Examining the Long-term performance effects of early IFRS adoption

Published: Last Edited:

This essay has been submitted by a student. This is not an example of the work written by our professional essay writers.

For years beginning on or after January 1, 2005, all publicly listed European firms were required to prepare IFRS compliant consolidated financial statements (Regulation (EC) 1606/2002). Nearly 7,000 listed European Union (EU) firms were affected by this regulatory enforcement.

In Turkey, as part of the harmonization process with the EU, all listed firms were also obliged to prepare IFRS compliant consolidated financials from 01.01.2005 onwards. During the transition period, firms were given the option of abiding by this regulation from 31.12.2003 onwards, hence adopting it 'early'.

However, we do not know how such a bold move affected the Turkish firms' performance. This provides the main motivation for the paper. The study explicitly investigates the long-term performance effects of such early IFRS adoption and looks into the factors that could be driving the results.

The remainder of the paper is organized as follows. Section 2 outlines the theoretical background, Section 3 describes the data and methodology and Section 4 presents the findings and finally Section 5 concludes.


This section explains the theoretical background the paper is based on. It looks at the issue from the Positive Accounting Theory (PAT) and from the Agency Theory perspectives and points out some common predictions of both theories. Additionally, it provides a brief review of the studies in the area.

According to Positive Accounting Theory (PAT) "there is a relation between firms' accounting choice and other firm variables" (Watts and Zimmerman, 1990, p.132). The foundations of this theory goes back to the pioneering works of Watts and Zimmerman (1978, 1979) exploring the links between accounting choice and contracting costs. Later on, Fama and Jensen (1983a, 1983b) suggested that minimizing contracting costs was crucial in firm survival. Additionally, the term 'contracting costs' was used by Klein (1983) to specifically refer to those agency costs that are relevant in explaining the contract which shape organizational choice.

In the empirical test of this theory, the political cost hypothesis has received considerable attention. This hypothesis predicts that large firms are more likely to use profit-reducing accounting methods. The size effect and other predictions of this theory received severe criticisms and led to a multitude of research papers [1] .

This debate still remains popular. Basu (2001) reports that small firms in the US are more conservative and are more reluctant to adopt new accounting methods. More recently, Missonier-Piera (2004) provide empirical evidence showing that accounting choices can be explained by a positive approach, in the Swiss context.

Empirical works based in the Agency Theory, find decreases in the cost of capital (Francis et al., 2004) and positive economic effects (Bartov et al., 2005) associated with International Accounting Standards (IAS) adoption. Daske and Gebhart (2006) and Barth et al. (2008) also find that adoption of IFRS improves the accounting quality of publicly listed firms in Europe.

Extending this line of research, one would expect a major change in accounting method, such as voluntary IFRS adoption, to affect firm performance. This is explicitly tested in the following sections.


The sample is constructed from the Istanbul Chamber of Industry 500 (ICI 500) list which periodically ranks the top 500 industrial firms in Turkey. There are 48 firms which have opted to prepare IFRS compliant consolidated financials (early-adopters), comprising the 'test group' of firms. Also, from the ICI500, a size, sector and performance matched sample is constructed as the 'control group' of firms.

In order to test whether voluntary (in other words 'early') adoption of IFRS affected long-term performance of firms, variables used in prior studies to proxy for firm size, short-term performance and ownership were employed as control variables. A number of dummy variables were also used to control for quotation status, operating sector and city of incorporation.

Initially, differences in medians between the test and control group are investigated through a Mann-Whitney U test and the statistically significant performance variables are further explored through a multivariate linear regression. A sensitivity analysis involving the use of alternative variables to measure long-term performance are considered, followed by a detailed examination of high performers through a univariate and multivariate logistic regression framework.


Table 1 presents the size, short-term performance, long-term performance, ownership and categorical variables used in the study. Sample firms had average total assets of 436 million TL for 2003, which was the year firms had the option of voluntarily adopting IFRS. For the same year, sample firms had average sales of 645 million TL, average profit before interest and taxes of 45 million TL, and average shareholders' equity of 213 million TL. While, in terms of short-term performance, sample firms had sector-adjusted average return on equity (ROE), return on assets (ROA) and return on sales (ROS) of 7.3%, 0.6% and 1%, respectively, in 2003.

The mean value of the average ROS (ROA) over the next 3-yr period following voluntary adoption was 6.5% (6.8%). Majority of the firms in the sample were privately owned. The average level of private ownership was 77%, while the average level of foreign ownership was 19%. The average state ownership of 6%, was due to firms which have been previously privatized or on the privatization agenda for the current year [2] .

In the sample there were 48 IFRS early-adopters, which comprised the 'test firms'. Out of these 48 test firms, 41 were quoted at the ISE, hence resulting in a total of 82 quoted firms in the overall sample. Looking at the distribution of the sample firms' primary sectors, it is fair to say that they are fairly representative of the firms on the ICI500 list. The only exception is the utilities sector, with a very high level of state ownership and specific business environment. In terms of the city of incorporation, Istanbul is over-represented, not only in our sample but also in the population (ICI500) that the sample was drawn from. Overall, the distribution of the city of incorporation resembles that of the ICI500.

Table 2 presents the results of the Mann-Whitney U test for the difference in medians of the short-term and long-term performance variables for the test versus the control group of firms. There is no significant difference between the short-term performance of the test and the control group.

However, the long-term performance of the test group is statistically significantly higher, for both ROA and ROS. The average 3-yr ROS is 6.5% for the test group and 2.7% for the control group (p:0.048). Similarly, the average 3-yr ROA is 7.6% for the test group and 4% for the control group of firms.

In order to find out which factors could be driving these findings, multivariate regression models are constructed, using alternative specifications of long-term performance. In Table 3, for each dependent variable, AvgROS-3yr and AvgROA-3yr, there are 2 different multivariate regression models using the enter method and the backward stepwise method.

Using the Enter method, the independent, control and dummy variables together explain 2.5% of the variation in AvgROS-3yr (p-value of F: 0.316) and 18.8% of the variation in AvgROA-3yr (p-value of F: 0.012). The only significant relationship is between the short-term performance variable and AvgROA-3yr.

Using the backward method, there is a significant and positive relationship between the size variable and AvgROS-3yr, as well as the short-term performance variable and AvgROA-3yr. There is also a positive relationship between the foreign ownership variable and AvgROA-3yr, which is marginally insignificant (p:0.055).

The IFRS-EarlyAdopt dummy is not significant in any of the regression models. Overall, the findings are qualitatively similar using alternative specifications for the size and short-term performance variables.

Table 4 presents the findings of the sensitivity analysis, using growth variables to measure long-term performance. The test firms outperform the control firms in all three measures of long-term performance using growth variables. The 3-yr average for sales growth is 21.5% for test firms, compared to 20.4% for control firms. However, the difference is marginally insignificant. Using 3-yr average income growth and asset growth, there is a statistically significant difference between test and control firms.

Table 5 presents the results of the logistic regression using HIGH-AvgAssetGrwth-3yr as the dependent variable, which takes on a value of 1 if the 3-yr average asset growth of the firm is higher than the median (0.229) and zero otherwise. The independent, control and dummy variables are the same as the multivariate linear regression analysis in Table 3.

In the univariate analysis each variable is regressed on the dependent variable one-by-one and only the statistically significant variables are presented in the first 2 columns of Table 5 for space reasons. There is a negative relationship between the size variable and the growth variable (0.024). In other words, firms that are already large are less likely to have average asset growth higher than the median in the long run. This finding makes sense intuitively.

There is a positive relationship between the IFRS-EarlyAdopt variable and the growth variable (p:0.000). This relationship holds in both the univariate and the multivariate logistic regression model. In fact, the IFRS-EarlyAdopt variable alone explains 18.4 % of the variability in the growth variable (P of model F value:0.000). The multivariate enter model, including all the independent, control and dummy variables, explains 52.4% of the variability in the growth variable. However, the only significant variable in both the enter and backward stepwise model is the IFRS-EarlyAdopt variable.


Overall, comparing IFRS early-adopters with a sector, size and performance matched control sample of firms, there is a statistically significant difference (at the 5% level) between the two groups. In investigating the drivers behind the early-adopters better performance we find evidence in support of the size effect and the short-term performance effect. In other words, large firms and firms currently performing better than their peers are more likely to outperform the others in the long-run. At this point it looks as if being an early-adopter has no effect on firm performance when all the other independent, control and dummy variables are taken into account.

However, when long-term growth variables are used to proxy for firm performance, the early adopters clearly outperform their peers. Furthermore, early-adopters are more likely to be in the high performance group even when all the other factors that could potentially have an impact on performance are taken into account. Therefore, the variables used to measure performance are crucial in assessing the impact of early IFRS adoption. Overall, the findings support the positive accounting theory.

A major limitation of this work is that it focuses on accounting measures of performance. It would be very interesting to see the stock market performance of these early-adopters. Unfortunately, it was impossible to conduct such analysis since the sample included firms that are not publicly traded. Future work measuring performance for longer time periods and using alternative variables would contribute to our understanding of the 'early adoption' issue.