reporting
requirements due to each category;
establishment of audit registry;
higher attention and resources dedicated to professional training as well as
materials’translation; and
stricter enforcement of the law.
Pirveli and Shughliashvili (2019, p. 2) note that the currently ongoing regulatory changes,
including professional certification and continuing education standards, align with the EU
framework: “the processes are governed, managed, administered and financially supported
by foreign authoritative parties. This may already represent a crucial tool to achieve sundry
results”. Anticipating solid international support from the World Bank and the EU, as well
as Georgia’s political will to join the EU and an easiness (or cheapness) of “copy-pasting”
international standards, it is likely that Georgia may well harmonize its reporting standards
to the European experience. A more relevant question here is: how well these standards are:
adjusted to country-specific settings; and
brought to the reality (enforced) and what will be the final outcomes of the reforms?
The topic becomes even more interesting as we move towards the levels of enforcement
of the law. We have seen the literature assessing the enforcement levels of the accounting
law in Georgia [4]. This literature, however, has been based on rather a limited sample –
about 700 large entities of Categories I and II [5] plus the PIEs who had to submit their
financial statements already by 1 October of 2018. Pirveli and Shughliashvili (2019)
descriptively reveal that almost all the required entities (more than 90%) have submitted
their reports of 2017. Only 68 entities have been sanctioned based on the first year of
“going public”, from which 6 entities’financial statements have not been audited, 6 of
them have not fully published disclosure and 56 entities have not published the
statements at all (Kvintradze, 2019). Pirveli (2019a) argued that in some cases entities
have not submitted their financial statements on purpose. Bringing an example of a case
study, he argued that among these 56 entities, there are some state-owned and state-
subsidized company(ies) with assets above GEL 700m (e200m), working on losses. One of
such company preferred to two times pay a sanction of GEL 10,000 (e3,000) but not to
report. This happened because the assets re-evaluation within the company has past time
happened several decades ago and the management was a priori aware that the audit
verification would anyway fail.
JFRA
18,3
570
Overall, the enforcement level is high; considerably higher than an average picture
across the EU countries (Hope, 2003;Pope and McLeay, 2011;Hitz et al.,2012;Zaidi and
Huerta, 2014). Particularly so, if anticipating the fact that the reform has been recently
implemented. Pirveli (2019b), however, puts his attention on the timeliness of the
compliance. As such, the transparency has been increasing not only till 1 October 2018 but
also continued even after 1 March 2019 (Pirveli, 2019b):
The Ministry of Finance has multi-times cultivated that the enforcement levels in Georgia has
been above 90%, though the question of timeliness has been muted. Opposed to this, Pirveli
(2019b) highlighted that the reports delivery process was delayed in time; in some cases, either
warning or sanctioning from the ministry have been used and only afterwards the rule
compliance has been reached (Pirveli, 2019a).
Inline to this conclusion, just fivedays before the next deadline of transparency as of 1
October 2019 –the Minister of Finance of Georgia has officially postponed the deadline for
the fourth category entities (about 80.000 entities) until 1 October 2021 (Metskhvarishvili,
2019). The preliminary observations have revealed that only a very minuscule share of the
Category IV enterprises have had published their financial reports by this official
announcement –26 September 2019. As such, enforcement of the accounting law is high but
delayed in time.
2.3.1 Research gap. What has not been considered yet, is the quality of the provided
financial information. Without trusty and valuation-useful information, the reform would
not lead to capital market development or corporate government improvement.
Quality financial information embodies the principle that financial statements should be
as helpful as possible to investors and other capital providers in making their resource
allocation decisions (FAS Board, 2010). No matter how perfect law is or how well it is
enforced, if the provided information is not worthy to users, we are never allowed to call a
reform successful. This paper attempts to provide the first assessment of the time-series
properties of accounting components such as earnings and cash flows. Assessing
persistence and predictability of earnings is important to detect the dynamics of reform
outcomes in future research.
2.3.2 Hypothesis development. Several reasons make us to expect for rather lower time-
series properties of earnings and cash flows within the private sector of Georgia:
The world is striving towards a unified system of accountability already for a long
time. In this way of convergence towards a unified system, the developed world is
moving much faster, while emerging economies still struggle with some of the basic
issues. While the developed world discusses the need of advanced accounting tools
such as the use of disruptive technologies –blockchain and robotics –emerging as
game-changers [6], the developing world still experiences the challenges related to
fair value accounting (He et al., 2012;Farooq, 2018). These concerns are particularly
detrimental within the countries rooted back to Soviet times accounting practices,
where transition towards a free market has been delayed in time (McGee and
Preobragenskaya, 2006;McGee,2008, 2014;McGee and McGee, 2008).
As Georgia is a predecessor of the Soviet era, the accounting fundamentals are also
importantly rooted to that times reporting practices where accounting has been perceived as
a vital pillar of monitoring (Lenin, 1964, pp. 71-72): “accounting and control –that is the
main thing required for the “setting up”and correct functioning of the first phase of
communist society”. These all root back to the centrally-planned economy, where
accounting served the role of taxation and formal authorities. As the entities have been held
Earnings
persistence
and
predictability
571
by the authority, there have not been the shareholders who may use accounting data for
decisions (Ash and Strittmatter, 1992, p. 21). Similarly, as the assets were governed by the
authorities, there were no market (with sellers and buyers), and thus, no risk existed of the
appraisal or devaluation of an asset (Athukorala and Reid, 2003). Consequently, Soviet times
accounting used a historic cost method, while the revaluation of assets almost never
happened. To this end, it comes to no surprise that Georgian local accountants bring less
experience of working with accruals, going concerned or fair value measurement. As for the
auditors, prior literature shows that in the absence of outsider-oriented reporting, not many
of the companies are willing to pay for the costlier services of the big international audit
companies (Pirveli, 2019b).
Having the soviet experience of bookkeeping, GA represents a shining example where
the existing environment does not stimulate demand on high-quality financial reporting and
development of related professional services; term “audit”is often misinterpreted by the
society and trust to the quality of audit work is limited; there is a lack of competent
professionals within the field; functions assigned to the respective professional
organizations are not very well fulfiled; and self-regulation can hardly accomplish its
mission:
While the international standard-setters (Financial Accounting Standards Board
and International Accounting Standards Board) at the deepest level of their
conceptual framework put the focus on capital owners such as investors and
creditors, we need to never forget that Georgia’s capital markets are outstandingly
limited. For example, Georgian Stock Exchanges, based on market capitalization
and the volume of stocks traded, ranks among the four most illiquid and smallest
capital markets worldwide (among about 110 countries with available market
information within the World Bank indicators data) (Pirveli, 2015;Pirveli and
Zimmermann, 2015). Pirveli (2015) notes that stock market capitalization (relative to
Gross Domestic Product) of Georgia was higher than only in the Kyrgyz Republic,
Armenia and Uruguay.
The local capital markets are unlikely to influence the corporate manager’s incentives to
disclose investor-oriented information. Pirveli (2015) argued that “corporate managers [in
developing economies] do not input particularly high efforts in providing highly decision-
useful accounting information as the overall demand on accounting numbers is moderate”.
Moreover, Pirveli and Shughliashvili (2019) detect that only about 22% of the large
Georgian entities use the audit service of “big 4”audit firms. This may indicate that even
large entities within the developing economies do hesitate to use the costlier services of
bigger audit firms (DeAngelo, 1981;Francis, 2004;Gvaramia, 2014;Pirveli, 2015). In the
absence of capital market incentives, firms’reporting incentives transfer from investors
towards creditors or tax authorities. Creditors, because of having more expedient, private
channels for understanding firm vitality, in general, are assumed to formulate a more liberal
demand on financial statements. In Georgia, even though the submission of financial reports
is required within the loan application process, banks, instead of reliance on financial
statements, mostly base their debt covenant decisions on the amount of collateral, future
expectations and website visits (World Bank,C, 2008, 2013;Alagardova and Manuilova,
2015;Pirveli, 2015). Overall, the demand on financial statements is weak both from investors
and creditors. Thus, the major focus shifts towards tax authorities, who had access to entity
transactions even without a recently declared high and obligatory transparency. As such,
we are in need to focus on the possibility of rising the funds internally, and thus, need to
learn the ownership structures of the entities, which according to Pirveli and Shughliashvili
JFRA
18,3
572
(2019c) are highly concentrated. So, to whom is the increased transparency directed to? If
there are no actual consumers of this information, these unprecedented accounting/audit
reforms of Georgia may seem politically rational –bringing Georgia closer to EU
membership, but economically less efficient (Pirveli, 2019c)[
7].
Georgia (and its private sector), as a country with limited demand on financial
information, less experience of working with accruals accounting (subject to personal
judgements and/or estimations), lower use of international auditing practices and a
turbulent business environment, are less likely to experience external-user-oriented
reporting, where accrual component of earnings would be properly accounted and reported.
This leads to the following hypotheses:
Ha. Earnings are likely to be poorly persistent and predictable within the Georgian
private sector.
Hb. Due to accruals’low quality, earnings are likely to be less persistent and predictable
compared to the cash flows from operations within the Georgian private sector.
3. Research design
3.1 Data and sample
This research analyses non-yet examined data. Based on the data collection period –from
October 2018 to January 2019 –this work is based on the reported financial statements of
about 700 larger (Categories I and II, plus the PIEs) entities. The examined data is
automatically collected from two sources as follows:
(1) First, based on an official letter, the author has withdrawn the systematized financial
information of the entities through the Ministry of Finance of Georgia. Financial
information covers balance sheet, income statement, cash flow statement and the
statement of changes in equity. This is systematic information, available in excel
format. The information is public in general, but can only be requested through an
official letter written to the ministry, explaining the purpose of the usage.
(2) Second, the author has collected the descriptive data of the entities from an open
governmental public source of https://reportal.ge/ by using the “link clicker”and
“Scrapestorm”techniques. This work has elaborated a time-efficient approach to
obtain this information. As long as the open government portal requires authenti-
cation, at first, a “Link Klipper”has been used to collect the website addresses sep-
arately for each entity. As next, the collected web addresses have been pooled
within the AI scraper (“ScrapeStorm”). This approach enabled the author to timely
collect all the necessary descriptive information such as entity category, legal sta-
tus, sphere of operation, year of registration, audit status, audit firm name and
more (details of the data withdrawal could be found in Pirveli, 2019c).
Before proceeding to the methodology and data analysis, we need to question the validity of
the obtained data and whether an authentic conclusion could be made based on this data.
How qualitative and reliable could the data be within a developing economy’s private sector,
particularly if this data is reported for the first time ever? The difficulty of measuring the
size of an enterprise (measured by total assets) is important. The evidence shows that asset
measurement is accompanied by technical flaws because a significant portion of enterprises
has technically incorrectly entered its financial information scale online –whether the
information is given in Georgian Lari (GEL) or in GEL 1,000.
Earnings
persistence
and
predictability
573
Entities are required to provide their four financial statements online by filling up the
specific forms. The verification of this information, however, does not take a place. That is, it
is a requirement from the supervisory body, but neither there is a sanction of providing
technically incorrect information, nor this information is a subject of audit verification.
Accordingly, entities seem to put less efforts to be as precise as possible in their reporting
processes.
The evidence shows that technically defective reports have not been detected and
corrected by the supervision service and have been published publicly [8]. It is likely that the
regulator did not pay much attention (time) to this factor. For example, a look at one of the
company’sfinancial statements reveals that the company’s total assets [9] amount to GEL
4,482m. According to the information provided, the company states that its data is given in
GEL 1,000 (not in GEL), which makes the company to a position with assets of GEL 4.482bn
and falls into the top 10 largest entities. To showcase how prevalent are the technical errors
in scale, this work has checked the firm sizes grouped by their Categories (I-IV and PIEs).
According to the results, no tendency was detected that Category I entities are larger than
Category II entities; that the latter are larger than Category III entities; and the latter are
larger than Category IV entities. This evidence strengthened the argument that the effect of
technical errors has been significant, limiting us to draw fair descriptive analysis (Pirveli,
2020).
To mitigate this deficiency, one possibility would be to drop the entities, which have
violated the category definition thresholds (Figure 1) by total assets. This approach would,
however, significantly decrease the given sample and neither it would be completely precise
for three reasons, namely, first, dissimilar to the Categories I-IV entities, there is no total
assets thresholds defined by the law for the PIEs; second, the probability of technical errors
would still remain for Categories I and IV entities as they are bordered only from one side
(either from above or from below); and third, matching the entities’asset sizes with law
thresholds does not uniquely identify our sample as total assets are just one of the three
factors (asset size, revenues and an average number of employees) affecting on the
categorization, whereas, according to the law, satisfying two of the three factors is good
enough to be allocated in a certain category (i.e. a company might be belonging say to
Category I not only because of assets size but also because of revenues and the average
number of employees). As such, this approach would have lowered our number of
observations, while we would still face Types I and II errors (i.e. some of the entities would
be dropped with no reason and some others would remain in the sample but should be
dropped).
To overcome the scaling problem, this paper went hand by hand across the all “doubtful”
entities’pdf financial reports to check the preciseness in scale. The “doubtful”entities have
been those ones with total assets above GEL 1bn and still indicating GEL 1,000 as a scale of
reporting or those with total assets below GEL 1,00,000 and still indicating GEL 1 as a scale
of reporting. The author has handed by hand-corrected the scale of reporting for 60 entities,
constituting about 15.6% of the final sample observations. Having again checked the
category tendency based on total assets, the picture has changed as we now got the picture
as expected: Category I entities are larger than Category II entities; the latter is larger than
Category III entities; and the latter is larger than Category IV entities.
As the technical problem of scaling is mitigated, we can now move towards the final
sample selection process. Table 1 shows the sample selection process. In total, from the
initial sample of 768 entities derived after merging the above-mentioned two sources of
information, after filtrations, the final sample consists of 450 entities. The author has
removed 39 enterprises due to missing the identification code and/or fiscal year information;
JFRA
18,3
574
135 entities with no “full”availability of necessary financial variables (assets, liability,
equity, revenue, net income and cash flows) have additionally been removed; due to the lack
of information for the 2017 fiscal year, 26 entities have been dropped; 69 entities with no
sphere and category details have been also dropped; 8 entities from the agricultural sector
have been also filtered to allocate our observations in 4 large sectors (manufacturing, retail,
service and finance); finally, to maintain the sample homogenously, the author has dropped
41 entities of Categories III and IV who have voluntarily (where audit verification is not
required) submitted their statements. This enabled the work to maintain 450 large entities of
PIEs and Categories I and II.
3.2 Methodology
The potential list of applicable methodologies for the detection of properties of accounting
information needs to be chosen with caution in general but for the emerging economies with
hardly available and less reliable data sets in particular (Pirveli, 2015). Time-series
properties such as persistence and predictability are often elaborated to detect time-series
properties of accounting information. Under the time-series properties of earnings, two
concepts –persistence and predictability –are meant. Earnings persistence attributes to the
extent at which current earnings are able to remain in the next round of earnings series.
Earnings predictability detects the extent at which current earnings are able to estimate
next year’s earnings (Lipe, 1990). Highly sustainable and well-estimated earnings serve as
the basis for information retrieval and promote accurate equity valuation; thus, both
attributes are positively linked with the quality of accounting information (Nissim and
Penman, 2001;Dichev and Tang, 2008;Dechow et al.,2010)[
10].
Based on the limited data available and, additionally, anticipating the accruals less
reliable working process in Georgia, this paper attempts to indirectly measure accruals.
That is, the paper evaluates time-series properties of earnings and cash flows from
operations and attempts to reveal the quality of accruals as a variation between the results
for earnings and cash flows from operations (Sloan, 1996). To be able to observe the quality
of accruals directly, the work would necessitate at least three-years of “full”information
(two years data for the lagging purposes and the third-year data for calculating the changes
in accruals components such as changes in current assets, changes in current liabilities,
changes in cash and changes in short-term-debt) for each entity.
According to Kothari (2001, p. 149), a simple model of Foster (1977) is of as good use as
more sophisticated Box-Jenkins autoregressive integrated moving average models. If a
Table 1.
Sample selection
Sample selection No. of observations
Initial Sample 768
Filtrations:
Entities without identification code and/or fiscal year 39
Entities without financial information 135
Entities without the 2017 year reports 26
Entities without sphere and category information 69
Entities from the agricultural sector 8
Entities from Categories III and IV 41
Final Sample 450
Notes: This table illustrates the sample selection procedure. The final sample consists of 450 entities
Source: Author’s own
Earnings
persistence
and
predictability
575
typical univariate regressive model would stand for the causality between two different
(independent and dependent) variables, a univariate autoregressive model would determine
the causality between the two different values of the same variable (X) in different times (t
and t1). In a model, autoregression is the tendency for observations made at lagged time
points to be related to each other. As such, past values of the variable should decreasingly
influence current values as the power of correlation decreases hand by hand with an
increase in a time lag. Consequently, this work tests time-series properties of earnings
following a basic regression between the current and lagged earnings as suggested by
Foster (1977) [Freeman et al. (1982); and Lev, (1983)][
11]:
NIt¼
g
0þ
g
1NIt1þ
«
t(1a)
where:
NI
t
= current year’s net income (scaled by the tyear’s total assets) [12]; and
NI
t1
= previous year’s net income (scaled by the t1year’s total assets).
Regressing current earnings on the previous year’s earnings enables us to know at what
extent the current earnings could be explained by the previous year’s earnings. Earnings
persistence can be revealed by observing the coefficient of scaled earnings (
g
1
) within the
autoregressive Model (1a).
g
1
, that is a mean-reverting and varies between zero and one,
speaks for earnings persistence, whereby, the variance of the residuals indicates on the
power of predictability (Beaver, 1970;Freeman et al., 1982). High values indicate highly
persistent earnings; thus, past (current) earnings’ability to accurately determine current
(future) earnings. To test predictability, researchers focus on the variance of the residuals of
a model. Highly volatile earnings show a high absolute value of the stochastic term. In this
case, current earnings can scarcely proxy for subsequent earnings. The variance of the
residuals is an inverse function of accounting quality. That is, the higher the variance of
residuals, the lower is the predictability –indicating poor accounting quality.
This work extends the basic model by firm-level characteristics such as size (logarithm
of total assets), return on assets (ROA) and financial leverage (liabilities over total assets)
and growth rate (change in net income under net income). As the given data set is not a real
time-series, rather a cross-sectional (the author basically has one observation for each
entity), he extends the basic model by category and industry (sphere of operation) fixed
effects (Xie, 2001;Dechow et al., 2010)[13]:
NIt¼
g
0þ
g
1NIt1þ
g
2LEVt1þ
g
3SIZEt1þ
«
t(1b)
LEV
t1
=financial leverage (total liabilities under the book value of equity) in year t1; and
SIZE
t1
= total assets in year t1.
Following Sloan (1996) and the hypothesis of this work, we aim to detect the quality of
accruals by separate assessment of earnings and cash flows from operations. While we are
limited in running solid time-series models, we are able to go into detailed analysis of
separate accounting components, namely, earnings and cash flows. Hand by hand with the
earnings analysis, we run similar tests for cash flows from operations:
CFOt¼
g
0þ
g
1CFOt1þ
«
t(1c)
CFOt¼
g
0þ
g
1CFOt1þ
g
2ROAt1þ
g
3LEVt1þ
g
4SIZEt1þ
«
t(1d)
where:
JFRA
18,3
576
CFO
t
= current year’s operating cash-flows (scaled by the beginning year’s total assets),
taken directly from the cash flow statement;
CFO
t1
= previous year’s operating cash-flows (scaled by the beginning year’s total
assets), taken directly from the cash flow statement;
ROA
t1
= return on assets (t1year’s net income under t1 year’s total assets).
Prior literature normally uses the ordinary least squares (OLS) method while running the
above regressions. In some cases, however, other statistical tools are more preferential. Xie
(2001), for example, uses a generalized least squares model (Aitken, 1935) instead of the OLS
method. Some of the authors (F. Dormann et al.,2007) argue that the confidence intervals for
the OLS regressions have (almost) the same centres as derived by various spatial methods,
including the WLS. According to them, different p-values (and therefore t-statistics) are
derived due to OLS and WLS, the difference between the two should be relatively minor, as
the OLS estimator is an unbiased estimator for WLS. In those cases where dependent
variables are not normally distributed and the model is a subject of concern about
heteroscedasticity, a WLS method is preferred (Asparouhov and Muthén, 2010). To make a
rational choice on the method of regression, the paper needs to run some tests on the
normality distribution of the variables andheteroscedasticity concerns of the model.
4. Analysis of results
4.1 Descriptive analysis
This paper provides descriptive analysis because of two major criteria, namely, entity
categorization and sphere of operation.
Figure 2 shows the distribution of 450 entities by their categories as defined by the law of
2016. The entities are distributed in three categories, namely, Categories I and II and the
PIEs. As we can see, each category has more than 70 observations. This is an important
prerequisite for making statistically reliable conclusions based on category analysis. The
most widely represented category is Category II, 279 observations, 61% of the sample.
There are 98 PIEs, constituting 23% of the sample. In total, 73 entities are from the Category
I, representing 16% of the sample.
Figure 3 shows the distribution of 450 entities by their sectors of operation. The entities
are distributed in four sectors, namely, manufacturing, retail, service and finance. As we can
see, each field has more than 90 observations. This is an important prerequisite for making
statistically reliable conclusions because of sectors. The most widely represented field is
service industry with 33% of the entities; the service industry is followed by the retail sector
(27%), finance (20%) and manufacturing (20%) sectors. Figure 2 shows that the number of
entities operating within the manufacturing sector is the smallest compared to other sectors.
Figure 2.
Distribution of the
number of entities by
categories
Earnings
persistence
and
predictability
577
Table 2 shows descriptive statistics because of entity category. Descriptive statistics
cover ROA, size (total assets in mln GEL) and financial leverage (liabilities under total
assets). We see that PIEs on average show negative ROA, while Categories I and II entities
are profitable. PIEs are on average about two times larger than Category I entities, while the
latter is on average about four times larger than the Category II entities. Financial leverages
are about equal across all categories, standing at around 60-65% of total assets. We need to
mention that an overwhelming majority of the PIEs are the financial sector players (banks,
microfinance organizations, insurance sector), who due to the recent banking sector
regulations implemented from the National Bank of Georgia had a more or less tough
financial year in 2018. More to this, for the financial sector players because of their structure
of capital and balance sheet, a more relevant measure would be return of equity (instead of
ROA).
Table 3 shows descriptive statistics because of entities’sphere of operation, namely,
manufacturing, retail, service and financial. Return is negative for the financial sector that,
again, should be measured by return on equity or return on operational assets. We see a
significant advantage of financial sector entities compared to other sectors based on asset
size. Financial sector players on average are at least four times larger than any other sector
representatives. It comes to no surprise that financial leverage is highest within the financial
sector, followed by the retail sector.
Figure 3.
Distribution of the
number of entities by
sector
Table 2.
Descriptive statistics
by category
Category ROA (%) Size (mln GEL) Lev (%)
PIEs 2.7 326.2 65.2
Category I 2.4 179.0 65.5
Category II 9.1 47.6 59.8
Table 3.
Descriptive statistics
by sectors
Sphere ROA (%) Size (mln GEL) Lev (%)
Manufacturing 3.6 72.9 59.4
Retail 9.2 56.9 69.5
Service 6.1 93.8 49.9
Financial 1.3 337.9 73.4
JFRA
18,3
578
4.2 Correlation analysis
Table 4 provides the results of Pearson correlation matrix. 2018-year earnings are
significantly correlated with 2017-year earnings (corr = 0.39, p-value = 0) and 2018-year
CFOs (corr = 0.181, p-value = 0), while 2017-year CFOs cannot explain 2018-year earnings at
all (corr = 0.036, p-value = 0.444). Financial leverage of the prior year is negatively linked
to current earnings, while size of an entity has no significant association to current earnings.
2018-year CFOs are by 54% linked to 2017-year earnings; firm-specific factors such financial
leverage and size of the prior year are negatively (but significantly) linked to current year’s
cash flows, while return is significantly positively linked to it.
4.3 Graphical analysis of earnings and cash flows
Figure 4 provides a graphical illustration on the association of current and past year
earnings an dcash flows. Figure 4 part a) illustrates the association of 2017 and 2018 year
earnings and Figure 4 part b) illustrates the association of 2017 and 2018 year cash flows
from operations. The variables are scaled by total assets and winsorized at 1 and 99
percentiles. Both graphs leave the impression that there is an association between 2017- and
Table 4.
Pearson correlation
Variable NI (2018) NI (2017) CFO (2018) CFO (2017) LEV (2017)
NI (2017) 0.392
0.000 ***
CFO (2018) 0.181 *** 0.122 ***
0.000 0.009
CFO (2017) 0.036 0.030 0.541 ***
0.444 0.522 0.000
LEV (2017) 0.199 *** 0.319 *** 0.081 * 0.038
0.000 0.000 0.086 0.425
SIZE (2017) 0.036 0.051 0.141*** 0.195 *** 0.060
0.442 0.285 0.003 0.000 0.207
Notes: p-values are given in italics below the correlation values. ***, ** and *stand for significances at 1, 5
and 10% significance levels (respectively) using two-tailed tests
Figure 4.
Scatter plot diagram
for earnings and cash
flows from operations
Earnings
persistence
and
predictability
579
2018-year data and that the association is positive. The strength of the association, however,
is likely to be rather weak as the observations are quite widely distributed, indicating on a
high variance. A closer examination may leave us with a suspicion that cash flows seem to
be more concentrated compared to earnings. The following regression analysis should shed
the light.
4.4 Regression analysis
Kernel density distribution can be used to make a visual detection on the distribution of the
used variables. Figure 5 shows Kernel density distribiton of earnings hand by hand with a
normal distribution line. The graph detects net income’s distribution is visibly deviated
from its normal distribution (bandwidth = 0.031). A similar deviation is observable for the
cash flows from operations.
Having additionally checked the normality of dependent and independent variables’
distributions, Shapiro-Wilk test showed that the assumption on variables’normal
distribution can be rejected at 1% significance level. As next, the work has checked a
Cameron-Trivedi test on heteroskedasticity. The results showcase that the Models (1a-1d)
encounter a heteroskedasticity problem in the context of the Georgian private sector.
Consequently, due to non-normality of the data distribution and an issue of
heteroscedasticity, the work elaborates a WLS regression instead of an OLS.
WLS-results on time-series properties of earnings are reported in Table 5. The table
reports the properties of earnings based on four models. The first model is a simple WLS
model where current earnings are regressed on previous year’s earnings. In the second
model, the same model is extended by firm specific variables such as: financial leverage, size
and growth rate. Each of the two models are weighted by category and sphere of the entities
separately.
Lagging represents the main restriction of the sample. The number of observations per
regression is 450. While the number of observations is smaller than in similar studies
conducted for developed capital markets, this data is rather luxurious in the context of
Georgia. With no surprise, there are significant effects of lagged earnings on current
earnings –the coefficients of the lagged earnings are significant at 1% in each model. The
persistence levels of earnings, however, vary between 0.21 and 0.24, indicating on a rather
lower persistence of earnings. The R
2
vary between 16-17%, speaking about relatively lower
predictability of earnings. The result indicates financial information users’limited ability of
Figure 5.
Kernel density
distribution of
earnings in 2018
JFRA
18,3
580
predicting the earnings of 2017 based on 2016-year information. In general, firm-specific
extended models do not augment the properties of earnings. While weighted by the sphere
of operation, we see that larger firms are characterized with slightly lower persistence of
earnings compared to their counterparts. We also observe, that the effect of “weighting”in
WLS model does not differentiate among weighting by “sphere”and “category”factors.
International studies (Sloan, 1996;Dechow and Ge, 2006;Richardson et al.,2006;Dichev and
Tang, 2008;Frankel and Litov, 2009) conducted for developed capital markets detect the same
indicator of persistence (
g
1
) varying between 60-85%. The accounting components of
2016 played a significant role in the formation of the 2017-year components within the
Georgian Categories I and II private sector entities, plus the PIEs. This indicator,
cannot be directly compared to the international findings as hereby we only consolidate
the data of two-years financial information for each entity, enabling us of having a
single regression observation for each entity. Having this statistical limitation in mind,
though, we are able to have the comparisons between earnings and cash flows from
operations, which follows as next.
Table 6 reports the properties of cash flows from operations based again on four
models. The analysis of persistence and predictability of cash flowsfromoperations
should shed more light on the reasons behind low the above-detected poor properties of
earnings. The first model is a simple WLS model where current cash flows are
regressed on previous year’scashflows. In the second model, the same model is
extended by the firm specific variables such as: ROA, financial leverage and size. Both
of these models are weighted by “category”and “sphere”of the entities, in sum totalling
four models. The number of observations per regression stands again at 450. There are
significant effects of lagged cash flows on current cash flows in each model (F>7, p<
0.01). The coefficients of one-year lagged cash flows are significant at 1% in each
model. The persistence levels of cash flows stand at 0.31-0.36. The R
2
of all four models
vary between 29% and 32%, speaking about relatively higher predictability of cash
flows compared to earnings. The firm-specific variables do not have significant
Table 5.
WLS results on time-
series properties of
earnings
Variables
NI
t
(1a)
NI
t
(1b)
NI
t
(1c)
NI
t
(1d)
NI
t1
0.236 *** 0.211 *** 0.231 *** 0.218 ***
3.91 3.15 3.95 3.47
LEV
t
0.051 0.040 *
1.45 1.36
SIZE
t
0.018 * 0.001
1.94 0.18
Weighting factor Category Category Sphere Sphere
N450 450 450 450
F15.29 *** 12.23 *** 15.59 *** 8.12 ***
R
2
(%) 15.50 17.01 17.27 17.75
Notes: This table reports coefficients, t-values (of two-tailed tests, in italic below) and Adj. R
2
of the
following pooled weighted least square (WLS) regressions with Huber-White robust standard errors:
NIt¼
g
0þ
g
1NIt1þ
«
t(1a) WLS weighted by category NIt¼
g
0þ
g
1NIt1þ
g
2LEVt1þ
g
3SIZEt1þ
«
t(1b) WLS weighted by category NIt¼
g
0þ
g
1NIt1þ
«
t(1c) WLS weighted by sphere
NIt¼
g
0þ
g
1NIt1þ
g
2LEVt1þ
g
3SIZEt1þ
«
t(1d) WLS weighted by sphere Intercepts remain
unreported for the sake of brevity. Lagging represents a main restriction for the sample size. The variables
are winsorized at 1 and 99 percentiles. ***, ** and * stand for significances at 1, 5 and 10% significance
levels (respectively) using two-tailed tests
Earnings
persistence
and
predictability
581
influence on the cash flows’properties; consequently, their addition does not improve
much of the models’fitness.
In general, we detect low persistence and predictability for both, earnings and cash flows
from operations. This finding is in line with a general expectation of low accounting quality
in Georgia. Comparing the results of Tables 5 and 6, we observe higher persistence and
predictability of cash flows compared to earnings. Persistence of earnings across four
models vary between 0.21 and 0.24 for earnings and between 0.31 and 0.36 for the cash
flows. Predictability of earnings across four models vary between 16% and 18% for
earnings and between 29% and 32% for cash flows. As earnings is the accounting
component that is a sum of cash flows from operations and accruals, the work attributes its
lower quality to the low quality of accruals. As such, we cannot reject the first (Ha)
hypothesis that within the Georgian private sector, earnings are lowly persistent and
predictable. Neither we can reject the second hypothesis (Hb) that earnings are less
persistent and predictable compared to cash flows from operations, that is likely to be due to
accruals law quality. The findings are in line to prior literature’s estimation about financial
information’s low orientation on its external use, limited accounting/audit profession and a
turbulent business environment, reducing investors’ability to provide efficient earnings-
based estimations.
5. Conclusion
This work provides the first assessment of persistence and predictability, based on a non-
yet examined data of 450 large Georgian private sector entities. Financial data has been
officially withdrawn from the Ministry of Finance of Georgia, while the descriptive data has
been obtained by the use of Link Klipper and ScrapeStorm techniques through the official
“Reportal”website. The result suggests that earnings are poorly persistent and predictable.
Additional analysis further indicates that cash flows from operations are evidently more
Table 6.
WLS results on time-
series properties of
cash flow from
operations
Variables
CF
t
(1a)
CF
t
(1b)
CF
t
(1c)
CF
t
(1d)
CF
t1
0.363*** 0.350*** 0.318*** 0.313***
5.33 5.08 4.73 4.70
ROA
t
0.012 0.042
0.43 1.22
LEV
t
0.015 0.009
1.29 0.66
SIZE
t
0.008 0.003
1.64 0.67
Weighting factor Category Category Sphere Sphere
N450 450 450 450
F28.37*** 11.47*** 22.37*** 7.32***
R
2
(%) 31.55 32.31 29.32 30.78
Notes: This table reports coefficients, t-values (of two-tailed tests, in italic below) and Adj. R
2
of the
following pooled weighted least square (WLS) regressions with Huber-White robust standard errors:
CFt¼
g
0þ
g
1CFt1þ
«
t(1a) WLS weighted by category CFt¼
g
0þ
g
1CFt1þ
g
2ROAt1þ
g
3LEVt1
þ
g
4SIZEt1þ
«
t(1b) WLS weighted by category CFt¼
g
0þ
g
1CFt1þ
«
t(1c) WLS weighted by sphere
CFt¼
g
0þ
g
1CFt1þ
g
2ROAt1þ
g
3LEVt1þ
g
4SIZEt1þ
«
t(1d) WLS weighted by sphere Intercepts
remain unreported for the sake of brevity. Lagging represents a main restriction for the sample size. The
variables are winsorized at 1 and 99 percentiles. ***, ** and * stand for significances at 1, 5 and 10%
significance levels (respectively) using two-tailed tests
JFRA
18,3
582
persistent and predictable compared to earnings. Relatively higher time-series properties of
cash flows from operations compared to earnings hints on accruals’low quality. This
finding is in line to prior literature’s estimation about financial information’s low orientation
on its external use, limited accounting/audit profession and turbulent business environment,
limiting investors’ability in providing efficient earnings-based estimations.
This paper enriches accounting literature by providing the first evidence on earnings
persistence and predictability from an emerging market’s private sector perspective. By
describing the accounting/audit reform development in Georgia, the work additionally
demonstrates how the transformation towards international accounting standards is
adopted, implemented and contextualized within a developing economy. This adds up to a
debate of a long literature on accounting standards adoption in emerging countries (Zaman
Mir and Shiraz Rahaman, 2005;Chand and White, 2007a,2007b;van Helden and Uddin,
2016).
Georgia is an interesting example to examine as the country has adopted international
accounting standards back in 2005 (accounting law, supplemented as of 6/04/2005, Reg. @11)
though it now undergoes unprecedented changes of public transparency for which the
standards adoption was aimed to. Increased public transparency is likely to promote towards
the local capital markets’development if and only if a quality (valuation-useful) financial
information enters the playing field. Data analysis has revealed that more than 10% of the
entities mechanically incorrectly enter the scale of their financial items. Beyond the technical
violations, however, the reported financial numbers are not useful for the valuation purposes
either. The regulator may wish to promote entities’reporting incentives to be transposed from
taxation towards valuation purposes. Logically, at the outset of the reform, for an emerging
economy like Georgia, it is difficult to ensure high accruals quality. Georgia, as a predecessor of
the soviet era, along the years has been used to report for the internal and/or taxation purposes.
On the one hand, along with its developing capital markets, entities’reporting incentives lack
the perspective of a potential investor. On the other hand, the banking sector majorly basis its
crediting decisions on the amount of collateral and/or site visits of the entities. As there is a
lower incentive for high quality reporting among accountants, the auditors may potentially
save the game, however, prior literature has also evidenced that many of the even larger
entities avoid a more costly services of larger audit firms. Encountering these circumstances, a
simultaneous and tireless efforts should be implemented by the regulators (Ministry of
Finance –supervisor of accounting/audit reform and the National Bank of Georgia –supervisor
of the local capital markets and the financial sector) to promote accounting/audit profession
development, entities’orientation on the external use of the reports and an ultimate
development of the capital markets.
The provided analysis is a subject to several caveats. One of the limitations the work
faces is the inability of testing the hypothesis based on a longer time-series data. This
reduces the econometrical soundness of the reached conclusion. Dealing with a limited
data reduces the work’s ability to detect earnings persistence and predictability at a
country-level. Opposed to these disadvantages, however, the work, through the use of
sophisticated filtration mechanisms, has reached a well-balanced cross-sectional data,
covers 450 entities that represent a solid sample to work with particularly if anticipating
the case of an emerging and relatively smaller economy of Georgia. It needs to be also
noted that the given sample is more heavily driven by the Category II enterprises. Even
though the work uses a weighted least square regression where the number of
observations is weighted by entity categories (size) and industries, the findings need to be
cautiously generalized beyond the sample.
Earnings
persistence
and
predictability
583
Notes
1. From literature we learn that earnings are the construct of its two parts such as permanent
(expected) and transitory (unexpected) earnings, both of which affect the properties of earnings.
As found by Beaver and Morse (1978), transitory component of earnings (a result derived from a
sale of a fixed asset) may only affect current earnings. While so, future series of earnings are
influenced only by its permanent component.
2. Accounting literature has also revealed a bilateral association between accounting earnings and
stock returns. Kothari (2001) highlighted that a significant portion of the changes in earnings are
anticipated within the stock price movements prior to their release. Collins et al. (1987) show that
stock prices lead earnings and that stock returns predict earnings growth.
3. A recent assessment of the local law made by the joint efforts of SARAS and Hellenic Accounting
and Auditing Standards Oversight Board of Greece under a twinning programme of “Enhancing
Accounting and Audit Quality in Georgia”has revealed several inconsistencies compared to
international standards. These inconsistencies mainly relate to term definitions of statutory audit
of consolidated financial statements, international cooperation, liability systems, requiring a
good reputation for registering auditors and audit firms, etc (EU Directive 2006/43 and 2014/56).
A local simplified IFRS also has been revealed in some incompliance related to the definitions of
the terms of audit service fee, prohibition of non-audit services, reporting to registered auditors
on transparency, composition and independence of the audit committee and additional
requirements (EU Directive 2006/43).
4. The data enabling us to detect the levels of enforcement are yet to load for other countries.
5. Entity categorization is done based on the size of assets, revenues and an average number of
employees. Larger entities are categorized within the Category I, less large –Category II, etc.
6. For example, AI is able to automatically retrieve an invoice from an e-mail, detect the necessary
information within it and record the consequent details within an accounting software, step by
step expelling the need of human resources in traditional bookkeeping (demo can be found here).
7. Among the macroeconomic factors, which could potentially drive earnings and/or cash flows
from operations throughout the observing period (2016-2017 years) Georgia’s taxation system’s
shift towards the Estonian model should be mentioned. Started from 2017 the Estonian model of
taxation has been enforced, that exempted taxation obligations for those entities re-invested their
income back to the entity to expand its operations. The author recognizes this change could bring
significant change to earnings volume in 2017, though he argues that this change could not
violate the conclusion of this work.
8. It should be noted that for the second year of transparency (deadline: 1 October 2019), the
supervision service has changed the approach of data collection and an anecdotal evidence
suggests the share of potential technical errors should be milder.
9. We should outline that the scaling impreciseness is important for descriptive analysis. In the
regression part of our analysis, however, variables consist both of numerator and denominator,
and thus, the scaling inconsistency is automatically eliminated.
10. Sadka (2007) and Dichev and Tang (2008) argue that earnings persistence and predictability are
correlated with each other. An amelioration of predictability implies that current earnings can
better explain the changes in future earnings. If subsequent earnings are perfectly predicted, then
the error term should equal zero. As the error term biases the coefficient
g
1
(coefficient for
persistence) towards zero, the lower the error term, the higher the R
2
(predictability); therefore,
g
1
is likely to be increased. A correlation between predictability and persistence is positive. This
assertion was juxtaposed by Frankel and Litov (2009, pp. 182-183), who write: “Dichev and Tang
(2008) identify an interesting empirical relation with potential practical import and this
contribution is sufficient to outweigh the underlying lack of a causal theory. ...The algebra does
JFRA
18,3
584
not provide a mathematical connection between these parameters. Such an explanation is
province of economic theory as applied to accounting”.
11. The firm subscript iis intentionally omitted from all models.
12. The work bases on the same year’s total assets as against of the beginning year’s total assets to
maintain higher number of observations.
13. This model does not include ROA as, by its definition, this variable is measured in a same
way as scaled NI that represents an independent variable in this model. ROA will be
additionally added to a cash flow model given below. The firm-specific characteristics do
not include the growth rate that should be calculate as a change in net income under net
income. In the variable of growth, there is a necessity of three-year data for net income –
two years to calculate change in net income and the third year to calculate a one-year
lagged growth rate.
References
Aitken, A.C. (1935), “On least squares and linear combinations of observations”,Proceedings of the
Royal Society of Edinburgh, Vol. 55, pp. 42-48.
Alagardova, G. and Manuilova, N. (2015), “Report on the observance of standards and codes on
accounting and auditing”,World Bank Report, available at: http://documents.worldbank.org/
curated/en/254551468189259272/pdf/ACS13210-REVISED-Georgia-ROSC-AA-Report-FINAL-09
2015-Box393232B-PUBLIC-ENGLISH-Dissemination.pdf
Amidu, M. and Issahaku, H. (2019), “Do globalisation and adoption of IFRS by banks in africa lead to
less earnings management?”,Journal of Financial Reporting and Accounting, Vol. 17 No. 2.
Ash, E. and Strittmatter, R. (1992), Accounting in the Soviet Union: Praeger.
Asparouhov, T. and Muthén, B. (2010), “Weighted least squares estimation with missing data”, Mplus
Technical Appendix, No. 1-10.
Athukorala, S.L. and Reid, B. (2003), “Accrual budgeting and accounting in government and its
relevance for developing member countries”.
Ball, R. and Brown, P. (1968), “An empirical evaluation of accounting income numbers”,Journal of
Accounting Research, Vol. 6 No. 2, pp. 159-178.
Bananuka, J., Tumwebaze, Z. and Orobia, L. (2019), “The adoption of integrated reporting: a developing
country perspective”,Journal of Financial Reporting and Accounting, Vol. 17 No. 1.
Basu, S. (1997), “The conservatism principle and the asymmetric timeliness of earnings1”,Journal of
Accounting and Economics, Vol. 24 No. 1, pp. 3-37.
Beaver, W.H. (1968), “The information content of annual earnings announcements”,Journal of
Accounting Research, Vol. 6, pp. 67-92.
Beaver, W.H. (1970), “The time series behavior of earnings”,Journal of Accounting Research, Vol. 8,
pp. 62-99.
Beaver, W. and Morse, D. (1978), “What determines price-earnings ratios? ”,Financial Analysts Journal,
Vol. 34 No. 4, pp. 65-76.
Bernstein, L. (1993), Financial Statement Analysis, Irwin, Homewood, IL.
Board, F.A.S. (2010), “Conceptual framework for financial reporting: Chapter 1, the objective of general
purpose financial reporting, and chapter 3”,Qualitative Characteristics of Useful Financial
Information, Vol. 8.
Chand, P. and White, M. (2007a), “A critique of the influence of globalization and convergence of
accounting standards in Fiji”,Critical Perspectives on Accounting,Vol.18No.5,
pp. 605-622.
Earnings
persistence
and
predictability
585
Chand, P. and White, M. (2007b), “Embracing international financial reporting standards (IFRSs) and
SME IFRSs: implications for emerging economies”, Paper presented at the Accounting
Association of Australia and New Zealand Conference (2007).
Collins, D.W., Kothari, S.P. and Rayburn, J.D. (1987), “Firm size and the information content of prices
with respect to earnings”,Journal of Accounting and Economics, Vol. 9 No. 2, pp. 111-138.
Darby, M.R. (1972), “The allocation of transitory income among consumers’assets”, Vol. 62 No. 5,
pp. 928-941.
DeAngelo, L.E. (1981), “Auditor size and audit quality”,Journal of Accounting and Economics, Vol. 3
No. 3, pp. 183-199.
Dechow, P.M. and Ge, W. (2006), “The persistence of earnings and cash flows and the role of special
items: Implications for the accrual anomaly”,Review of Accounting Studies, Vol. 11 Nos 2/3,
pp. 253-296.
Dechow, P., Ge, W. and Schrand, C. (2010), “Understanding earnings quality: a review of the proxies,
their determinants and their consequences”,Journal of Accounting and Economics, Vol. 50
Nos 2/3, pp. 344-401.
Dichev, I.D. and Tang,V.W. (2008), “Matching and the changing properties of accounting earnings over
the last 40 years”,The Accounting Review, Vol. 83 No.6, pp. 1425-1460.
F. Dormann, C., M. McPherson, J., B. Araújo, M., Bivand, R., Bolliger, J., Carl, G., ...Daniel Kissling, W.
(2007), “Methods to account for spatial autocorrelation in the analysis of species distributional
data: a review”,Ecography, Vol. 30 No. 5, pp. 609-628.
Fama, E.F. (1970), “Efficient Capital markets: a review of theory and empirical work”,The Journal of
Finance, Vol. 25 No.2, pp. 383-417.
Farooq, K. (2018), “Robotics and distributed ledger technology (DLT) in PFM: Preparing for GovTech”,
World Bank Group.
Feltham, G.A. and Ohlson, J.A. (1995), “Valuation and clean surplus accounting for operating and
financial activities”,Contemporary Accounting Research, Vol. 11 No.2, pp. 689-731.
Foster, G. (1977), “Quarterly accounting data: Time-series properties and predictive-ability results”,
Accounting Review, pp. 1-21.
Foster, G. (1986), Financial Statement Analysis, 2/e, Pearson Education India.
Francis, J.R. (2004), “What do we know about audit quality?”,The British Accounting Review, Vol. 36
No. 4, pp. 345-368.
Frankel, R. and Litov, L. (2009), “Earnings persistence”,Journal of Accounting and Economics, Vol. 47
No. 1-2, pp. 182-190.
Freeman, R.N., Ohlson, J.A. and Penman, S.H. (1982), “Book rate-of-return and prediction of earnings
changes: an empirical investigation”,Journal of Accounting Research, Vol. 20 No. 2,
pp. 639-653.
Galimberti, J.K. and Cupertino, C.M. (2009), “Explaining earnings persistence: a threshold
autoregressive panel unit root approach”.
Graham, B. and Dodd, D.L. (1951), Security Analysis, 3d ed., McGraw-Hill Book Co., New York, NY.
Graham, B. Cottle, S. Tatham, C. and Dodd, D.L. (1962), Security analysis: Principles and technique.
Group, W.B. (2007), Report on the Observance ofStandards and Codes(ROSC) Georgia, Accounting and
Auditing, World Bank Report.
Gvaramia, N. (2014), “The issues of activity accounting and audit regulation in Georgia”,Economics
and Management, Vol. 18 No. 4, pp. 619-624.
He, X., Wong, T. and Young, D. (2012), “Challenges for implementation of fair value accounting in
emerging markets: evidence from China”,Contemporary Accounting Research, Vol. 29 No. 2,
pp. 538-562.
JFRA
18,3
586
Healy, P.M. and Palepu, K.G. (2012), Business Analysis Valuation: Using Financial Statements, Cengage
Learning.
Hitz, J.-M., Ernstberger, J. and Stich, M. (2012), “Enforcement of accounting standards in Europe:
Capital-market-based evidence for the two-tier mechanism in Germany”,European Accounting
Review, Vol. 21 No. 2, pp. 253-281.
Hope, O.K. (2003), “Disclosure practices, enforcement of accounting standards, and analysts’
forecast accuracy: an international study”,Journal of Accounting Research, Vol. 41 No. 2,
pp. 235-272.
Kaciashvili, V. (2003), “აუდიტორული საქმიანობის ფორმირებისა და რეგულირების
თავისებურებები საქართველოში გარდამავალ ეტაპზე”, Finance Research Institute.
Kothari, S. (2001), “Capital markets research in accounting”,Journal of Accounting and Economics,
Vol. 31 No. 1-3, pp. 105-231.
Kvintradze, N. (2019), “ფინანსური ინფორმაციის წარუდგენლობისთვის კომპანიები 735 000
ლარით დაჯარიმდნენ”, available at: https://forbes.ge/news/6498/kompaniebi-735-000-lariT-
dajarimdnen
Law of Georgia on Accounting, Reporting and Auditing (2016), “Law of Georgia on accounting,
reporting and auditing”.
Lenin, V. I. I. (1964), The Development of Capitalism in Russia, Vol. 3, Progress Publishers, Moscow.
Lev, B. (1983), “Some economic determinants of time-series properties of earnings”,Journal of
Accounting and Economics, Vol. 5, pp. 31-48.
Lev, B. and Ohlson, J.A. (1982), “Market-based empirical research in accounting: a review,
interpretation, and extension”,Journal of Accounting Research, Vol. 20, pp.249-322.
Lipe, R. (1990), “The relation between stock returns and accounting earnings given alternative
information”,Accounting Review, pp. 49-71.
McGee, R.W. (2008), “Reforming accounting education in a transition economy: a case study of
Armenia”,Accounting Reform in Transition and Developing Economies, Springer, pp. 237-255.
McGee, R.W. (2014), “A comment on the accounting and auditing law of Georgia”,Business and Law,
No. 3, pp. 11-13.
McGee, R.W. and McGee, R.W. (2008), Accounting Reform in Transition and Developing Economies,
Springer.
McGee, R.W. and Preobragenskaya, G.G. (2006), “Accounting reform in Ukraine”,Accounting and
Financial Systems Reform in Eastern Europe and Asia, pp. 45-73.
Metskhvarishvili, S. (2019), “Postponement of financial reporting deadline for IV category enterprises
in Georgia”,BM.ge, available at: https://bm.ge/ka/article/meotxe-kategoriis-kompaniebs-
angarishgebis-axleburad-wardgena-gadauvaddat-/40949
Nissim, D. and Penman, S.H. (2001), “Ratio analysis and equity valuation: from research to practice”,
Review of Accounting Studies, Vol. 6 No. 1, pp. 109-154.
Ohlson, J.A. (1995), “Earnings, book values, and dividends in equity valuation”,Contemporary
Accounting Research, Vol. 11 No. 2, pp. 661-687.
Penman, S.H. (2013), Financial Statement Analysis and Security Valuation, McGraw-Hill.
Pimentel, R.C. and Lima, I.S. (2010), “Time-series properties of earnings and their relationship with
stock prices in Brazil”,Business and Economics Research Journal, Vol. 1 No. 4, p. 43.
Pirveli, E. (2014), “Accounting quality in Georgia: Theoretical overview and development of
predictions”,International Journal of Business and Social Science (USA), Vol. 5 No. 3,
pp. 283-293.
Pirveli, E. (2015), Financial Statement Quality: First Evidence from the Georgian Stock Exchange,
Shaker Verlag.
Earnings
persistence
and
predictability
587
Pirveli, E. (2019a), “Assessment of the ongoing accounting and audit reform in Georgia: Regulatory
base, enforcement level and the quality of financial statements”,Caucasus University Journal
(Conference Proceedings), Vol. 9 No. 1, pp. 8-27.
Pirveli, E. (2019b), “საქართველოში მიმდინარე ბუღალტერიისა და აუდიტის რეფორმა:
პირველი შეფასება”,Forbes Georgia, Vol. 9 No. 2, pp. 82-85.
Pirveli, E. (2019c), “ფინანსური გა(უ)მჭვირვალობა საქართველოში:ინფორმაციის მოპოვების
სირთულის გაზომვა”,The Economist, Vol. 11 No. 4, pp. 1-20.
Pirveli, E. (2020), “Sectoral analysis of the economic model of Georgia. Economics and business (TSU)”.
Pirveli, E. and and Bendeliani, P. (2020), “Ownership structure analysis within georgian private sector”,
Working Paper.
Pirveli, E. and Shughliashvili, T. (2019), “Accounting and audit reform in Georgia: theoretical and
descriptive analysis”,Economics and Business (TSU), Vol. 10 No. 2, pp. 163-184.
Pirveli, E. and Zimmermann, J. (2015), “Time-series properties of earnings: the case of georgian stock
exchange”,Journal of Business and Policy Research, Vol. 10 No. 1, pp. 70-96.
Pope, P.F. and McLeay, S.J. (2011), “The European IFRS experiment: objectives, research challenges
and some early evidence”,Accounting and Business Research, Vol. 41 No. 3, pp. 233-266.
Richardson, S.A., Sloan, R.G., Soliman, M.T. and Tuna, I. (2006), “The implications of accounting
distortions and growth for accruals and profitability”,The Accounting Review, Vol. 81 No. 3,
pp. 713-743.
Sabauri, L. (2018), “Approval and introduction of the international financial reporting standards
(IFRS) in Georgia: challenges and perspectives”,Journal of Accounting and Marketing,
Vol. 07 No. 02.
Sadka, G. (2007), “Understanding stock price volatility: the role of earnings”,Journal of Accounting
Research, Vol. 45 No. 1, pp. 199-228.
Shiller, R.J. (2003), “From efficient markets theory to behavioral finance”,Journal of Economic
Perspectives, Vol. 17 No. 1, pp. 83-104.
Sloan, R.G. (1996), “Do stock prices fully reflect information in accruals and cash flows about future
earnings?”,Accounting Review, pp. 289-315.
Stigler, G.J. (1963), “Capital and rates of return in manufacturing industries: a study by the national
bureau of economic research, NY”,
Tawiah, V. (2019), “The state of IFRS in Africa”,Journal of Financial Reporting and Accounting, Vol. 17
No. 4, pp. 635-649, doi: 10.1108/JFRA-08-2018-0067.
Thiagarajan, S. (1989), “Economic determinants of earnings persistence”,
van Helden, J. and Uddin, S. (2016), “Public sector management accounting in emerging economies: a
literature review”,Critical Perspectives on Accounting, Vol. 41, pp. 34-62.
Watts, R.L. and Zimmerman, J.L. (1978), “Towards a positive theory of the determination of accounting
standards”,Accounting Review, pp. 112-134.
World Bank, C. (2007). “Report on the observance of standards and codes (ROSC) Georgia”,Accounting
and Auditing., available at: http://documents.worldbank.org/curated/en/863431468771552971/
pdf/389770ENGLISH01unting0ROSC01PUBLIC1.pdf
World Bank, C (2008), “Ukraine - Report on the observance of standards and codes (ROSC): accounting
and auditing (in ukrainian)”, available at: http://documents.worldbank.org/curated/en/
367341468337802893/pdf/626030WP00UKRA0e0Box0361486B0PUBLIC.pdf.
World Bank, C (2013), “Moldova 2013 a&a ROSC”, available at: http://siteresources.worldbank.org/
EXTCENFINREPREF/Resources/4152117-1232359743213/5751608-1279318020896/Moldova_
ROSC-AA_en.pdf
Wumburidze, L. (2013), “მოსაზრებები ბუღალტრული აღრიცხვისა და ფინანსური
ანგარიშგების აუდიტის შესახებ”,
ბუღალტრული აღრიცხვა
, Vol. 1, pp. 1-5.
JFRA
18,3
588
Xie, H. (2001), “The mispricing of abnormal accruals”,The Accounting Review, Vol. 76 No. 3,
pp. 357-373.
Zaidi, S.K.R. and Huerta, E. (2014), “IFRS adoption and enforcement as antecedents of economic
growth”,International Journal of Accounting and Financial Reporting, Vol. 4 No. 1, pp. 1.
Zaman Mir, M. and Shiraz Rahaman, A. (2005), “The adoption of international accounting standards in
Bangladesh: an exploration of rationale and process”,Accounting, Auditing and Accountability
Journal, Vol. 18 No. 6, pp. 816-841.
Further reading
Bernstein, L.A. and Wild, J.J. (1993), Financial Statement Analysis: Theory, Application, and
Interpretation, Irwin, Homewood, IL.
Dechow, P.M. and Schrand, C.M. (2004), “Earnings quality”,CFA Digest, Vol. 34 No.4, pp. 82-85.
Richardson, S.A., Sloan, R.G., Soliman, M.T. and Tuna, I. (2005), “Accrual reliability, earnings
persistence and stock prices”,Journal of Accounting and Economics, Vol. 39 No.3, pp. 437-485.
Corresponding author
Erekle Pirveli can be contacted at: epirveli@cu.edu.ge
For instructions on how to order reprints of this article, please visit our website:
www.emeraldgrouppublishing.com/licensing/reprints.htm
Or contact us for further details: permissions@emeraldinsight.com
Earnings
persistence
and
predictability
589
