This study analyzes the influence of national culture, more specifically by look at Hofstede’s cultural dimensions (2017), on the accounting conservatism level of companies from different countries. We sought to test empirically the theoretical approach draft by Gray (1988) that the culture influences accounting development, which can occur differently in each country.

Surveys that verify the countries cultural diversity in the business, companies, and accounting are proposed as greatly important by several national and international studies (Gray, 1988; Weffort, 2005; Chand, 2012; Riahi & Omn, 2013; Kitching, Mashruwala & Pevzner, 2016; Rodrigues et al., 2017), because they make a contribution by identifying divergences and similarities in the way accounting is performed in the countries studied.

Therefore, this study brings another perspective of analysis upon the quality of the accounting information, where focus often falls upon agency problems, dealing with an exogenous aspect of countries’ cultures, and being capable of influencing the levels of accounting conservatism presented by companies around the world.

Accounting conservatism can influence accounting information quality (AIQ), improving internal control in organizations (Goh & Li, 2011), serve as a corporate management mechanism, lower the managers’ opportunistic behavior (Watts, 2003) to counterbalance the managers’ optimism excess about overestimation of profits, which is more dangerous than their underestimation in terms of penalties in their disclosure (Hendriksen & Van Breda, 1999), or to reduce the discretion and subjectivity from managers during the accounting profit disclosure (Ball, Kothari, & Robin, 2000).

On the other hand, it can skew the recognition of facts, reducing the AIQ, because its practice increases the informational asymmetry of contracts, raising external capital costs and favoring the increase of financial constraint of companies that need to raise funds (Hendriksen & Van Breda, 1999).

Several variables can affect the usage of the accounting conservatism in the financial statements, such as the adoption of a new accounting standard (Santana & Klann, 2016); the votes concentration (Sarlo Neto, Rodrigo & Almeida, 2010), because the agreement executed between the shareholders impacts on the reduction of the companies conservatism level; as well as the proper culture, since the historical development of each country, the level of the government control, the structure of the business property, in conjunction with the cultural influence received are the possible causes of the different accounting practices used by different countries (Cieslewicz, 2014).

The term ‘culture’ equates to a set of meanings and signs created inside a society within which individuals interact with each other. This group embraces the forms of work, leisure, housing, the language, the religion, the cuisine, the power relations, the family structure, and social systems, among others. (Chaui, 2008). According to Hofstede (1980), the culture does not consist of something that can be inherited; it is, however, obtained through a social environment that involves the individuals. It can be built, rebuilt and transformed upon the environment’s needs because it does not relate with just one individual, but with all the organization that it is engaged with (Hofstede, 1980; 1991; 2001).

Accounting affects the social and cultural environments in which it takes place, and one can view the culture as a fundamental element for the comprehension of the regulations and values that govern those environments (Hopwood & Miller, 1994). Thus, the cultural differences between the individuals can affect the relations inside of the organization; the higher the difference is, more disparate are the administrative, organizational, and strategic practices (Ariño & De La Torre, 1998).

According to Gray (1988), culture influences the accounting progress in companies, besides impacting on their distinct practices from country to country. The literature, in turn, recognizes the influence of the cultural factor in accounting without excluding the eventual impacts that it can cause in the financial report (Justino, Albuquerque, Quirós & Rodrigues 2017).

This way, the divergences among countries earn space as an object of investigation and the realization of these studies become important because they allow the visualization of the environmental factors that describe the divergences in accounting (Verma, 1998). Based on the theoretical predictions that point out a relation between the national culture and the accounting, survey research was drawn up: What is the influence of national culture on the level of accounting conservatism of companies? In order to answer the survey research, this study aims to evaluate the influence of Hofstede’s cultural dimensions (2017) on the level of the accounting conservatism of companies from different countries.

In general, the results indicate a high dispersion of accounting conservatism observed in the companies analyzed. Furthermore, such companies are stemming, in most cases, from countries with greater power distance, with a culture more concerned with female and collective values, with a greater aversion to uncertain situations, long-term oriented culture, and less likely to indulge. As to the cultural influence in the accounting conservatism, the results show that companies from countries that are more individualistic and with an indulgent behavior can present less conservative accounting practices, while those from countries with a more long-term oriented culture show more conservative behavior.

From a theoretical perspective, the realization of this study aims at contributing to the investigations about the AIQ, seeking to expand the limits of knowledge regarding the cultural factors that can influence the existence and level of accounting conservatism in different countries. This study sheds light on the discussion of the influence of accounting principles on the quality of accounting information, by showing that external factors, to companies, such as the country’s culture, can influence the level of accounting conservatism in accounting statements of companies from different countries.

Regulatory bodies must consider this in their quest for greater comparability of accounting information internationally, as well as analysts and investors when analyzing companies from different countries. The results of this study point out that companies from different countries might present different levels of accounting information quality, even though both use international accounting standards or US accounting standards, for example.

As for the literature breakthrough, Gray’s assumptions (1988) about the accounting conservatism and the national culture have already been analyzed by several authors (Schultz & Lopez, 2001; Doupnik & Richter, 2004; Doupnik & Riccio, 2006; Tsakumis, 2007), but only in the form of experiments and interviews, which evaluated the interpretations and judgments by accounting professionals. This study overlaps with them, when utilizing data from accounting statements, just like Salter and Lewis (2011), who has already presented the preparers’ judgment at the moment of its elaboration.

The difference between this study and Salter and Lewis’ (2011) is that this one required 32 countries and took into account the six cultural dimensions of Hofstede (2017), while the former relied on 14 countries and considered just three cultural dimensions.

Watts (2003) says that the practice of conservatism originated in the 19th century, since the accounting professionals had to provision all the probable loses of companies before performing any distribution of assets. However, for Basu (1997), its origin was in the 15th century, and, even with numerous criticisms to its use, conservatism is still widely used in the last three decades (Watts, 2003).

Gray (1980) defines conservatism as the preference, when using a cautious approach, to measure future events in opposition to the competitive and optimistic market as a form of risk prevention. He forked the concept of conservatism into two sub-concept groups: conditional conservatism and unconditional conservatism (Beaver & Ryan, 2005).

Conditional conservatism represents the skewed recognition of bad news, occurring earlier than good news (Basu, 1997; Ahmed et al., 2000; Ball et al., 2000; Ball & Shivakumar, 2005; Brown Jr, He & Teitel, 2006). Its use benefits the production of more transparent and loyal information to the investors, without being overly optimistic (Lopes, 2002). The asymmetric condition to the recognition of gains and losses brought by the practice benefits the users of the accounting statements, essential to mitigate the manager’s opportunistic habits (Watts, 2003).

Unconditional conservatism results from the choice between two forms of measurement and recognition of events, both being equally valid, where the chosen option is the one that results in the lowest valuation of assets. It is based on uncertainty, as it affects the continuity of companies to future demands (Hendriksen & Van Breda, 1999). In this study, the focus is on the conditional conservatism, taken from Ball e Shivakumar’s model (2005).

Accounting conservatism, as a feature of AIQ, has been widely investigated in the academic environment, with researchers attempting to look at which factors determine its use. One of the first researchers to consider the culture as an important variable for the presence of conservatism among countries was Gray (1988). From theoretical conceptions, the author related conservatism to needs and desires, so that if a country aims to reduce uncertainties and be less individualistic, it will show a great tendency to be conservative.

Gray’s study (1988) addressed three of the four dimensions identified by Hofstede (1980): Uncertainty Avoidance (UAI), individualism versus collectivism (IVC), and masculinity versus femininity (MAS). The author theoretically maintained that countries with higher UAI, with more collectivist and female values, were more conservative in preparing their financial statements.

To Hofstede (1980), the dimension of UAI relates to the level of social stress when facing the unknown future. Thus, this dimension is not about social behavior, but about the human search for the truth, and indicates, culturally, to what degree members of a society are comfortable or uncomfortable in different and unknown situations. When dealing with UAI, Gray (1988) claims that this dimension is the most positively related to accounting conservatism. Consequently, the preference for adopting more conservative measures ahead of profits is consistent with a scenario of high UAI (H₁), as a concern with prudence and safety leads to cautious approaches in recognizing future events.

The IVC dimension is characterized by Hofstede (1980) as the integration of individuals into primary groups and reveals the degree of integration of individuals within a group. In an individualistic organization, its members have a more competitive, more assertive view and there are no ties between its components; each one takes care of their own business and their own immediate family. On the other hand, the collectivist view, where it is noticeable an integration and cohesion among groups in society, the family vision extends to uncles, aunts, and grandparents, and the meaning of the collective has nothing to do with the politic sense, referring to the state, but it has a connection with the group.

One can link the individualism defined by Hofstede (1980) to the overconfidence of the individual; in other words, in societies with high IVC, more decisions are made by the individual driven by the overconfidence (Chaui, 2008).

Gray (1988) linked accounting conservatism with the IVC, noting that societies with more individualistic traits have less accounting conservatism. (H₂). This negative relation was also verified by Salter and Niswander (1995) and Salter and Lewis (2011), while Sudarwan and Fogarty (1996) observed positive relations between the IVC and the accounting conservatism derived from a study in Indonesia.

The dimension called MAS engages in, according to Hofstede (1980), the division of emotional roles between men and women, which may reflect different visions within a society. According to the author, women differ less among societies than men, of which they retain values such as modesty and affection, while men are more assertive and competitive. The assertive pole has been called “masculine” while the modest and caring has been called the “feminine” pole (Hofstede, 1998).

For Gray (1988), the MAS is negatively associated with conservatism, because masculine societies seek for more earnings, performance, and efficiency, exerting a stronger influence on the managers and statement preparers, which in this case, would tend to have a less conservative bias in the earnings measures (H₃). In line with the assumptions described by Gray (1988) are the results found by Salter and Niswander (1995) and Salter and Lewis (2011), while Sudarwan and Fogarty (1996) did not find a significative relation between these variables.

Hofstede (1980) identifies problems involving human inequality inside the power distance dimension (PDI), which reflects the extent that members with less power inside organizations accept and even expect a power that is unequally distributed. Goodwin, Goodwin, and Fiedler (2000) identified that individuals from countries with high PDI were less likely to be involved in the practice of earnings management, representing a more faithful and realistic portrayal of the company.

Almeida, Lopes, and Corrar (2011) point out that the higher the GR is, the lower the level of accounting conservatism observed. Thus, assuming a negative relationship between GR and a high PDI and a negative relationship between GR and accounting conservatism, a positive relationship between the high PDI and the accounting conservatism practice is assumed (H₄).

The fifth dimension identified by Hofstede et al. (2008), the long-term orientation versus short term orientation (LTO), relates with the choice of focus to people’s efforts: the future or the present and the past. According to the authors, a society focused on cultural values that care about LTO has a forward-looking vision, saves economically, and shows values such as persistence and austerity, unlike the short-term orientation, which focuses on the past and the present.

This way, in a society, where values are short term oriented, the most important results are those perceived and evidenced in the short term. Thus, in the presented context, the higher the LTO of a country or society is, the higher the level of accounting conservatism will be presented by it (H₅).

The sixth dimension described by Hofstede, Hofstede, and Minkov (2010), indulgence versus restraint (IND), has not yet been inside the scope of studies in accounting literature involving the practice of accounting conservatism. These authors define this dimension as a tendency for individuals to allow themselves to be human figures, enjoying life.

Linking the dimension with the practice of conservatism, it is believed that more indulgent societies, which are more likely to enjoy life, are less conservative, because according to Gray (1988), the conservatism leads to the use of prudence and caution, where the more conservative accountants opt for more cautious accounting, in order to deal with the uncertainty of future events (H₆).

Based on the above, Figure 1 presents the expected behavior for accounting conservatism in the analyzed companies to each dimension of the national culture of the countries.

Figure 1. Source: own elaboration

We expect that the dimensions of uncertainty avoidance (UAI), power distance (PDI), and long-term orientation (LTO) are positively related to the accounting conservatism level, while individualism (IVC), masculinity (MAS) and indulgence (IND) present the opposite behavior.

However, neither the previous literature nor this study analyzes the interdependence of these cultural dimensions and its effect on the accounting conservatism. Despite this limitation, it is possible to assume that the strongest cultural dimension of a certain nation can impose itself upon others regarding the accounting conservatism.

For example, one can assume that an extremely male country, even though it has moderate cultural traits of uncertainty aversion, will tend to behave less conservatively. On the other hand, a country with a high-power distance, even though if it has individualistic traits, for example, will tend to have greater traits of accounting conservatism. Therefore, the cultural arrangement between the dimensions in each nation can be very diverse, implying possible different effects on the accounting conservatism in each country.

Through the Thomson Reuters®’ database, the study population includes 8.769 public companies from 32 countries, with data extracted from 2010 to 2016, totalizing 61.383 observations. We excluded from the sample outliers (above 3 standard-deviations) and companies with total assets less than US$ 100 million, as detailed in Table 1:

Table 1.

Description			Companies	Observations		% Population
Population			8.769	61.383		100%
(-) Exclusion outliers (above 3 standard-deviations)			(29)	(201)		0,3%
(-) Companies with Total Assets < US$ 100 million			(942)	(6.698)		10,9%
(=) Final sample			7.798	54.484		88,8%
Country		Population		Sample		%	% Sample
		Observations	Companies	Observations	Companies
1	USA	1.099	157	1.095	157	99,6%	2,0%
2	Brazil	1.519	217	1.400	200	92,2%	2,6%
3	Canada	5.068	724	1.808	259	35,7%	3,3%
4	Mexico	546	78	546	78	100,0%	1,0%
5	Argentina	280	40	254	37	90,7%	0,5%
6	Colombia	301	43	287	41	95,3%	0,5%
7	Chile	693	99	636	91	91,8%	1,2%
8	Peru	616	88	582	84	94,5%	1,1%
9	China	11.739	1.677	11.679	1.670	99,5%	21,4%
10	Japan	1.547	221	1.537	220	99,4%	2,8%
11	India	147	21	147	21	100,0%	0,3%
12	South Korea	8.806	1.258	8.790	1.257	99,8%	16,1%
13	Indonesia	1.197	171	1.136	163	94,9%	2,1%
14	Saudi Arabia	623	89	556	80	89,2%	1,0%
15	Taiwan	9.471	1.353	9.282	1.332	98,0%	17,0%
16	Thailand	2.429	347	2.316	331	95,3%	4,2%
17	Philippines	588	84	576	83	98,0%	1,1%
18	Singapore	2.128	304	1.623	232	76,3%	3,0%
19	Nigeria	266	38	237	34	89,1%	0,4%
20	Egypt	322	46	292	42	90,7%	0,5%
21	South Africa	420	60	381	55	90,7%	0,7%
22	Morroco	98	14	91	13	92,9%	0,2%
23	Germany	1.981	283	1.160	166	58,6%	2,1%
24	United Kingdom	896	128	888	127	99,1%	1,6%
25	France	2.268	324	1.294	185	57,1%	2,4%
26	Italy	1.008	144	851	122	84,4%	1,6%
27	Russia	1.358	194	1.357	194	99,9%	2,5%
28	Turkey	1.127	161	1.126	161	99,9%	2,1%
29	Netherlands	364	52	286	41	78,6%	0,5%
30	Switzerland	686	98	592	85	86,3%	1,1%
31	Poland	1.351	193	1.346	193	99,6%	2,5%
32	Belgium	441	63	333	48	75,5%	0,6%
	Total	61.383	8.769	54.484	7.798	88,8%	100%

Source: Survey Data.

Table 2 shows the scores of each country in the sample, to each of the six dimensions searched by Hofstede, identified in surveys developed by the author inside the countries, through the questionnaires applied.

Table 2.

Country	PDI	IVC	MAS	UAI	LTO	IND
USA	40	91	62	46	26	68
Brazil	69	38	49	76	44	59
Canada	39	80	52	48	36	68
Mexico	81	30	69	82	24	97
Argentina	49	46	56	86	20	62
Colombia	67	13	64	80	13	83
Chile	63	23	28	86	31	68
Peru	64	16	42	87	25	46
China	80	20	66	30	87	24
Japan	54	46	95	92	88	42
India	77	48	46	40	51	26
South Korea	60	18	39	85	100	29
Indonesia	78	14	46	48	62	38
Saudi Arabia	95	35	60	80	36	52
Taiwan	58	17	45	69	93	49
Thailand	64	20	34	64	32	45
Philipinnes	94	32	64	44	27	42
Singapore	74	20	48	8	72	46
Nigeria	80	30	60	55	13	84
Egypt	70	25	45	80	7	4
South Africa	49	65	63	49	34	63
Morroco	70	46	53	68	14	25
Germany	35	67	66	65	83	40
United Kingdom	35	89	66	35	51	69
France	68	71	43	83	63	48
Italy	50	76	70	75	61	30
Russia	93	39	36	95	81	20
Turkey	66	37	45	85	46	49
Netherlands	38	80	14	53	67	68
Switzerland	34	68	70	58	74	66
Poland	68	60	64	93	38	29
Belgium	65	75	54	94	82	57

Source: Hofstede (2017).

Measured from 0 to 100, a high score on the PDI dimension indicates how much the country accepts the distance between the members of a society/organization; the individuals expect the powerful members to guide them in their work. For this dimension, the country that achieved the highest score was Saudi Arabia (95), while Switzerland presented the lowest score.

South Korea achieved the highest score (100) on the LTO dimension. It is necessary to highlight that the country reached the maximum score on the Hofstede’s scale and according to the author, this is because Korea is one of the most pragmatic and long-term societies in the world.

The scores to the IND dimension also presented important divergences, with the country with the highest score, indicating the disposition of “enjoying life”, was registered in Mexico (97), while the lowest score was Egypt (4). The score verified in Mexico indicates a definitive cultural tendency toward indulgence, which shows a desire to accomplish the impulses, showing a positive attitude, and a tendency towards optimism.

We used Ball and Shivakumar’s model (2005) to measure the conditional conservatism, which measures the asymmetry of recognition between gains (good news) and losses (bad news). The model assesses the conservatism according to the occurrence of accounting earnings reversal, which allows the identification of transitional components on gains or losses. In this context, the lowest frequency of recognition of timely loss means the lowest quality of the accounting statements.

ΔNIit= α0 + α1DΔNIit-1+ α2ΔNIit-1+ α3DΔNIit-1 × ΔNIit-1+ ɛt (1)

In which:

ΔNI _it : variation in the accounting operation profit of the company i of the year t-1 to the t weighted by total asset value at the beginning of the year t;

ΔNI _it-1 : variation in the accounting operation profit of the company i of the year t-2 to the t-1 weighted by total asset value at the beginning of the year t-1;

DΔNI _it-1 :dummy variable assuming value 1 if ΔNI_it-1<0, and 0 in other cases;

ɛ _t : standard error of the regression.

Initially, equation 1 was run for each country, separately, in order to identify the conservatism present in the sample. According to the model, which was tested for each country, the anticipation of economic gains for when there is cash flow allows one to recognize gains as a persistent component of accounting profit, which tends not to be reversed and implies on α₂ being zero. However, when the recognition of gains occurs in a timely manner, they are seen as a factor of increase in the transitional component of profit, with a tendency to be reversed in subsequent periods, which is perceived by α_2<0.

Furthermore, the recognition of losses represents earnings’ transient decreases, where α₂ + α₃ < 0, and if the losses are recognized timelier than gains, it is considered as a proxy of the conditional accounting conservatism, then α₃< 0 (Ball & Shivakumar, 2005). For better data visualization, we inverted the dependent variable, so in this context, α₃ > 0 indicates conservative accounting practice.

Ball and Shivakumar’s model (2005) was adjusted to measure the effect of the national culture on accounting conservatism, by the national culture variables, represented in Equation 2 by the initials CULT, representing the score of each dimension obtained by the country in the Hofstede’s survey.

ΔNIit = α0 + α1 DΔNIit-1 + α2ΔNIit-1+ α3 DΔNIit-1 × ΔNIit-1+ α4CULTit + α5CULTit × DΔNIit-1 + α6CULTit × ΔNIit-1 + α7CULTit × DΔNIit-1 × ΔNIit-1+ ɛit (2)

Equation 2, which covers each one of the six cultural dimensions of Hofstede’s sample, was tested separately for each dimension, using observations in a period from 2010 to 2016. In this equation, the variables of interest are α₆ and α₇. The identification of the timely recognition of economic losses related to each cultural dimension implies their recognition with a decrease in transient profits verified at α₆+ α₇<0. The hypothesis of the existence of conservatism related to each one of the cultural dimensions tested, which indicates that the economic losses are recognized faster than the gains, results in α₇<0. Thus, α₆ refers to the early recognition of gains and α₇ to the relation between the accounting conservatism and the cultural dimension surveyed.

For control variables, IFRS was used (1 for adopting companies and zero otherwise) and size (log of Total Assets). The IFRS adoption by countries may influence the level of companies accounting conservatism; therefore, it would be prudent, when comparing such levels among the companies of the sample, to consider a possible effect of the IFRS. As for the size, it is a common control variable in accounting conservatism studies.

The present study had a goal to evaluate the influence of Hofstede’s cultural dimensions (2017) on the level of accounting conservatism of companies from different countries. To this end, we researched 7.798 companies and 54.484 observations extracted from 32 countries.

This study sought empirical evidence that could contribute to the reflection and discussion on AIQ. Understanding how conservatism appears in financial statements is important because it assists in the economic and financial analysis of companies, contributes to the regulatory processes of accounting, and promotes the issuance of higher quality financial statements.

In general, through the findings and corroborating the literature, we conclude that cultural factors from country to country do influence accounting information. Specifically, in the case of this study, certain cultural dimensions, especially IVC, LTO, and IND are related to accounting conservatism, the former positively, and the latter negatively.

One of the main implications of this study is that the comparability of the companies’ financial statements from different countries may not be a simple task to achieve, even if accounting standards such as IFRS is used due to cultural differences among countries, affecting the AIQ.

Another important implication of the study is that the analysis of accounting quality at the international level, specifically in the case of the accounting conservatism, has to take into account the cultural factors of each country, which should be considered by the respective national regulatory bodies, in order to produce accounting principles more appropriate to the reality of each nation.

Finally, it is important to highlight that this study has its limitations. The first one is the assumption of perennial national culture. This way, as recommendations for future studies, it may be interesting to consider that a country’s national culture may change in the long-term. Thus, studies with longer time horizons could identify countries’ cultural changes and how this would affect a long-term AIQ.

Another limitation of this study is that it does not consider regional differences within each country. Thus, a point to be discussed in future research is the condition that, even within a single country, different cultures may exist, and these may influence accounting conservatism differently.